Piano World Home Page

128 note polyphony question

Posted By: Stephano

128 note polyphony question - 06/05/16 02:40 AM

Is there any known piece in classical music which would require more than 128 note polyphony? Is there any piano sonata, or piano concerto, or etude, or any other composition, where 128 polyphony of a digital piano would not be sufficient?

To me, it doesn't seem like there is, but I still want to hear some opinions. Some digital pianos praise their higher polyphony numbers, which is why I'm asking.

Perhaps this is only an advantage if you play a piano concerto with the orchestra playing in the background on the piano, or similar?
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/05/16 02:58 AM

It depends on how much you use the pedal... wink

I suspect very high polyphony numbers are largely a marketing gimmick (like ultra high sampling rates in digital recording).

On the other hand, if you seriously intend using a MIDI rendition of orchestral music through the piano (rather than mixing in an analogue or digital track), then 128 won't be enough by a long shot if you want to have a serious orchestra with reasonably realistic decay timing (20 or so voices playing together means you run out of polyphony 6 notes into the piece, never mind the piano).
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/05/16 03:31 AM

It does depend on how much you use the pedal. You can easily hit 128 voices by putting the pedal down and doing a glissando. An acoustic piano does not suffer from voice limitations, therefore it is a reasonable objective of a digital piano not to suffer from note-cutoff/voice stealing.

Some digital pianos and keyboard instruments allow you to do splits and layers. For example - a string section sound along with the piano samples. The greater the number of available voices - the more capable the instrument is at performing splits and layers without note-cutoff/voice stealing. I've done many gigs with one keyboard where I layer acoustic piano, electric piano, strings, and brass instruments and use a volume fader on each layer to bring in or take out each of these sound groups. If I didn't have the poly, this would not be possible.

Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 03:48 AM

I wouldn't be using layered sounds, MIDI or anything like that. It'd be piano playing only.

Now that both of you mention that it really depends on the pedaling, it does give me something to think about.

AFAIK, one stereo sample = two notes of polyphony. That means 64 would be the actual limit.

Can anyone name a highly advanced composition where this would be a problem? I need to decide between two instruments, one with 128 and the other with 196 note polyphony. I'm an advanced player.

I really thought 128 would be enough, until this thread.
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/05/16 04:25 AM

Originally Posted by Stephano
I wouldn't be using layered sounds, MIDI or anything like that. It'd be piano playing only.

Now that both of you mention that it really depends on the pedaling, it does give me something to think about.

AFAIK, one stereo sample = two notes of polyphony. That means 64 would be the actual limit.

Can anyone name a highly advanced composition where this would be a problem? I need to decide between two instruments, one with 128 and the other with 196 note polyphony. I'm an advanced player.

I really thought 128 would be enough, until this thread.
You mentioned "playing an orchestra through the piano". That's layering sound, though there are ways of doing it without affecting polyphony limits.

Stereo has nothing to do with polyphony; it's a recording/reproduction technique, not a characteristic of a musical instrument (a piano is neither mono nor stereo nor 7.1 - you can record it and play it back in all three and more - the number of notes you can hear at any one moment is very largely independent of how you record them).

Easy solution to your dilemma: play both instruments in fast pieces that require a fair amount of pedalling, and see if you can hear notes being lost. Some Chopin or Rachmaninoff Études would probably work well, though I don't know them well enough to suggest which ones. Gershwin piano concerto (3rd mov. especially) and Rhapsody in Blue too, I suspect.

If you can't test play either instrument... I would suggest you don't buy it independent of any polyphony limits.
Posted By: Charles Cohen

Re: 128 note polyphony question - 06/05/16 05:37 AM

I once tried Pianoteq, pressing my forearms down on the keyboard and holding the pedal.

. . . I couldn't even reach polyphony=100 by doing that.

My guess is that polyphony=128 is enough to play anything that you'd ever want to play, or that you _could_ play, with a "straight piano" setup. But I can't prove it.


Posted By: 8 Octaves

Re: 128 note polyphony question - 06/05/16 06:29 AM

Listen to the famous pianist Andras Schiff on Beethoven - http://download.guardian.co.uk/sys-audio/Arts/Culture/2006/11/22/03_14CSharpMin.mp3

He talks about playing sonata no. 14 in C# minor first movement with the sustain pedal held one-third down throughout the entire first movement. You may disagree with Andras Schiff, but if you take him at his word, then the sustain dampers are off the strings 6-7 minutes long through the whole of the 1st movement more than any DP's polyphony capabilities.

I think the important question is how many notes could the human ear (brain) really continue to hear as they fade with much louder new notes being made? I personally don't know, but at some number we simply lose track I'm guessing. If you cannot hear it anymore does it matter?
Posted By: Macy

Re: 128 note polyphony question - 06/05/16 07:24 AM

Measured by apples and oranges. Polyphony is defined differently by various VST's and DP's depending on how they count stereo samples, release samples, sustain samples (on some VSTs), etc.
Posted By: Bellicapelli

Re: 128 note polyphony question - 06/05/16 09:52 AM

Originally Posted by Charles Cohen
I once tried Pianoteq, pressing my forearms down on the keyboard and holding the pedal.

. . . I couldn't even reach polyphony=100 by doing that.

My guess is that polyphony=128 is enough to play anything that you'd ever want to play, or that you _could_ play, with a "straight piano" setup. But I can't prove it.




Well, i guess this math is easily done, since a piano only has 88 keys to be played at once, so you could have had more forearms handy, to play even more keys, whitout actually reaching the limit. Though that is, without pedal.

Problem rather arises playing complex chord pieces with pedal. Think Rach, or anything comparable to 23 5.
Posted By: JoeT

Re: 128 note polyphony question - 06/05/16 10:02 AM

Note stealing has been perfected long before polyphony increased to the hundreds. You can play pure piano with just 32 notes polyphony without audibly losing notes. In fact the software piano Ivory II starts you up with a default setting of 24 notes polyphony.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/05/16 11:45 AM

I have tried to play the Chopin's Fantaisie Impromptue with a virtual piano which displays the number of used voices. The number of voices hardly get over 100 in the heaviest measures.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/05/16 11:54 AM

Originally Posted by JoeT
Note stealing has been perfected long before polyphony increased to the hundreds. You can play pure piano with just 32 notes polyphony without audibly losing notes. In fact the software piano Ivory II starts you up with a default setting of 24 notes polyphony.


My GranTouch had 32 note polyphony and I never noticed notes dropping out.

I suppose there will always be someone who thinks 512 note polyphony is better than 256 note polyphony because the number is bigger, and bigger is always better.

If you can't hear the difference, is there a difference? smile
Posted By: EssBrace

Re: 128 note polyphony question - 06/05/16 12:08 PM

Originally Posted by Stephano
...I need to decide between two instruments, one with 128 and the other with 196 note polyphony. I'm an advanced player.

I really thought 128 would be enough, until this thread.


Almost everything else, when comparing two digital pianos, is more important than polyphony in my opinion. Enough is enough and once you have 'enough' it ceases to be an issue. For stand alone piano playing, without layering etc, I think 64 note polyphony should suffice for advanced repertoire. I suspect 32 would probably be enough too in the vast majority of circumstances. Polyphony has just become a meaningless number whereby manufacturers can differentiate (and justify higher prices) between their own models or make boastful claims about their products versus other makers.

Choose on sound/touch/features/appearance/after-sales support and of course price. Not polyphony.

Cheers,

Steve
Posted By: propianist

Re: 128 note polyphony question - 06/05/16 12:27 PM

The more advanced software pianos and digital pianos require extra polyphony to generate release samples, per note, sustain pedal resonance layers, and most importantly sympathetic string resonance which may add dozens of "extra" sound layers if individual "harmonic tone" samples are used whenever needed.

On some instruments, stereo samples require 2 notes of polyphony, whereas mono sampled versions are more economical with limited polyphony.

If you watch the polyphony count in Kontakt, you can see up to 3 poly voices for a single played note. It shows what's going on behind the scenes with the piano's layering to create a realistic sound.

Repeated strikes of the same key building up, playing with sustain pedal, glissando etc, all consume large polyphony. I've regularly seen over 200 poly on the counter when I'm playing software pianos in Kontakt, eg. Vintage D.

Posted By: MacMacMac

Re: 128 note polyphony question - 06/05/16 01:23 PM

It's refreshing to see a dose of sanity brought to this subject.
Originally Posted by Dave Horne
I suppose there will always be someone who thinks 512 note polyphony is better than 256 note polyphony because the number is bigger, and bigger is always better. If you can't hear the difference, is there a difference? smile
Originally Posted by EssBrace
Almost everything else, when comparing two digital pianos, is more important than polyphony in my opinion. Enough is enough and once you have 'enough' it ceases to be an issue. For stand alone piano playing, without layering etc, I think 64 note polyphony should suffice for advanced repertoire. I suspect 32 would probably be enough too in the vast majority of circumstances. Polyphony has just become a meaningless number whereby manufacturers can differentiate (and justify higher prices) between their own models or make boastful claims about their products versus other makers.

Choose on sound/touch/features/appearance/after-sales support and of course price. Not polyphony.
Posted By: JoeT

Re: 128 note polyphony question - 06/05/16 04:05 PM

Originally Posted by propianist
The more advanced software pianos and digital pianos require extra polyphony to generate release samples, per note, sustain pedal resonance layers, and most importantly sympathetic string resonance which may add dozens of "extra" sound layers if individual "harmonic tone" samples are used whenever needed.

But you won't notice those missing once you reach polyphony limits and those subtle special effects diminish in favor of new notes. This is the same how MP3 compression works by removing the inaudible parts of the sound.

Quote
On some instruments, stereo samples require 2 notes of polyphony,

Technical details are irrelevant. There is only one interpretation for "32 note polyphony". If using stereo samples halves it, then it's simply "16 note polyphony" for the matter of this discussion. A note on the staff (and sent through MIDI) played with a stereo sample is still a note, not two notes. Instruments being advertised wrong doesn't change this.
Posted By: propianist

Re: 128 note polyphony question - 06/05/16 04:24 PM

Thanks MacMacMac, but looks like you just spoke too soon...!

Originally Posted by JoeT
Technical details are irrelevant.


In that case Joe, I won't explain to you why you're wrong.
Posted By: JoBert

Re: 128 note polyphony question - 06/05/16 04:51 PM

I'm with Joe here. In it's original meaning, the polyphony count means the number of notes that can be played at the same time. Many DP manufacturers have changed this meaning in their marketing material (to be able to boast a higher polyphony count). The meaning they use, is that the polyphony count means the number of available sound generators.

But if you need two sound generators to play a single stereo note, then if the marketing material boast a 256 polyphony (because there are 256 sound generators), then, with the original meaning, the instrument actually only has a polyphony of 128 while playing a stereo voice. The full polyphony of 256 can only be reached when playing a mono voice.

Wikipedia explains it like this:

A "sound" (also called a "timbre" or "patch") may be generated by more than one oscillator or sound-source to allow more complicated sounds to be produced. A synthesizer with 16 oscillators may be capable of 16-note polyphony only when simple, single-oscillator sounds are produced. If a particular patch requires four oscillators, then the synthesizer is only capable of four-note polyphony.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 05:03 PM

Ah, a variety of opinions on this matter. I didn't expect it to be that way.

To clarify, I'm not interested in any orchestra playing on the piano, or any kind of layering sounds whatsoever. My question about 128 note polyphony only pertains to playing complex classical music using the piano's highest quality stereo sample.

Yes, I use lots of pedal sometimes. It depends on the piece, but generally with Beethoven and Chopin, I need to use it frequently.

The instrument in question is the Yamaha YDP-162 with 128-note polyphony.

I inquired on this forum at the end of 2014 about buying a digital piano. Well, some unforeseen have caused me to put that aside at the time, so now I'm back and researching as much as possible before I buy an instrument.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/05/16 05:14 PM

To clarify, I'm not interested in any orchestra playing on the piano, or any kind of layering sounds whatsoever. My question about 128 note polyphony only pertains to playing complex classical music using the piano's highest quality stereo sample.
...

The instrument in question is the Yamaha YDP-162 with 128-note polyphony.




If you like that piano, buy it, 128 note polyphony is more than adequate.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 05:31 PM

Originally Posted by Dave Horne
If you like that piano, buy it, 128 note polyphony is more than adequate.

Thanks, Dave.

Okay, so let's say that some complex, fast passages with lots of pedalling somehow lead to the maximum of 128 notes being played. In that case, would the 129th note be nonplayable (silent), or would it only cause the oldest note (1st note) to disappear? If it's the latter, then that's acceptable I guess, since those oldest notes would have almost faded out already. If it's the former, that'd be really bad, pressing keys with no sound coming out.

The reason I ask this question is because of the users who wrote that it all depends on the pedalling.
Posted By: JoeT

Re: 128 note polyphony question - 06/05/16 05:40 PM

Originally Posted by Stephano
Okay, so let's say that some complex, fast passages with lots of pedalling somehow lead to the maximum of 128 notes being played. In that case, would the 129th note be nonplayable (silent), or would it only cause the oldest note (1st note) to disappear? If it's the latter, then that's acceptable I guess, since those oldest notes would have almost faded out already. If it's the former, that'd be really bad, pressing keys with no sound coming out.

It's neither the oldest nor the newest note. It's the one you can't hear anymore below the entire soundscape.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/05/16 05:56 PM

The note from the bass last longer than these from the trebble, then in case of a polyphony limit, it is not wise to shut off the earlier one. The quiter should be preferable.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/05/16 06:01 PM

It would have been better to get the group's opinion on which of the two keyboards your looking at rather than discussing only one spec of both instruments - polyphony. The sample set, action, amplifiers and speakers, build quality are just if not more important. But since we are buried in a polyphony debate instead I'll say 128 poly for solo piano is good for most rep you may throw at it... I suspect if the material you are playing is so advanced as to push the limits of current digital pianos you'd be looking at models other than the YDP anyway.

As far as note stealing... It's not desirable, IMHO. YMMV. But Poorly handled cut-off sounds terrible. There is a sound you expect a solo acoustic piano to make when pedaling and doing effects like gliss, tremelo, rapid passages with pedal down. Instruments with low poly let go of notes using a technique to make cutoff as unnoticeable as possible. Fading notes under a certain volume to recover the voice so it can be used by newly struck notes. Complex blending of sustained notes, fundamentals, overtones, and resonance that happens between them is an area digital pianos are only beginning to do well. In addition to better polyphony specs they are trying to use modeled algorithms to account for sounds digital pianos didn't make before. While none of this matters much for pop music in a band setting running through a PA system. Solo piano is a different animal and I don't think it's wise to suggest to digital piano makers that their work was done 15 years ago when they figured out how not to make a DP with 32 note polyphony sound like carp. So, yeah... I'm thrilled DPs can handle 128 or 256 note poly, note on and release sounds, string resonance and beyond. Today's digital pianos play, sound look and feel much better than the previous decades. There's nothing wrong with that.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 06:03 PM

Originally Posted by Frédéric L
The note from the bass last longer than these from the trebble, then in case of a polyphony limit, it is not wise to shut off the earlier one. The quiter should be preferable.

So the piano's technology would choose the most silent note to drop?

I wonder if this was actually tested somewhere.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 06:09 PM

Originally Posted by ElmerJFudd
I suspect if the material you are playing is so advanced as to push the limits of current digital pianos you'd be looking at models other than the YDP anyway.

That's a really good point, and an interesting view to look at it. Thanks.

Quote
I'll say 128 poly for solo piano is good for most rep you may throw at it

It's only the underlined that bothers me a bit. Perhaps I'm just obsessing. laugh

Quote
It would have been better to get the group's opinion on which of the two keyboards your looking at rather than discussing only one spec of both instruments - polyphony. The sample set, action, amplifiers and speakers, build quality are just if not more important.

I opened this separate thread for that discussion.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/05/16 06:10 PM

You're obsessing. Indeed.

And I'll rephrase... without a doubt 128 poly + the silent note stealing techniques used in the YDP you are looking at will handle any piano rep a beginning player all the way through many years of successful study may throw at it. At which point you will have outgrown the YDP and be looking at a more professional instrument for many reasons other than or perhaps in addition to polyphony.
Posted By: JoBert

Re: 128 note polyphony question - 06/05/16 06:24 PM

Originally Posted by Stephano
Quote
I'll say 128 poly for solo piano is good for most rep you may throw at it

It's only the underlined that bothers me a bit. Perhaps I'm just obsessing. laugh


I think you can safely say that the polyphony will be enough. Even if the actual polyphony for stereo samples is half of that, i.e. only 64, then that still means that you can play all 52 white keys plus 12 of the black keys at the same time (leaving only 24 black keys unplayed) and there still aren't any note drops. Only if you then play another black key will one of the other notes be dropped. To be honest, if you can hear that dropped note over all the other 64 notes, then you won't be happy with a sampled digital anyway, as sampled digitals have other sound shortcomings that are much easier to hear (looping, short decay, etc.). You should then get either an acoustic or a fully modeled digital.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/05/16 06:25 PM

Originally Posted by Stephano

So the piano's technology would choose the most silent note to drop?

I wonder if this was actually tested somewhere.

I do remember a software virtual instrument from 2000. And there was something like this : a loop process the decrease of the sample volume with the time, and this volume was compared to the others to decide which note is the quieter.

It was not that complicated, and could be integrated easily in every piano. It is perfectible (not each sample has the same loudness at the same volume : the sensitivity of the ear depends of the frequency), but with a 128 polyphony piano, I suppose the stolen note will be quiet enough.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 07:31 PM

Originally Posted by ElmerJFudd
And I'll rephrase... without a doubt 128 poly + the silent note stealing techniques used in the YDP you are looking at will handle any piano rep a beginning player all the way through many years of successful study may throw at it. At which point you will have outgrown the YDP and be looking at a more professional instrument for many reasons other than or perhaps in addition to polyphony.

Not a beginner though. I have played the Pathetique, the Appassionata, several Schubert Impromptus, Chopin pieces such as Fantasie Impromptu, and looking forward to learning some of his Scherzos, among other things. Have played quite a few Preludes and Fugues from Bach's WTC as well. I've played piano a long time ago and have decided to come back to it, and also post recordings on YouTube eventually.

Right now I can only practice on my two 61-key KORG Triton Studio and KORG PA800, which is nothing for classical music.

Beethoven's Emperor concerto (piano part) is also something I'd be practicing on the YDP-162. Again, the orchestra coming from the piano is not needed.

Do you still think 128 polyphony is fully sufficient? And I really apologize if I'm bothering you with my questions.
Posted By: JoeT

Re: 128 note polyphony question - 06/05/16 07:44 PM

Solo piano: 32 notes is fine even for advanced repertoire, 64 is fully sufficient. Anything above is just marketing.

You have ten fingers, Stephano.

Digital pianos, even current ones, have various issues, when you try practicing advanced pieces, like when repeating notes without pedal using upright technique. But polyphony is solved once and for all since more than a decade.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/05/16 07:58 PM

Originally Posted by JoeT
Solo piano: 32 notes is fine even for advanced repertoire, 64 is fully sufficient. Anything above is just marketing.

You have ten fingers, Stephano.

Digital pianos, even current ones, have various issues, when you try practicing advanced pieces, like when repeating notes without pedal using upright technique. But polyphony is solved once and for all since more than a decade.


I agree.

I think if more people said this Stephano might start believing it.

In the future when there's another thread about polyphony can we refer everyone to this thread and be done with it?

Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 07:59 PM

Originally Posted by Dave Horne
I think if more people said this Stephano might start believing it.

I have to say, you're right. laugh
Posted By: JoBert

Re: 128 note polyphony question - 06/05/16 08:01 PM

To be honest, with that repertoire, it's probably not the polyphony that will be the first thing you will feel limited by. It will be the other limitations of the two entry level Pianos (the Kawai and the Yamaha) that you are considering. Meaning the keyboard (especially the fine control) and the sound generator (stuff like harmonic vibrations, short sustain, etc.).

It's possible to play such pieces on these entry level DPs, but you'll soon feel their limitations. And then you'll want to upgrade. smile

If it's a hard budget restriction that you can't get around, then no problem, go for it (an entry level DP is still better than no piano at all, if you want to get back into playing).

But if it's rather a consideration of "well, I could afford something more expensive, [but I'm not sure that I will stick with it/that I'm still good at it/I can't convince my wife/I can't justify the expense after not playing for so long/whatever other reason], so I'll get something entry level to begin with", then my suggestion would be: Don't. I made that mistake myself last autumn, so now I have a good DP and an entry level DP I'm not really using anymore, an expense I could have saved myself. Get something proper right away, it will save you money and at the same time increase your enjoyment from the get go. My recommendation for the next step up would be the Kawai CA-17. A very good piano for its price (but 1000€ more than your current choices).
Posted By: Frédéric L

Re: 128 note polyphony question - 06/05/16 08:15 PM

Originally Posted by JoeT
Solo piano: 32 notes is fine even for advanced repertoire, 64 is fully sufficient. Anything above is just marketing.

You have ten fingers, Stephano.

With the pedal, you can trigger far more than 10 notes simultaneously.

Then with the Fantaisie Impromptu, 100 voices were used on some heaviest measures. However with a quite good note stealing algorithm, we need less than 100 voices. (The 100th is at the limit to be cut because of being inaudible).

I never heard drop outs with my 24 voices Ivory VST (default setting), but I don't play advanced repertoire.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 08:18 PM

Originally Posted by JoBert
To be honest, with that repertoire, it's probably not the polyphony that will be the first thing you will feel limited by. It will be the other limitations of the two entry level Pianos (the Kawai and the Yamaha) that you are considering. Meaning the keyboard (especially the fine control) and the sound generator (stuff like harmonic vibrations, short sustain, etc.).

It's possible to play such pieces on these entry level DPs, but you'll soon feel their limitations. And then you'll want to upgrade. smile

I know what you are saying, and it makes sense, but I'm able to adapt very well. I've practiced on a very poor "AE" (Action Effect) key action on an old Yamaha Clavinova a very long time ago, for years, and was still able to adapt to an upright when performing, and even to a concert grand. Basically, I practiced once a week one a Steinway Grand, and the rest was on my Clavinova with "AE" keys, and I didn't have much trouble. Later I was able to practice on an acoustic for a longer period.

I don't intend to be a concert pianist. That's a thing of the past. Right now I want to play for myself, and to make some recordings. A small recital from time to time would be ok, but nothing more than that.

And thank you for your input.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 08:26 PM

Originally Posted by JoBert
it's probably not the polyphony that will be the first thing you will feel limited by

This is somewhat worrysome, because it appears that you think that 128 note polyphony would indeed be a limitation.

It's interesting that there isn't some kind of test that has been done at some point, with results posted online, about how 128 note polyphony handles fast, chord-rich music with lots of sustain. That would really be ideal.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/05/16 08:34 PM

Originally Posted by Stephano
It's interesting that there isn't some kind of test that has been done at some point, with results posted online, about how 128 note polyphony handles fast, chord-rich music with lots of sustain. That would really be ideal.

I have used the number of used voices of the Fantaisie Impromptu which is a quite fast piece of music with sustain (but no chords), but if you have a better idea, you're welcome.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 08:36 PM

Originally Posted by Frédéric L
Originally Posted by Stephano
It's interesting that there isn't some kind of test that has been done at some point, with results posted online, about how 128 note polyphony handles fast, chord-rich music with lots of sustain. That would really be ideal.

I have used the number of used voices of the Fantaisie Impromptu which is a quite fast piece of music with sustain (but no chords), but if you have a better idea, you're welcome.

No, that sounds really good. I was just thinking that maybe not all 128-note polyphony is the same. Many speculate that one stereo piano sample takes up two notes, or something.

Ah, I have no idea anymore. It's starting to irritate me.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/05/16 08:40 PM

Originally Posted by JoeT
Solo piano: 32 notes is fine even for advanced repertoire, 64 is fully sufficient. Anything above is just marketing.

You have ten fingers, Stephano.

Digital pianos, even current ones, have various issues, when you try practicing advanced pieces, like when repeating notes without pedal using upright technique. But polyphony is solved once and for all since more than a decade.



I disagree, emphatically. I guess with David on this topic as well. Oh well. If this were true, 10 note poly would be sufficient. And it isn't. 128 is in fact better than 10 at recreating the buildup of sound when pedaling on a DP. Note stealing is a solution to the limitations of the silicon. Acoustic pianos don't do this - they do more than this when the pedal is down and the harp and case are vibrating.

But regardless of how entry level digital pianos sound playing advanced repertoire, the YDP will handle the issue the way it was designed to... whenever you strike a key a pitch will sound (with the mentioned caveat of very fast repeated notes - unrelated to polyphony).

Some people feel that all the digital piano manufactures' action designs are just marketing and they all play the same from one to the next. I entirely disagree with that notion as well. There is a very short list of current DPs that come close to playing like the mechanism in a fine grand piano. Definitely an area they can still improve on.

But again, all unimportant when your budget is entry level. Get the YDP and enjoy playing a full size keyboard again. You'll know wether you need something better or not as time goes by. And if so you'll do something about it. The 128 poly is not going to be the reason you move up from the YDP.
Posted By: Stephano

Re: 128 note polyphony question - 06/05/16 08:47 PM

Thank you for your advice, ElmerJFudd. I appreciate it.
Posted By: JoeT

Re: 128 note polyphony question - 06/05/16 08:50 PM

Originally Posted by ElmerJFudd
I disagree, emphatically. I guess with David on this topic as well. Oh well. If this were true, 10 note poly would be sufficient.

Back in the Yamaha OPL3 FM synth days one had as few as 9 voices. That was limiting.
Posted By: petes1

Re: 128 note polyphony question - 06/05/16 08:55 PM

Originally Posted by Stephano
Originally Posted by JoBert
it's probably not the polyphony that will be the first thing you will feel limited by

This is somewhat worrysome, because it appears that you think that 128 note polyphony would indeed be a limitation.


An emphatic NO is deserved here. As I've seen folks posting and as I agree, 128 is more than sufficient.

Quote
It's interesting that there isn't some kind of test that has been done at some point, with results posted online, about how 128 note polyphony handles fast, chord-rich music with lots of sustain. That would really be ideal.


There's no need. Use your ears, and instead concentrate on more pertinent issues.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/05/16 09:23 PM

There's no need. Use your ears, and instead concentrate on more pertinent issues.

I agree.

There seems to be a lot of folks who look to numbers instead of using their ears.

I owned a GranTouch (as I've written here many times) before I bought my N3. The GranTouch had 32 note polyphony and I never noticed notes dropping out.

I'm sure 64 note polyphony will suffice and even if there is a limitation, we won't hear it. Now we're discussing 128 note polyphony.

Are there software pianos where the user can dial in the polyphonic limitation? We could then 'play' a midi file (or whatever) of a very demanding piece and compare the outcomes.

I really can't think of anything more to add to this discussion.

smile
Posted By: Frédéric L

Re: 128 note polyphony question - 06/05/16 09:30 PM

Yes, Garritan CFX and Ivory have a polyphony limitation which can be selected. But I don't know if it affects "offline" (non realtime) rendering. If the limitation affect this rendering it will have an impact on the rendering time, it will be easy to notice it.

EDIT: Both displays the number of used voices. Ivory restrict the number of voices even in offline rendering. According to the displayed number of "layers", CFX doesn't seem to limit the polyphony.
Posted By: emenelton

Re: 128 note polyphony question - 06/05/16 09:34 PM

Originally Posted by Dave Horne
[i].

I really can't think of anything more to add to this discussion.

smile


Except get a Kawai ES8 and you won't have to worry about buying the wrong 1/8" adapter cable either!
Posted By: Frédéric L

Re: 128 note polyphony question - 06/05/16 09:55 PM

I have rendered the Fantaisie Impromptu with Ivory :

24 voices : http://www.sinerj.org/~loyer/piano/fantaisie_24.mp3
32 voices : http://www.sinerj.org/~loyer/piano/fantaisie_32.mp3
64 voices : http://www.sinerj.org/~loyer/piano/fantaisie_64.mp3
128 voices : http://www.sinerj.org/~loyer/piano/fantaisie_128.mp3

(And just to test the effectiveness of the limit : 4 voices : http://www.sinerj.org/~loyer/piano/fantaisie_4.mp3)
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/05/16 10:05 PM

Originally Posted by Frédéric L
Hah! That's what I call "evidence based posting". And confirmation that posters on PianoWorld really are special.
Posted By: MacMacMac

Re: 128 note polyphony question - 06/05/16 11:24 PM

The car's gas tank holds 18 gallons. But I only have 10 gallons right now.
The tire pressure should be 32 PSI, but it's at only 30 right now.
The windows are not tinted. I wish they were.
The battery is dead.
Which should I worry about?

Piano polyphony is the tinted glass. Forget about it.

Look at something more important.
Like sound. Like touch.
Like anything other than polyphony.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/05/16 11:30 PM

Originally Posted by Frédéric L


Now compare to an acoustic grand piano with infinite ability to resonate and blend with liberal use of the pedal.


So, Ivory one way or another does what it has to do to allow a note attack or trigger to occur, most obviously it opts to chop off ends of notes and/or skip sounding the release samples and notes the player is no longer concentrating on as expected. But in listening back, can we agree that the 4 note sounds poo, the 24 does better and it improves from there eventually reaching a point of diminishing returns? Maybe we can't, I guess it depends on what ones expectations are.

Regardless, back to OP - with a solo sampled piano 128 notes is surely plenty... off topic, only layering with additional timbres would require yet more.

...and yet for other reasons a modern product as amazing as Ivory and the like are, it still falls short of the real deal. But it's so exciting to see advancements in digital piano technology and the developers' efforts to bring us closer and closer to the sound of the instrument we have grown to love with the benefits of affordability, space constraints, the ability to practice silently, to have the mechanism of the action translate to when we do have opportunity to play the real thing, to record ourselves and immediately play back, have backing tracks, etc. etc. etc. So please, advance, do better. Improve. Next I hope to see the amplifier and speaker systems improve and trickle down to instruments for all.
Posted By: Kawai James

Re: 128 note polyphony question - 06/06/16 12:28 AM

Originally Posted by Frédéric L
(And just to test the effectiveness of the limit : 4 voices : http://www.sinerj.org/~loyer/piano/fantaisie_4.mp3)


Nice! wink

Reminds me of listing to 4 channel MODs on an Amiga. wink

Cheers,
James
x
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/06/16 12:54 AM

OT: Haha, terribly geeky James and right up my alley. I miss those days... I had the 64, 128, Amiga 500, and a 1200. Loaded with games and all sorts of music software. Even had it synced to smpte stripe on my 8 track reel-to-reel. Now that was making the most of whatever tech i could get my hands on.
Posted By: Kawai James

Re: 128 note polyphony question - 06/06/16 02:08 AM

Nice setup Elmer! wink

Yep, I too miss those 8-bit Amiga tracker days, although towards the end of my Amiga 'life' I was using OctaMED SoundStudio with MIDI and 14-bit virtual track mixing. wink

</geek>

Cheers,
James
x
Posted By: Frédéric L

Re: 128 note polyphony question - 06/06/16 04:48 AM

I was sure the 4 voices setting would please some people. wink
Posted By: JoBert

Re: 128 note polyphony question - 06/06/16 11:16 AM

Originally Posted by Stephano
Originally Posted by JoBert
it's probably not the polyphony that will be the first thing you will feel limited by

This is somewhat worrysome, because it appears that you think that 128 note polyphony would indeed be a limitation.

That was not my intention. I too think that you will be fine with 128 polyphony. Check my earlier posts here in the thread, I argued that even if 128 boils down to "actual" 64, because of stereo, you likely wouldn't be able to hear it as a limitation.

(Although from my own Kawai/Yamaha experience, I would probably prefer the Kawai anyway, if I tried the two DPs you are considering. Then the 128 polyphony question would be moot anyway...)
Posted By: propianist

Re: 128 note polyphony question - 06/06/16 05:22 PM

Originally Posted by Stephano
Is there any known piece in classical music which would require more than 128 note polyphony? Is there any piano sonata where 128 polyphony of a digital piano would not be sufficient?


[Linked Image]

Consider the last few bars of Beethoven, 3rd movement of Sonata in C# minor, Opus 27, No 2.
I expect loads of pianists probably know this famous piece. I played it at school concert when I was a kid.
Anyway, if we depress sustain pedal down at beginning of this excerpt, and hold sustain pedal down through the whole arppeggio section up to the bass C# octaves at start of penultimate bar, and release it where the crotchet rest comes, I would say that section strikes 103 notes in total, all of which sustain simultaneously until the pedal release cuts them off, whereupon you may even trigger 103 release samples too!!!!

But - you can count them yourself - there's undoubtedly 103 main note samples to be sounded.

This means that 64 note poly is definitely insufficient.

At dynamic strike of forte = loud, the three bars duration of a piece marked "Presto agitato" isn't nearly long enough for any of these notes to have faded out to silence already, so they all must keep sounding. All 103 of them.

PLUS...
Any serious modern digital piano would add sympathetic string resonance overtone layers - for every strike of every note - and here, there is clearly going to be a lot of harmonic inter-relationship and reinforcement between all the C#'s, G#'s and E's over the 4 octave range of this arpeggio and extended by further harmonics beyond this to the extreme of the piano range. You could reasonably expect at least 3 or 4 layers of harmonic overtones to be triggered by a modern digital piano sympathetic string resonance algorithm (F, 2F, 3F, 4F etc.) so there could be 3 or 4 times as many samples sounded, in addition to the global sustain pedal resonance sample too.

I'd say this would easily exceed 128 voice polyphony of a modern day digital piano's sound generator with sympathetic string resonance overtones, release samples, etc, if fully implemented.

If stereo samples do require two voices of polyphony technically then you can obviously double your estimate.
Posted By: JoeT

Re: 128 note polyphony question - 06/06/16 05:47 PM

Originally Posted by propianist
so they all must keep sounding. All 103 of them

You won't be able to hear all 103 of them. This is why MP3 compression works so well. Same goes for subtle side effects.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/06/16 06:18 PM

I wouldn't depress the sustain pedal down for three bars in that example.

I'm surprised that this thread is still going. smile
Posted By: Stephano

Re: 128 note polyphony question - 06/06/16 06:24 PM

I was just about to order the YDP-162 with 128 note polyphony... and then I read propianist's post.

Blah.
Posted By: propianist

Re: 128 note polyphony question - 06/06/16 06:24 PM

Originally Posted by JoeT
You won't be able to hear all 103 of them. This is why MP3 compression works so well. Same goes for subtle side effects.


MP3 compression does not "work so well" - it is a lossy compression which degrades sound quality and also introduces compression artifacts into the audio that were not present in the original.
It's easy to hear the difference between a 320kbps MP3 and the same source 1440kbps linear WAV of "CD quality" 16 bit 44.1KHz.

Of course you can hear all the notes ringing in this 5 second Beethoven excerpt. They're all struck forte - what else do you expect?!
If the notes were to disappear the total energy of the excerpt would be much less. The level in dB which the passage crescendos up to, by addition of sounding notes, would be a lot lower and you could notice the difference. Play it on an acoustic piano with sustain pedal, then play it staccato without pedal and hear the difference in total dB volume!
Adding all the notes together makes a big dramatic effect! The WHOLE summed effect is what you hear, not necessarily each individual note sustain tail in stark clarity and definition, but the sum total waveform.

Psychoacoustic masking of sounds may prevent you from clearly hearing whether a note was ringing or silenced, but you can still hear it even if it's not as blatantly crystal clear as when in isolation. Your ears don't stop hearing details just because other sounds are present. You can focus your attention. You can listen to one voice in a crowded room.

eg. Consider an orchestra which normally has 20 violins.

Psychoacoustic masking concepts might make you think you couldn't hear one individual violin in the mix, but if one violin was slightly flat or sharp you would hear it, and if one player abruptly stopped (like a polyphony note cut-off), you theoretically could hear that subtle dB level change also.

Adding together 20 equal volume violins raises volume +12.97dB louder than one solo violin.
If one player drops out (now 19 violins) the volume becomes only +12.75dB which is a -0.22dB change. Not massive, but it can be noticeable. EQ controls are often calibrated in +/-0.1dB steps. I know I can hear changes of +/-0.1dB in my digital mixer and EQ, and I believe a level change of -0.22dB is perceivable by a discriminating listener too. Especially if it's a sudden dropout - like a note cutoff due to polyphony limit being reached suddenly, rather than a soft gradual fadeout.

If you limited polyphony to 64, vs original = 256 voice poly, you could actually measure the difference in dB between the two excerpts in a DAW audio editor. That's a fact.
Whether you can hear slight dB differences or not, or whether you even care, depends on your experience as a skilled listener and the quality of your monitoring setup / loudspeakers / headphones, etc.

Obviously if you spend your time listening to low bit rate MP3s, then don't expect to hear all the subtleties that exist in audio!

PS. Before anyone else says it - Beethoven did NOT hear all 103 notes! Yes, thankyou. hahaha
Posted By: propianist

Re: 128 note polyphony question - 06/06/16 06:29 PM

Originally Posted by Dave Horne
I wouldn't depress the sustain pedal down for three bars in that example.


You would if you were using it to try and prove the point about polyphony in digital pianos...!

Originally Posted by Dave Horne
I'm surprised that this thread is still going. smile


Yeah, it's like arguing with people who say, "But I've only got ten fingers..." or "there's only 88 keys on a piano..." and don't know the difference between musical notes and digital piano sound generator's voices of polyphony.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/06/16 06:32 PM

Interesting piece of music. Thanks!

I have render it with both settings (still Ivory ACD):

64 voices : http://www.sinerj.org/~loyer/piano/beethoven_64.mp3
128 voices : http://www.sinerj.org/~loyer/piano/beethoven_128.mp3

Ivory doesn't seem to count extra voices for the sympathetic resonance. The voice counter go up to 111 only. But it has a distinct "sustain resonance" feature (a global filter ?) like some digital piano : CLP545, CA97 to name a few.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/06/16 06:46 PM

Originally Posted by Stephano
I was just about to order the YDP-162 with 128 note polyphony... and then I read propianist's post.

Blah.

But the YDP162 has no string resonance.
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/06/16 06:51 PM

Originally Posted by JoeT
This is why MP3 compression works so well.
Are you seriously suggesting this?
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/06/16 07:00 PM

Originally Posted by propianist
PS. Before anyone else says it - Beethoven did NOT hear all 103 notes!
He probably did - and I mean physically, not just in his mind's ear. His hearing troubles started around 1798 and Op. 27 No. 2 was published in 1802.

Stephano - unless you live in the back of beyond, go out and play a couple of modern digital pianos. See if you can hear notes dropping out of a few fast passages with sustain. No point in agonising "theoretically" about something when empirical proof is very easy to obtain.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/06/16 07:40 PM

The resonance algorithms they are using today are modeled FX and would not come into play with regard to voice polyphony. However, it does use additional DSP or system resources. It adds a, great deal of realism and acoustic like behaviour to a sampled piano and wasn't available on DPs in decades past. There is nothing wrong with keeping your older digital piano if the specs like polyphony and fx like reverb/chorus satisfy you. But there is also no reason for the digital piano industry to halt advancement because of it. Today's software pianos are cutting edge and driving the DP manufacturers to do better. I think we could all hear the difference between a rendering of Ivory and the like when compared to the Korg SG1 or Kurzweil Micro Piano which at one time were considered so "real".
Posted By: propianist

Re: 128 note polyphony question - 06/06/16 07:41 PM

Hi Frédéric,

Your "64 poly" MP3 file measures louder than the "128 poly" MP3 file!!!!
That cannot be right, empirically.
Are the settings truly identical for both?
Something must be inconsistent in your rendering process.
If you make two offline (non-realtime) renders of the same MIDI file on exactly the same settings as a null test, do they match together as perfectly identical?
Are there round robin samples in Ivory II American Concert D which make the results different each time?
Is sympathetic string resonance turned on?
Also, if the audio clip is only 7 seconds long, why not upload as linear WAV rather than MP3. Maybe MP3 compression is degrading results slightly.

Ultimately, although this debate is barely of interest to most people (including me - since I readily understood many, many, many years ago that more polyphony was always a good thing!) if a repeatable scientific test can be demonstrated that shows a measurable dB difference between two otherwise identical rendered piano files of different polyphony limits, then it finally proves there is a measurable / audible difference, and puts a quantifiable dB value on how large or small that is, which each reader can then ascertain for himself to what degree such a given dB difference might really affect them or not.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/06/16 07:50 PM

I have just made MP3 files again. (To be sure I have not done any error)

The Wav files:

64 voices : http://www.sinerj.org/~loyer/piano/beethoven_64.wav
128 voices : http://www.sinerj.org/~loyer/piano/beethoven_128.wav

Yes it is Ivory II American Concert D (Steinway). Sympathetic string is on (0dB), but there is also a sustain resonance (0dB clean Soundboard 1).
Posted By: MacMacMac

Re: 128 note polyphony question - 06/06/16 08:04 PM

Me too.
Originally Posted by Dave Horne
I'm surprised that this thread is still going. smile
It's a non-issue for me. And, for the beginners who are asking "which piano should I buy", the question is a non-starter. It just doesn't matter. It mattered hardly at all in the days of 32. It's meaningless in today's world of polyphony 128 and up.

propianist has a point ... in some music, high polyphony may be present. But contrast that with modern/pop music. Here, polyphony runs to the opposite extreme. Quite low.
Posted By: propianist

Re: 128 note polyphony question - 06/06/16 08:05 PM

Originally Posted by ElmerJFudd
The resonance algorithms they are using today are modeled FX and would not come into play with regard to voice polyphony.


From where do you come up with this blanket statement of fact???

Galaxy II in Kontakt 4 is clearly adding overtone sample notes, and the polyphony indicator onscreen shows this. Press and hold keys down silently and strike others and watch the behaviour.
You can also see the "overtone" samples loading, by name, when you load the preset at the beginning.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/06/16 08:13 PM

Originally Posted by propianist
Ultimately, although this debate is barely of interest to most people (including me - since I readily understood many, many, many years ago that more polyphony was always a good thing!) if a repeatable scientific test can be demonstrated that shows a measurable dB difference between two otherwise identical rendered piano files of different polyphony limits, then it finally proves there is a measurable / audible difference, and puts a quantifiable dB value on how large or small that is, which each reader can then ascertain for himself to what degree such a given dB difference might really affect them or not.

It is simple... I select with Audacity the whole file minus the last two chords, and select the Amplification effect to get the headroom before saturation. The level of the 64 voices version is -8.2dB, the level of the 128 is -7.7dB.

The 128 voices version is 0.5dB louder.
Posted By: Stephano

Re: 128 note polyphony question - 06/06/16 08:35 PM

Originally Posted by MacMacMac
Me too.
Originally Posted by Dave Horne
I'm surprised that this thread is still going. smile
It's a non-issue for me. And, for the beginners who are asking "which piano should I buy", the question is a non-starter. It just doesn't matter. It mattered hardly at all in the days of 32. It's meaningless in today's world of polyphony 128 and up.

propianist has a point ... in some music, high polyphony may be present. But contrast that with modern/pop music. Here, polyphony runs to the opposite extreme. Quite low.

The thing is, I am not a beginner, and modern/pop music on the piano is of very little interest to me. I've played fast, complex passages with lots of sustain, and intend to play and learn even more. The reason I'm worrying about 128 note polyphony is because of so many conflicting views, as you can see from this very thread.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/06/16 08:50 PM

Stephano, The reason I'm worrying about 128 note polyphony is because of so many conflicting views, as you can see from this very thread.

There's a difference between something being measurable and something being discernible in everyday life.

Your question has been answered and then some.
Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/06/16 09:00 PM

I haven't read the entire thread yet but I think it's pretty easy to surpass 128 polyphony in many pieces. For example I just looked at a piece I'm currently playing where there is a passage held by pedal from a trill that goes into doubled handed arpeggio runs up and down the keyboard. The arpeggios themselves are 63 notes per hand but it is doubled handed so that brings the notes to 126 already, however, the trill leading up to the run and the single handed notes after which one may decide to re-pedal (or not) depending on the pianist, would definitely cause the 128 polyphony to run out.

First 3 bars of page 4
http://conquest.imslp.info/files/im...___Zhavoronok___Balakirev__for_piano.pdf

However, as to wether or not you would notice the notes dropping, I'm not sure. I've heard that the digital pianos are now designed in such a way that it would not be very noticeable and very subtle.

That being said, the digital piano salesperson at my local store said that when Lang Lang was here for a masterclass, he played around on the digitals for a while and while he played certain arpeggios very fast the salesperson heard a small popping noise which he attributed to notes cutting out from the lack of polyphony. I wasn't there myself so I don't know how true this account is, but that is what I was told.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/06/16 09:09 PM

Originally Posted by Stephano
The thing is, I am not a beginner, and modern/pop music on the piano is of very little interest to me. I've played fast, complex passages with lots of sustain, and intend to play and learn even more. The reason I'm worrying about 128 note polyphony is because of so many conflicting views, as you can see from this very thread.


You were ready to buy an YDP162 which lacks string resonance. Such a piano will not need more voices than the number of notes in a set a measures with the sustain despressed. Here, with the end of the Sonata 14 Op27, you need no more than 103 voices. (111 according to the Ivory count).

And even if you need more for an even more demanding piece, you can hear on my renderings the impact of the polyphony limitation and make your opinion.

If you prefer a piano which deals with string resonance, you will have to get an higher end model and you will get more voices (Kawai CA17 : 192 voices, Yamaha CLP535 : 256 voices... well the Roland HP504 is the exception : 128 voices).
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/06/16 09:20 PM

Originally Posted by propianist
Originally Posted by ElmerJFudd
The resonance algorithms they are using today are modeled FX and would not come into play with regard to voice polyphony.


From where do you come up with this blanket statement of fact???

Galaxy II in Kontakt 4 is clearly adding overtone sample notes, and the polyphony indicator onscreen shows this. Press and hold keys down silently and strike others and watch the behaviour.
You can also see the "overtone" samples loading, by name, when you load the preset at the beginning.


Hi, fair enough, I don't believe all the instruments out there use the exact same methods to create these additional realisms. However, I do know a lot of research has been done on the modeling of pedal down and key down resonance fx and that many digital piano manufacturers are using DSP to achieve these results - Casio, Kurzweil, Nord, etc.

Here's a few papers that google coughed up if the math or method is of interest.

https://hal.archives-ouvertes.fr/fi...dellingOfSympatheticStringVibrations.pdf

http://www.eurasip.org/Proceedings/Eusipco/Eusipco2015/papers/1570104271.pdf

Here is an example of how modeling string and sympathetic resonance has been added to the new Kurzweil Forte with the v1.4 update.



I think we can agree that the effects are quite dramatic and an advancement over reverb, delay, chorus, etc. on digital pianos of the past.
Posted By: Stephano

Re: 128 note polyphony question - 06/06/16 09:21 PM

Originally Posted by Michiyo-Fir
That being said, the digital piano salesperson at my local store said that when Lang Lang was here for a masterclass, he played around on the digitals for a while and while he played certain arpeggios very fast the salesperson heard a small popping noise which he attributed to notes cutting out from the lack of polyphony. I wasn't there myself so I don't know how true this account is, but that is what I was told.

If true, that's really terrible.
Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/06/16 09:32 PM

Originally Posted by Stephano
Originally Posted by Michiyo-Fir
That being said, the digital piano salesperson at my local store said that when Lang Lang was here for a masterclass, he played around on the digitals for a while and while he played certain arpeggios very fast the salesperson heard a small popping noise which he attributed to notes cutting out from the lack of polyphony. I wasn't there myself so I don't know how true this account is, but that is what I was told.

If true, that's really terrible.


This was around 10 or 12 years ago, so I'm sure the technology back then is different from now, but I do believe the salesperson said the piano he was trying was one of the highest models at the time, a Yamaha.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/06/16 09:35 PM

Stephan, you can only afford what's in your budget and even so, 128 note polyphony was common for a whole generation of keyboards prior to the most recent higher end models. You can hear from the Ivory demos 128 note poly does the job with Ivory which in many ways is more demanding on the hardware it runs on than the dedicated DSP in digital pianos.

The Kawai CA67/97 or a Yamaha Clavinova 5XX are out of your budget. Are they "better" DPs than the Yamaha YDP-X. Yes! But for many reasons beyond polyphony. Keyboard action, monitoring system, build quality, aesthestics, etc.

I think what you need to do is either work within your budget, or go shopping and see the difference for yourself and if you need a better DP, then open up a high yield savings account and put your pennies away for a while before you jump in.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/06/16 09:38 PM

Originally Posted by Michiyo-Fir
This was around 10 or 12 years ago, so I'm sure the technology back then is different from now, but I do believe the salesperson said the piano he was trying was one of the highest models at the time, a Yamaha.


If it is an high-end Yamaha 10-12 years ago, it could be a CLP270, or CLP280 (2005) : 128 voices.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/06/16 09:39 PM

Originally Posted by Stephano
Originally Posted by Michiyo-Fir
That being said, the digital piano salesperson at my local store said that when Lang Lang was here for a masterclass, he played around on the digitals for a while and while he played certain arpeggios very fast the salesperson heard a small popping noise which he attributed to notes cutting out from the lack of polyphony. I wasn't there myself so I don't know how true this account is, but that is what I was told.

If true, that's really terrible.


I think you need to spend some time playing digital pianos without first reading the specifications, just using your ears.

Perhaps this is time to start a discussion on MonsterCable. smile
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/06/16 09:42 PM

As long as we cover power and ethernet cables too... wink
Posted By: Stephano

Re: 128 note polyphony question - 06/06/16 10:32 PM

Ok, I've decided to expand my budget a little, but maybe that little will be sufficient:

http://www.thomann.de/de/yamaha_ydp_163_b_arius.htm

It's the successor to the YDP162.

GH3 action and 192 note polyphony.

I know that the GH3 action is just a third sensor added, but the added polyphony makes it more appealing to me, as crazy as that sounds.

For what it's worth, I was totally ready to order the YDP162 today, and then I read propianist's post. I got over that and was ready to order again, and then I read this post about popping notes. So now it looks like it will be the YDP163.

Sorry if I'm irritating anyone with my posts.
Posted By: JoeT

Re: 128 note polyphony question - 06/06/16 10:41 PM

I never heard popping notes with 32 note polyphony, which I had to deal with until 2014.
Posted By: Macy

Re: 128 note polyphony question - 06/06/16 11:56 PM

Originally Posted by propianist

PLUS...
Any serious modern digital piano would add sympathetic string resonance overtone layers - for every strike of every note - and here, there is clearly going to be a lot of harmonic inter-relationship and reinforcement between all the C#'s, G#'s and E's over the 4 octave range of this arpeggio and extended by further harmonics beyond this to the extreme of the piano range. You could reasonably expect at least 3 or 4 layers of harmonic overtones to be triggered by a modern digital piano sympathetic string resonance algorithm (F, 2F, 3F, 4F etc.) so there could be 3 or 4 times as many samples sounded, in addition to the global sustain pedal resonance sample too.


Most DPs and VSTs do not use samples for sympathetic resonances or sustain pedal resonances (but a few do), and use DSP processing instead which is not counted as polyphony. Many have release samples but not all count them as additional simultaneous polyphony either. Some count stereo as 2 notes of polyphony, some do not. That's why I said above that polyphony is an apples and oranges spec from one DP or VST to another. You can't judge anything by comparing polyphony specs. You have to try them to know what they really do. And most VSTs have adjustable polyphony limits that will be practically limited by computer/RAM/disk drive performance.
Posted By: siros

Re: 128 note polyphony question - 06/07/16 12:47 AM

I think there are things you should be more worried about than not having more than 128 polyphony.
For example, the lack of string resonance and release samples.

If you play a piece with a lot of notes with pedal down, you may sometimes hit 128 polyphony limit.
But you will miss string resonance and release samples 100% of the time in whatever piece you play, if the DP does not support them.
It looks like those YDP models do not have them.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 01:59 AM

Originally Posted by siros
I think there are things you should be more worried about than not having more than 128 polyphony.
For example, the lack of string resonance and release samples.

If you play a piece with a lot of notes with pedal down, you may sometimes hit 128 polyphony limit.
But you will miss string resonance and release samples 100% of the time in whatever piece you play, if the DP does not support them.
It looks like those YDP models do not have them.

The lack of string resonance and release samples won't bother me. I played on a very old Clavinova when I lived in the States, which had neither of those, and I didn't have problems.

Right now I'm seriously considering the YDP-163. The only thing holding me back is the fact that there are zero reviews for it out there. It's understandable, seeing that it came out in April 2016, but I'd love to be able to read at least one single review for the product.
Posted By: petes1

Re: 128 note polyphony question - 06/07/16 04:55 AM

Originally Posted by Stephano
I was just about to order the YDP-162 with 128 note polyphony... and then I read propianist's post.

Blah.


Hopeless. Go buy something with the highest polyphony spec you can find, and sound be darned, and be done with it, and hopefully be happy. As for myself, I'll stick with what I said before and will trust my ears.
Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/07/16 04:59 AM

You've gotten me very curious about the polyphony limit. I suspect you won't be able to hear any notes dropping noticeably after the 128 note limit. I can actually go to the music store tomorrow to test that passage of my piece on the 2 YDPs (hopefully they have them) to see if I can hear anything dropping...

I will report back tomorrow night PST.
Posted By: Dr Bekken

Re: 128 note polyphony question - 06/07/16 07:03 AM

https://youtu.be/RElarmDDVrE

Here's a Yamaha P-35.
32 note polyphony.
Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/07/16 07:15 AM

Originally Posted by DrBekken
https://youtu.be/RElarmDDVrE

Here's a Yamaha P-35.
32 note polyphony.


The video leads me to a ukelele performance?
Posted By: propianist

Re: 128 note polyphony question - 06/07/16 12:07 PM

Originally Posted by Stephano
Ok, I've decided to expand my budget a little, but maybe that little will be sufficient:

http://www.thomann.de/de/yamaha_ydp_163_b_arius.htm

GH3 action and 192 note polyphony.

For what it's worth, I was totally ready to order the YDP162 today, and then I read propianist's post. I got over that and was ready to order again, and then I read this post about popping notes. So now it looks like it will be the YDP163.


Hi Stephano,
If you're now looking for high polyphony vs low price...

Casio PX860BK = 256 note poly = €949 = £746

Kawai CN14 = 192 note poly = €819 = £643

Remember, as others have already said, the main piano sound quality and feel / ergonomics are more important than tech specs, especially if you're a beginner. 128 note poly is usually okay for most people, but I'd say it's about the minimum figure you should accept if you're buying a brand new digital piano today in 2016. Don't buy something with 64 note poly, for instance, or you'll outgrow it one day.

19 years ago, I bought GEM RealPiano Module which had 64 note poly, and I could hear note-stealing dropouts quite often. It's annoying, but it's not the end of the world! That module was the world's first ever with proper sympathetic string resonance - it would be far more annoying not to have that feature, even for 128 note poly or better.
My 10 year old Kawai MP8 is 192 note poly and I rarely notice note dropouts on solo piano playing. You can make it do it, but it's not the end of the world! I'd still prefer to have more polyphony always of course.

Remember keyboards / digital pianos are often used for other things besides solo piano. eg. "Piano + Strings" layers - very popular - immediately halves the polyphony, so if you still want your piano to sound decent over the string pad, having 256 note poly isn't a bad idea.

Some people also use keyboards / digital pianos as General MIDI multitimbral synths for playback of Standard MIDI File songs, with dense orchestral / band arrangements, and they play live piano parts over the top too! Having enough extra polyphony headroom in reserve can be very important there! It depends whatever your needs and activity might be.

Popular software like Ivory II only defaults to 24 note poly when first installed because Synthogy want it first to run okay on most average people's computers without risk of crashing or audio glitches, so they start factory default with very low CPU demand settings. Users would then increase poly setting up to something like 160+ for serious playing, if their computer can handle it safely. Same reason why Ivory II defaults with sympathetic string resonance turned off to start with.
Posted By: propianist

Re: 128 note polyphony question - 06/07/16 12:28 PM

Originally Posted by Frédéric L
The level of the 64 voices version is -8.2dB, the level of the 128 is -7.7dB.
The 128 voices version is 0.5dB louder.


Good, well done.
So that demonstrates a real, measurable, audible difference going from 64 note poly to 128 note poly. QED.

To me, this is axiomatic.

Posted By: JoeT

Re: 128 note polyphony question - 06/07/16 12:43 PM

Originally Posted by propianist
So that demonstrates a real, measurable,

Yes.
Quote
audible

No.
Quote
difference going from 64 note poly to 128 note poly. QED.

On a digital recording there is more level with more notes sounding: You mix samples by adding them up digitally and the result is the sum. The digital sum is higher with more notes. This is simple math.

But human hearing doesn't care at all. That's the reason why measuring voltage or unweighted deziBels is useless to assess audible volume. You have to weight them or use real loudness units (Sone, Phon) to get accurate results.

Nevertheless a 0.5dB level difference is still inaudible.
Posted By: MacMacMac

Re: 128 note polyphony question - 06/07/16 01:00 PM

Bravo, Joe.

Yes, 0.5 dB is inaudible.

Which brings the point: If you can't tell the difference, there is no difference.
In piano ... music/art ... perception is all that matters. Perception is reality.

Which brings the next point: Specs don't matter. Even when they're honest, and especially when they're not.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 01:01 PM

Originally Posted by propianist
Remember, as others have already said, the main piano sound quality and feel / ergonomics are more important than tech specs, especially if you're a beginner.

Again, as I've pointed out already, I'm not a beginner. Hence my worries about polyphony. Think about the ending of the first movement of the Appassionata for example. Right before it slows down, prior to the Piu Allegro, there are many many fast notes with lots of sustain, and this goes on for a while.

Is 128 polyphony sufficient for that? I'm just using that as an example, as it is something that I play.

On the other hand, I fully agree with JoeT. The question is whether there is an audible difference or not. If I have to analyze the recording through software to determine if the polyphony limit has been reached, then I'll tell you right now that I don't care about it. That being said, the other post about "popping noises" due to running out of notes was the only thing that prevented me from ordering the YDP-162.

And the way I see it, the true polyphony test would be to post several recordings with different polyphony limits without saying which is which, and see if someone can determine if the limit has been reached on some of them.

So the audible difference, audible to the human ear, is all I care about. But if there is a risk of "popping noises", or the 129th note being nonplayable/silent, then that would be a real problem. I just didn't expect so much conflicting information, to be honest. And I do appreciate all the help.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/07/16 01:05 PM

Anyone here grow up reading the Hirsch-Houck audio reviews? We could use some of that thinking here.

I agree, if you can't hear the difference, there is no difference ... and if you still think there's a difference, a double blind test will confirm it ... or not.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/07/16 01:18 PM

Perception is important and if you aren't used to playing good quality acoustic pianos and don't have an expectation of what they sound like you're less likely to be disappointed by what this or that DP offers until you set them side by side and play and listen. At which time you'll prefer this or that action, this or that timbre in various styles of play. Once you've narrowed it down to a few, you'll look at convenience features like ability to record and playback performances, multi-track record, export to wav on a USB stick, audio file playback, USB MIDI io so you can upgrade your sound engine with a software piano, backing rhythms, etc. Then lastly, how it will look in the Mrs. living room perhaps - although, depending on how particular your partner in crime is, this might be higher on your list.

As has been eluded to, double blind tests are far less important than sitting down and playing. The whole instrument has to feel and sound right to you. The interaction between keyboard and resulting sound. It is quite possible a piano with less polyphony checks more boxes for you.

Posted By: bennevis

Re: 128 note polyphony question - 06/07/16 02:11 PM

Originally Posted by Stephano
Originally Posted by propianist
Remember, as others have already said, the main piano sound quality and feel / ergonomics are more important than tech specs, especially if you're a beginner.

Again, as I've pointed out already, I'm not a beginner. Hence my worries about polyphony. Think about the ending of the first movement of the Appassionata for example. Right before it slows down, prior to the Piu Allegro, there are many many fast notes with lots of sustain, and this goes on for a while.

Is 128 polyphony sufficient for that? I'm just using that as an example, as it is something that I play.


So the audible difference, audible to the human ear, is all I care about. But if there is a risk of "popping noises", or the 129th note being nonplayable/silent, then that would be a real problem. I just didn't expect so much conflicting information, to be honest. And I do appreciate all the help.

You know, I didn't even know what "polyphony" meant in relation to a digital, when I bought mine six years ago grin.

I've just re-read my DP manual and it says it has only 128-note polyphony cry. OK, it's a bit long in the tooth now (though it was state-of-the-art at the time), but it still plays and sounds like new - and I've never heard any note dropouts. And my DP is unique in that not only has it got full string, soundboard, and cross-resonances, you can push all of them up to way beyond that of any acoustic, as well as the sustain level (in increments of one, from -100 to +100). So, you'd think that when swimming in a wash of sound, with pedal down, notes dropping out is a real possibility, when (apparently) only 128 notes can be heard at any one time (though my digital only has 88 keys cry).

Well, I can't hear any notes disappearing, or not sounding, if I play - or keep repeating notes - with pedal down. BTW, I play only classical, and my current repertoire includes Gaspard de la nuit, Chopin Études and Rachmaninov Preludes (for which I use all three pedals). As well as the Appassionata wink .

Does that help?
Posted By: EssBrace

Re: 128 note polyphony question - 06/07/16 03:08 PM

Originally Posted by bennevis


Well, I can't hear any notes disappearing, or not sounding, if I play - or keep repeating notes - with pedal down. BTW, I play only classical, and my current repertoire includes Gaspard de la nuit, Chopin Études and Rachmaninov Preludes (for which I use all three pedals). As well as the Appassionata wink .

Does that help?


It should help. But I fear it will not. The OP is very preoccupied (one could almost say obsessional) about this element of DP specification. I get the impression that if given a stark choice he'd rather it sound like crap but have huge polyphony rather than sound or feel good but have some theoretical limitation in regard to polyphony.

I can't relate to that at all I'm afraid.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 03:35 PM

Originally Posted by bennevis
Does that help?

It would definitely help if I was buying your instrument. The thing is, the "popping noises" due to running out of notes were allegedly occurring on high-end Yamahas with 128 note polyphony.

And the YDP-162 I'm interested in is not a high-end instrument.

I think I said everything I wanted to say in this thread, and others have, too.
Posted By: JoeT

Re: 128 note polyphony question - 06/07/16 03:47 PM

Originally Posted by Stephano
I just didn't expect so much conflicting information, to be honest.

That's what you get when asking forums. They are for opinions, not an expert system. There is simply no way for you to tell the correct answer unless you know it already.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/07/16 03:59 PM

Originally Posted by Stephano
The thing is, the "popping noises" due to running out of notes were allegedly occurring on high-end Yamahas with 128 note polyphony.

And the YDP-162 I'm interested in is not a high-end instrument.

The 10-12 years old high-end Yamaha like the one used by Lang-Lang in a shop is less advanced than a YDP162. But I don't know if the tone generator of the YDP162 is less prone to make a popping noise when you play a demanding piece. The only way to know is to get into a shop and try one of them.
Posted By: EssBrace

Re: 128 note polyphony question - 06/07/16 04:11 PM

Originally Posted by Stephano
Originally Posted by bennevis
Does that help?

It would definitely help if I was buying your instrument. The thing is, the "popping noises" due to running out of notes were allegedly occurring on high-end Yamahas with 128 note polyphony.


Don't express this is if it's a fact. It's a third hand report from something ten or twelve years ago on a completely different instrument. Note dropping simply doesn't cause popping sounds.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/07/16 04:15 PM

Is this thread coming to an end so soon?

I was hoping we could keep this thread alive until 1024 note polyphony was standard fare.
Posted By: JoeT

Re: 128 note polyphony question - 06/07/16 04:18 PM

Originally Posted by Dave Horne
Is this thread coming to an end so soon?

I was hoping we could keep this thread alive until 1024 note polyphony was standard fare.

I bet, someone will post the rumor, that a digital will go out of tune past its polyphony limit. wink
Posted By: EssBrace

Re: 128 note polyphony question - 06/07/16 04:24 PM

Originally Posted by JoeT
I bet, someone will post the rumor, that a digital will go out of tune past its polyphony limit. wink


Rumour? I tell can you that I know for a fact that one big name brand of DP self destructs once the polyphony is exceeded. It goes bang, big time. Just to keep you all on your toes and to encourage a certain 'economy of effort' when it comes to pressing too many keys at once I've decided that I'm not at liberty to disclose the maker in question.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 04:31 PM

Originally Posted by Dave Horne
Is this thread coming to an end so soon?

It has ended as far as I'm concerned. The way I see it, everything that needed to be said, has been said.
Posted By: TheodorN

Re: 128 note polyphony question - 06/07/16 04:56 PM

Why are people worrying about a thread coming to an end or not? Nobody is obliged to follow it, nobody is obliged to post, neither the OP nor others. Discussions sometimes go on for months, even when the ones who started them, have stopped following them a long time ago.

This is my first post in it, and for my sake, it can be the last. Just want to link to the video below, which I've posted before, in a similar discussion. It's a recording of a demanding Rachmaninoff prelude.

Watch the voices count at the top. It goes a little over hundred voices at most. If 128 polyphony is enough for this particular piece, which has many notes at once, and requires heavy use of the sustain pedal, it should be enough for most classical pieces.

Posted By: Nordomus

Re: 128 note polyphony question - 06/07/16 05:08 PM

Actually you are wrong, in the video you've provided I've noticed moment around 3:22 when there are 136 voices smile
Edit: Even above 150 around 7:50
Posted By: Dave Horne

Re: 128 note polyphony question - 06/07/16 05:25 PM

The software might state 136 or 150 voices but I'd bet the software is counting more than just one count for one note depressed.

Posted By: TheodorN

Re: 128 note polyphony question - 06/07/16 05:40 PM

Nordomus, thanks for pointing that out. Dave Horne, so the count is too high, if we were comparing to just playing on a digital piano, using it's internal voices? That's good news!
Posted By: Frédéric L

Re: 128 note polyphony question - 06/07/16 05:47 PM

You're right Dave, a single note on Galaxy Vintage D counts for 3 when the sustain pedal is depressed.

But with Ivory, I get the same number of voices and even more (173).

But if a 64 to 128 voices evolution just add 0.5dB... I don't think the extra voices needed (128 to 150) will add a lot.
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/07/16 05:59 PM

We are still well below 128 (posts).
Posted By: TheodorN

Re: 128 note polyphony question - 06/07/16 06:01 PM

How do we count polyphony on a digital piano, with the internal voices used, no VST(i)s?
Posted By: EssBrace

Re: 128 note polyphony question - 06/07/16 06:04 PM

Originally Posted by oldmancoyote
We are still well below 128 (posts).


Oh yes, I see where you're going with this...

If PW's 'message dropping' implementation is sophisticated then only the most useless and vacuous posts will be dropped once the thread reaches 128 posts....I've a feeling mine will be the first to disappear!
Posted By: oldmancoyote

Re: 128 note polyphony question - 06/07/16 06:05 PM

Originally Posted by TheodorN
How do we count polyphony on a digital piano, with the internal voices used, no VST(i)s?
With difficulty. However, one factor simplifies life: internal voices have usually much shorter decay than good VSTs, which may render the whole issue rather moot... since high voice counts depend on relatively long decays as well as fast playing.
Posted By: propianist

Re: 128 note polyphony question - 06/07/16 07:23 PM

Originally Posted by JoeT
Nevertheless a 0.5dB level difference is still inaudible.

Originally Posted by MacMacMac
Yes, 0.5 dB is inaudible.


No, you're both wrong. 0.5dB difference is audible.

It's a 6% increase in sound pressure.

Six percent is not negligible. Maybe 0.006% doesn't matter, but 6% is a small but valid quantity to be aware of and take some notice.
If you got a 6% pay rise, or had 6% of your blood drained out, you'd notice. If tax or petrol price suddenly went up by 6% people would start complaining.

At what percentage would you guys actually start to care about stuff??!

Would you advocate still not caring about even a 1dB (12%) or 2dB (24%) difference in output due to polyphony between two piano renders?
Would you still keep insisting they are subjectively identical, quality wise?
Would you be happy to keep the lesser version, knowing it doesn't measure up to the other file and delete the better one?
Are you advocating that you'd be happy with a 64 note poly piano?!
Are you saying because you can't hear any differences that nobody else can or should be able to hear either?

This dumb ignorance is a baffling attitude to me!
If you guys are prepared to state outright that you cannot hear any difference in even 6% sound pressure amplitude change, how do you expect to critically judge fine parameters in your software pianos, or be confident about hi-fi equipment choices or musical listening in general?
Mastering engineers make their livings from making adjustments probably smaller than 0.5dB most of the time in their professional career.
Go onto any of these forums...

https://www.gearslutz.com/board/mastering-forum/
http://www.digido.com/forum/index.html
http://recording.org/forums/mastering.18/
http://forums.stevehoffman.tv/forums/music-corner.2/
http://www.head-fi.org/f/133/sound-science
http://www.computeraudiophile.com/forum/

... and tell them that 0.5dB is inaudible and they'll shoot you down in flames!!!!

Now, if anybody wants to know for themselves, try this great website I found!!!! where you can test yourself to see if you can hear 0.5dB level changes in two consequtive 440Hz sinewaves...

http://www.audiocheck.net/blindtests_level.php?lvl=0.5

At the top of the page you can change test to easier 6dB or 3dB or 1dB increments to get the hang of test method, or if you're a "golden ear" you can try 0.2dB or 0.1dB test, which I have to admit is very tough, but I can personally score 10/10 on the 0.5dB test if I occasionally reference myself inbetween test attempts by listening again to the three known level buttons before pressing the next question mark button. After a while my ear gets "dialed in" to the tone variations and I can hear them reliably. It just takes a little practise.

Anyway, to those who (wrongly) say 0.5dB is inaudible, what dB change can you guys reliably hear on a test like that? Can you manage 3dB okay? Honestly perfect score, no cheating? Ability to do this well probably depends how much of your lives you've spent concentrating on listening and worrying about audio levels and audio engineering in general. If you never give audio levels and dB values much thought, and "sound" doesn't invoke any numerical signal analysis ideas from you, but only musical shapes, timbres, notes, chords, then you might not fare well in a technical audio test like this. It's a different kind of listening skill.

Personally, I know that if I can hear a 0.5dB difference in sound, then I know the polyphony benefits which produce a level change of similar proportion must be audible even by virtue of dB difference alone, let alone the musical benefit of having all the notes included as they would be on an acoustic piano. I can't see any reason why a pianist wouldn't want that benefit, given the straight choice.
Posted By: TheodorN

Re: 128 note polyphony question - 06/07/16 07:33 PM

Oldmancoyote, I thought I'd solved this issue of polyphony with this video. Turns out I was wrong, obivously, since software pianos count the notes differently than digital pianos alone do, according to Dave.

Sorry to those who were hoping this thread was over, that I didn't succeed in ending it, if anything, I started the never-ending story again. eek
Posted By: Dave Horne

Re: 128 note polyphony question - 06/07/16 07:56 PM

If you play an arpeggio with both hands up and down for three octaves and add a run for three octaves all with the pedal down and you count every time you depress a key, the number is nowhere as high as everyone thinks.

If you kept the pedal down the entire time you'd have soup.

Just for the heck of it a while back I counted the number of times I could press the keys and keep the pedal down. I don't remember the count but it was low. If you play with a heavy foot it's going to sound like mush.

Counting each depressed key (as one count) with the pedal down, you will not come close to 128. (My assumption is one depressed key equals one count. If I'm wrong, great. For me this entire thread is an exercise in futility. smile )




Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/07/16 08:15 PM

Originally Posted by Dave Horne
If you play an arpeggio with both hands up and down for three octaves and add a run for three octaves all with the pedal down and you count every time you depress a key, the number is nowhere as high as everyone thinks.

If you kept the pedal down the entire time you'd have soup.

Just for the heck of it a while back I counted the number of times I could press the keys and keep the pedal down. I don't remember the count but it was low. If you play with a heavy foot it's going to sound like mush.

Counting each depressed key (as one count) with the pedal down, you will not come close to 128. (My assumption is one depressed key equals one count. If I'm wrong, great. For me this entire thread is an exercise in futility. smile )



When each note is pressed again with the pedal down, it counts as a 2nd use of the polyphony count according to the music store sales staff.

Look at the link of Balakirev's The Lark I posted earlier. It definitely uses more than 128 polyphony in one section, I have yet to count the other sections I suspect might use more than 128 too. And in my version it specifically has pedal notations that indicates my Pedal needs to be down for more than 128 notes polyphony provided each repeated note counts for an additional use of the 128.

I'm on my way to the music store right now to see if I can hear any dropped notes with 128 polyphony. Will report back later with my findings/discussion with the digital piano staff.

Edit: in my mind there are countless pieces more advanced than the lark so there could be a possibility that many pieces that exist in the classical repertoire that uses more than 128 polyphony although I don't have any other specific examples.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/07/16 08:31 PM

The audio test is interesting, but the ability to recognize a 0.5dB change between two sound played one just after one doesn't equal the ability to recognize a difference between two long extract.

And my 0.5dB measure has been done between a 64 voices and a 128 voices limit. It is not quite relevant since there are no digital pianos nowadays which have a 64 limit (excepted the Yamaha P45... and the Casio CDP130 which has even less). But with the Fantaisie Impromptu and the Sonata op27 which has been proposed... comparing 128 and more would have no sense since these pieces of music doesn't need all the 128 voices.

An analog measure with the Rachmaninov Prelude in C#, but with 128 and 500 voices gives a 0.3dB difference (a peak level... which has no musical sense, but I have no tool which could give me a RMS measure).

128 voices : http://www.sinerj.org/~loyer/piano/rachma_128.wav
500 voices : http://www.sinerj.org/~loyer/piano/rachma_500.wav

I have to admit I can't hear a difference.
Posted By: pinkfloydhomer

Re: 128 note polyphony question - 06/07/16 08:37 PM

I might be missing something, but why is more than 88-note polyphony ever needed, even with full pedal all the time?
Posted By: Frédéric L

Re: 128 note polyphony question - 06/07/16 08:49 PM

Originally Posted by pinkfloydhomer
I might be missing something, but why is more than 88-note polyphony ever needed, even with full pedal all the time?

If you trigger a note twice with the pedal depressed, the notes have to be added.

The easiest way to do it is to allocate a new voice which will play the second note.

Another way is to have 88 voices : one per key. And each time we strike a key, we change the volume of the voice adding a new note. However, it is not that easy : the sound has an envelop (attack, decay...). And adding two samples which are not synchronized can't be done easily. It can't be a straight volume change. Things are far more complicated. I think the newest modeled SuperNatural sound engine of Roland does something like this (I have no info, but since it is claimed to provide an infinite polyphony, there should be something like this.) Since it is modeled and not sample based, things could be easier.
Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/07/16 09:27 PM

I don't know if Stephano is still following this thread. But if you are, I tried the ydp 142 and 163 and couldn't notice the sound of notes dropping from either. I don't think I hit the polyphony count of the 163 anyways.

The biggest difference I feel is that the 142 has very light keys and has a pretty boomy bass that gets pretty loud without much force. I find the bass to be boosted too much for the light touch response that is has, especially compared to the treble/mid range.

The 142 also has a very slippery keytop.

The 163 is heavier in touch and has a more balanced response compared to what I'm used to. Both sound pretty decent with the 163 having a deeper/more solid mid low and low range without being boomy. And the 163 has less slick keytops that is slightly more matte in feeling.
Posted By: Dr Bekken

Re: 128 note polyphony question - 06/07/16 09:37 PM

Originally Posted by Michiyo-Fir
Originally Posted by DrBekken
https://youtu.be/RElarmDDVrE

Here's a Yamaha P-35.
32 note polyphony.


The video leads me to a ukelele performance?


Now, that's a mess up...how about this link..?:

https://youtu.be/rwFFi1THJo0
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 09:40 PM

Originally Posted by Michiyo-Fir
I tried the ydp 142 and 163

I wasn't interested in the 142 though. That one has a significantly poorer key action, called Graded Hammer Standard.

I was choosing between 162 and 163, not the 142.

In any case, I really appreciate you taking the time to try these instruments. Thank you. It is comforting to hear that there weren't any "popping noises" on the 128 polyphony instrument (YDP-142), and it's also good to know that the YDP-163 seems to have left a good impression on you.

I think I'll go for the 163.
Posted By: Frédéric L

Re: 128 note polyphony question - 06/07/16 09:51 PM

Originally Posted by Stephano
I think I'll go for the 163.

Even if you don't need the extra voices, you may enjoy the better keyboard. I am satisfied with the YDP162 keyboard (GH), but my level is far lower than yours.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 10:06 PM

Originally Posted by Frédéric L
Even if you don't need the extra voices, you may enjoy the better keyboard. I am satisfied with the YDP162 keyboard (GH), but my level is far lower than yours.

Do you think it'd be wise to order right away, or wait for some reviews/tests to come out?

Like I said in this post: (click here), I just wonder how likely it is for brand new Yamaha instruments to have some kind of flaw(s), and whether it'd be a good idea to wait for reviews of that product to come out?

But maybe I'd be waiting too long then. Heck, I just wish I could find one or two reviews for that product.

That said, Michiyo-Fir played that piano today, and seemed to have liked it, which is very comforting.
Posted By: Dave Horne

Re: 128 note polyphony question - 06/07/16 10:16 PM

Stephano, don't read reviews, buy only what you've actually played in a store for more than 30 minutes.

I've made purchases after reading a positive review only to be disappointed.

For me the feel of the action comes first with the sound a close second. Everything else is gravy.



Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/07/16 10:17 PM

Originally Posted by Stephano
Originally Posted by Michiyo-Fir
I tried the ydp 142 and 163

I wasn't interested in the 142 though. That one has a significantly poorer key action, called Graded Hammer Standard.

I was choosing between 162 and 163, not the 142.

In any case, I really appreciate you taking the time to try these instruments. Thank you. It is comforting to hear that there weren't any "popping noises" on the 128 polyphony instrument (YDP-142), and it's also good to know that the YDP-163 seems to have left a good impression on you.

I think I'll go for the 163.


Oh sorry, I missed that. My store doesnt have the 162 so I wouldn't have been able to try that one anyways.

The other thing I wanted to add is that the decay on these instruments (including most of the clavinovas) are pretty quick so before dropped notes even happen, some of the notes played earlier may have decayed anyways and stopped sounding compared to my grand when I tested it at home. In this way it's even harder to hit the polyphony limit since some of the notes are naturally gone. I didn't change/try to see if there are settings for longer sustain/decay though since I didn't know how the and rep wasn't around.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 10:36 PM

Originally Posted by Dave Horne
Stephano, don't read reviews, buy only what you've actually played in a store for more than 30 minutes.

I've made purchases after reading a positive review only to be disappointed.

For me the feel of the action comes first with the sound a close second. Everything else is gravy.

Thanks for the advice. Unfortunately I can't try this instrument. The store I called doesn't carry it, and it would be inconvenient for me at this moment, for several reasons, to drive an hour to another store.

I'm not at all worried about not liking the action though. I'm just worried about a brand new product having some sort of flaw/defect that no one knows about, only because it's recently been released. Maybe it is a bit foolish of me to think this way?

Believe it or not, and I mentioned this before in this thread, I can adapt very well to any action, and I was able to do so many years ago when I lived in the States. For a quite lengthy period I was practicing on a Steinway Grand only once a week, and the other days on a Yamaha Clavinova with pretty poor "AE" key action. I still adapted very well to various upright pianos when I performed, and to other grands, and other digitals. So at this point the only thing I'm worrying about is how wise it would be to order a Yamaha product that has been released only two months ago.

Sound is important to me, although this is subjective, and I don't really nitpick. I am familiar with the Yamaha sound, and I really like it. What I hear on the various demos on YouTube is also totally okay with me. smile
Posted By: JoeT

Re: 128 note polyphony question - 06/07/16 10:40 PM

Don't buy a digital piano you didn't try out before.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 10:54 PM

Originally Posted by JoeT
Don't buy a digital piano you didn't try out before.

Appreciate this advice, too, but right now that's not possible, for various reasons. That said, I've always liked every Yamaha I've played in the States many years ago. Now if the YDP-163 turns out to be an exception (which I doubt), I'll just have to send it back, and that's it. No harm done.

My only concern is, as I said, the product being recently released with no extensive reviews. I worry about some unknown "flaws", which is probably foolish.
Posted By: Michiyo-Fir

Re: 128 note polyphony question - 06/07/16 11:22 PM

It seems pretty unlikely to have any sudden flaws or poor design since I think it's just a refresh of the 162.
It's not a completely new line of product that's being released for the first time. Besides, since you're buying a new instrument you are covered under warranty and if any problems should arise (very unlikely I think), warranty will take care of the problem.
Posted By: Stephano

Re: 128 note polyphony question - 06/07/16 11:34 PM

Originally Posted by Michiyo-Fir
It seems pretty unlikely to have any sudden flaws or poor design since I think it's just a refresh of the 162.
It's not a completely new line of product that's being released for the first time.

That's pretty rational, yes.

And it's also good to know you didn't hate it when you played it.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/07/16 11:36 PM

Yeah, it's a non-worry on Yamaha digital pianos. They've been building these things for so long, and this is barely a revision, most definitely uses silicon and mechanical parts that are by and large the same for the entire Arius line.
Posted By: pinkfloydhomer

Re: 128 note polyphony question - 06/08/16 07:29 AM

Originally Posted by Frédéric L
Originally Posted by pinkfloydhomer
I might be missing something, but why is more than 88-note polyphony ever needed, even with full pedal all the time?

If you trigger a note twice with the pedal depressed, the notes have to be added.


Why? In a real piano, if I trigger a note twice, whatever the strings are doing will stop and the new note will be played instead.

As far as I can see, there are only the 89 different sound providers in a real piano: The 88 different notes (let's ignore the two or three unison strings of a note) and the soundboard which indeed will continue "playing" even after strings are muted.

You're probably right since I am not an expert in digital pianos. But what you're saying is that the set of strings of a certain note (together with the soundboard) will behave differently when a second note is triggered while the first note is still playing, compared to how they would behave when the first note wasn't already playing?
Posted By: jokke

Re: 128 note polyphony question - 06/08/16 11:01 AM

Originally Posted by pinkfloydhomer

Why? In a real piano, if I trigger a note twice, whatever the strings are doing will stop and the new note will be played instead.


That's a good question, if we would simplify the piano be a set of 88 single strings cut to proper lengths to give out the proper frequencies, is one number per string (amplitude) enough? In this case we'd count out multiple strings, the soundboard and other resonating parts out of the equation.

Could someone with physics background remind me, if a 440 Hz string can resonate to multiple frequencies (like 880 Hz), or at multiple phases simultaneuously? Or could we store phase and amplitude with a single complex number (that is how adding different phase sine waves work if I recall correctly)?
Posted By: Frédéric L

Re: 128 note polyphony question - 06/08/16 11:16 AM

Yes, a string will resonate simultaneously at different frequency. It has multiple modes. If a string resonates only at one frequency, the timbre will be poor and we won't have sympathetic resonance.

It is possible that sinus decomposition of the sound made by the different strings is done by modelled pianos (Pianoteq and now modelled SuperNatural by Roland).

The Patent EP2047455 B1 (by Modartt which sells Pianoteq), says :

Quote
Device [...] characterised in that the presynthesis module is adapted to determine excitation parameters that are representative of the initial amplitude and of the phase difference of each partial of the notes.


If the patent is used in Pianoteq, this software calculates the different partials instead of looping a sample.
Posted By: pinkfloydhomer

Re: 128 note polyphony question - 06/08/16 11:19 AM

Originally Posted by jokke

That's a good question, if we would simplify the piano be a set of 88 single strings cut to proper lengths to give out the proper frequencies, is one number per string (amplitude) enough? In this case we'd count out multiple strings, the soundboard and other resonating parts out of the equation.

Could someone with physics background remind me, if a 440 Hz string can resonate to multiple frequencies (like 880 Hz), or at multiple phases simultaneuously? Or could we store phase and amplitude with a single complex number (that is how adding different phase sine waves work if I recall correctly)?


A piano string can and does resonate at multiple frequencies at once (integer multiples of the fundamental frequency, disregarding inharmonicity). But it doesn't matter in the case we're discussing. It doesn't matter if we're talking about full samples of a note (which will contain several frequencies at once) or physical modelling (in which case the audio will also contain several frequencies at once). What matters is what happens when I play a note that is already playing. I can't see why it would matter that the note is already playing.

Of course, there is the possibility that the strings will behave differently when hit by the hammer while they're already resonating. But even so, the correct way to model this would not be to just "add the second note". The already playing note would be replaced by "something else", still only requiring 88 note polyphony.
Posted By: JoeT

Re: 128 note polyphony question - 06/08/16 12:39 PM

Originally Posted by pinkfloydhomer
Of course, there is the possibility that the strings will behave differently when hit by the hammer while they're already resonating.

If you strike a string unison two times without dampening it in between, once with fortissimo and once with piano, the fortissimo sound won't stop ringing, because the energy put into the string previously doesn't suddenly vanish.

It won't even do that, when slightly dampened it in between and it's a common implementation flaw in both digital pianos and VSTs if they completely lose the previous tone in that case.

Quote
But even so, the correct way to model this would not be to just "add the second note". The already playing note would be replaced by "something else", still only requiring 88 note polyphony.

Adding another note is the simplest way to replicate the real sound.
Posted By: danielp11

Re: 128 note polyphony question - 06/08/16 12:41 PM

For me the action is the most important thing.

Besides owning an Schimmel upright (and playing a Steinay Grand of my teacher) i bought a yamaha p105 for weekdays practice in my other apartment. The main problem with the p105 was the short pivot point of the keys which made them very hard to play when playing "deep inside" the keys, which is often neccessary. I changed to the kawai CA97 which i think has the longest pivot point of all digital pianos (same as acoustic grands). Now everthing is fine... The 128 note polyphony of the p105 was never an obstacle.

Daniel
Posted By: Stephano

Re: 128 note polyphony question - 06/08/16 01:27 PM

Originally Posted by danielp11
For me the action is the most important thing.

Besides owning an Schimmel upright (and playing a Steinay Grand of my teacher) i bought a yamaha p105 for weekdays practice in my other apartment. The main problem with the p105 was the short pivot point of the keys which made them very hard to play when playing "deep inside" the keys, which is often neccessary. I changed to the kawai CA97 which i think has the longest pivot point of all digital pianos (same as acoustic grands). Now everthing is fine... The 128 note polyphony of the p105 was never an obstacle.

Daniel

The P105 you played has a significantly poorer Yamaha key action called Graded Hammer Standard (GHS).
Posted By: danielp11

Re: 128 note polyphony question - 06/08/16 02:53 PM

Originally Posted by Stephano
Originally Posted by danielp11
For me the action is the most important thing.

Besides owning an Schimmel upright (and playing a Steinay Grand of my teacher) i bought a yamaha p105 for weekdays practice in my other apartment. The main problem with the p105 was the short pivot point of the keys which made them very hard to play when playing "deep inside" the keys, which is often neccessary. I changed to the kawai CA97 which i think has the longest pivot point of all digital pianos (same as acoustic grands). Now everthing is fine... The 128 note polyphony of the p105 was never an obstacle.

Daniel

The P105 you played has a significantly poorer Yamaha key action called Graded Hammer Standard (GHS).


Hello Stephano.

For comparison look at these photos.
pivot point
The 200% force mark for a grand piano is about the end of the black key (long pivot point) The 200% mark for a yamaha GH Action (not ghs as the PH105) is still in the middle of the black key. The kawai GFII (CA97 and CA67) is very similar to a grand action in this respect http://www.kawaius.com/digital/Features/actions.html

Daniel
Posted By: Stephano

Re: 128 note polyphony question - 06/08/16 04:37 PM

Originally Posted by danielp11
Hello Stephano.

For comparison look at these photos.
pivot point
The 200% force mark for a grand piano is about the end of the black key (long pivot point) The 200% mark for a yamaha GH Action (not ghs as the PH105) is still in the middle of the black key. The kawai GFII (CA97 and CA67) is very similar to a grand action in this respect http://www.kawaius.com/digital/Features/actions.html

Daniel

Thanks for clarifying! I still think the GH or GH3 would have made your playing experience overall more enjoyable than the GHS instrument you tried. Regardless, both the CA97 as well as the CA67 are way above my budget, and with my current budget I can't really ask for extraordinary or near perfect grand piano emulation.

At this point, I can only choose between YDP-162 and YDP-163, and after some more pondering today, I finally decided on the newer YDP-163.

So I ordered it today.

Will report back with my thoughts as soon as I'm able to play it. I will post in that other thread though, because this one is about polyphony and it shouldn't go off-topic.
Posted By: ElmerJFudd

Re: 128 note polyphony question - 06/09/16 12:00 AM

Mission accomplished. Enjoy!
Posted By: R_B

Re: 128 note polyphony question - 07/09/16 11:41 AM

Originally Posted by ElmerJFudd
It does depend on how much you use the pedal. You can easily hit 128 voices by putting the pedal down and doing a glissando. An acoustic piano does not suffer from voice limitations, therefore it is a reasonable objective of a digital piano not to suffer from note-cutoff/voice stealing.

Some digital pianos and keyboard instruments allow you to do splits and layers. For example - a string section sound along with the piano samples. The greater the number of available voices - the more capable the instrument is at performing splits and layers without note-cutoff/voice stealing. I've done many gigs with one keyboard where I layer acoustic piano, electric piano, strings, and brass instruments and use a volume fader on each layer to bring in or take out each of these sound groups. If I didn't have the poly, this would not be possible.



"An acoustic piano does not suffer from voice limitations"
Hmmm, maybe, maybe not.
Certainly multiple strings "compete" to vibrate the soundboard, there are winners and losers.
Whether we call that harmonious interaction or voice limitation may be a matter of semantics.
In simulating a wooden piano it is a fairly simple question of dominance, not one of "running out of notes".

Posted By: Dave Horne

Re: 128 note polyphony question - 07/09/16 11:54 AM

... and for some listening pleasure ...

Posted By: ElmerJFudd

Re: 128 note polyphony question - 07/09/16 03:15 PM

Originally Posted by R_B
Originally Posted by ElmerJFudd
It does depend on how much you use the pedal. You can easily hit 128 voices by putting the pedal down and doing a glissando. An acoustic piano does not suffer from voice limitations, therefore it is a reasonable objective of a digital piano not to suffer from note-cutoff/voice stealing.

Some digital pianos and keyboard instruments allow you to do splits and layers. For example - a string section sound along with the piano samples. The greater the number of available voices - the more capable the instrument is at performing splits and layers without note-cutoff/voice stealing. I've done many gigs with one keyboard where I layer acoustic piano, electric piano, strings, and brass instruments and use a volume fader on each layer to bring in or take out each of these sound groups. If I didn't have the poly, this would not be possible.



"An acoustic piano does not suffer from voice limitations"
Hmmm, maybe, maybe not.
Certainly multiple strings "compete" to vibrate the soundboard, there are winners and losers.
Whether we call that harmonious interaction or voice limitation may be a matter of semantics.
In simulating a wooden piano it is a fairly simple question of dominance, not one of "running out of notes".



If it's semantics why start a conversation on it? The point is, I can hold the pedal down on an acoustic piano and gliss my hands up and down it. All the strings will ring together. Some louder, some softer of course, and my ear will be drawn to this or that relationship perhaps if I'm inclined to listen in that way. Or I may perceive it as just one thing, a sonic mish mosh. You delve into the psychology of perception now. Let's leave it as semantics and call it a day.
Posted By: R_B

Re: 128 note polyphony question - 07/19/16 11:06 AM

Originally Posted by ElmerJFudd
Originally Posted by R_B
Originally Posted by ElmerJFudd
It does depend on how much you use the pedal. You can easily hit 128 voices by putting the pedal down and doing a glissando. An acoustic piano does not suffer from voice limitations, therefore it is a reasonable objective of a digital piano not to suffer from note-cutoff/voice stealing.

Some digital pianos and keyboard instruments allow you to do splits and layers. For example - a string section sound along with the piano samples. The greater the number of available voices - the more capable the instrument is at performing splits and layers without note-cutoff/voice stealing. I've done many gigs with one keyboard where I layer acoustic piano, electric piano, strings, and brass instruments and use a volume fader on each layer to bring in or take out each of these sound groups. If I didn't have the poly, this would not be possible.



"An acoustic piano does not suffer from voice limitations"
Hmmm, maybe, maybe not.
Certainly multiple strings "compete" to vibrate the soundboard, there are winners and losers.
Whether we call that harmonious interaction or voice limitation may be a matter of semantics.
In simulating a wooden piano it is a fairly simple question of dominance, not one of "running out of notes".



If it's semantics why start a conversation on it? The point is, I can hold the pedal down on an acoustic piano and gliss my hands up and down it. All the strings will ring together. Some louder, some softer of course, and my ear will be drawn to this or that relationship perhaps if I'm inclined to listen in that way. Or I may perceive it as just one thing, a sonic mish mosh. You delve into the psychology of perception now. Let's leave it as semantics and call it a day.


It may only be semantics - in fact the term "polyphony" may be at the heart of the problem.
I don't remember when it started, or when it came into general use, but it became a spec sheet boast item.
A glis can be nice, but in a MUSICAL context it hardly matters how distinctly each note can be heard - and by the way I only hit 7 notes per octave, so a capacity to represent 128 tones is adequate.
In an electronic instrument they all come together as a complex wave and we HEAR the speaker - just as they all come together in a wooden piano and we HEAR a board of wood.

IMO 128 note "polyphony" is adequate for MUSIC.
It is for others to opine whether or not it is sufficient to compete in the spec sheet war.
Posted By: toddy

Re: 128 note polyphony question - 07/19/16 01:22 PM

That's a great point! This debate keeps going because there are facts on both sides. But I don't think facts, pure and simple, are going to decide the issue. I think it's more a sense of proportion. The reason we worry endlessly about polyphony is because the capacity exists, on the digital side, due to the effect of Moore's Law.

The equivalent in acoustic terms would require an endless supply of low cost mechanical gadgetry, piles of seasoned hard wood and a preposterous amount of free space. Completely out of the question! So, until the possibility of hundreds of voices of polyphony appeared with computer technology, no one even conceived of such things, apart possibly from crackpots like Scriabin

And as other people have already said, polyphony is a big feature on the specifications sheet where other points of design of greater value are ignored because they are all but impossible to quantify.

The fidelity of chromatic resonance using sustain pedal and other resonant effects will depend on very clever programming of DSP.....ie modelling, not so much on available polyphony.
Posted By: R_B

Re: 128 note polyphony question - 07/22/16 11:57 AM

As I recall "multi timbral" was the buzz word that preceded polyphony, or maybe that was the term for multiple voice splits. Either way there were only 8 or 16 to go around, so some algorithms based on LRU and bass dominance had to be used.

@toddy
Well, there have been pieces written for two, three, four and probably more pianos.
...and of course there is HUGE "polyphony" (multi timbrality ?) in ensembles and .....Oh, 120+ piece ORCHESTRAS

So the demand has been there since WAY before Moore.

Back to the music.
Posted By: Frédéric L

Re: 128 note polyphony question - 07/22/16 12:52 PM

Multi-timbral is often associated with the 16 MIDI channels. With USB a synthesizer can be seen as multiple MIDI device, this can enable 32 or 64 channels.

To play a 120+ Instruments orchestra, we don't use 1 channel per instrument (we can't), but 1 channel per ensemble. Some instrument libraries include ensemble of instrument playing all the same note (and using one voice of polyphony to play this note)
Posted By: toddy

Re: 128 note polyphony question - 07/22/16 07:29 PM

Originally Posted by R_B


@toddy
Well, there have been pieces written for two, three, four and probably more pianos.
...and of course there is HUGE "polyphony" (multi timbrality ?) in ensembles and .....Oh, 120+ piece ORCHESTRAS

So the demand has been there since WAY before Moore


Of course you are so right! I really slipped up there.

As for the word 'polyphonic', I think this use of the word (as opposed to the academic one describing pre-harmonic texture) goes back to the days of analogue synthesisers (70s) which were able to play more than one note. Those able to play only one note at a time as was the case with the early classics like the Mini Moog were called monophonic.
Posted By: Frédéric L

Re: 128 note polyphony question - 03/24/17 09:24 PM

My digital piano - Yamaha CLP150 - is given to produce 128 notes Max. (se http://usa.yamaha.com/products/musical-instruments/keyboards/digitalpianos/clp_series/clp-150/)

Then I have done the following experiment :

Sending a steady stream of 32 C4, then after a short time, 32 G4. I could hear both the C4 and the G4.

Sending a steady stream of 64 C4, then after a short time, 64 G4. The G4 doesn't pass !! All the voices are busy producing the C4.

Then the actual polyphony is 64 notes and not 128 as I could read on the specs. I found it weird, my meaning of a "note" is the 2 channels I can hear, not just one. But a "Yamaha note" seems to be half a note (left or right channel).
Posted By: Gombessa

Re: 128 note polyphony question - 03/24/17 09:31 PM

Originally Posted by Frédéric L

Sending a steady stream of 64 C4, then after a short time, 64 G4. The G4 doesn't pass !! All the voices are busy producing the C4.

Then the actual polyphony is 64 notes and not 128 as I could read on the specs. I found it weird, my meaning of a "note" is the 2 channels I can hear, not just one. But a "Yamaha note" seems to be half a note (left or right channel).


A couple of questions:
1. Is that the whole "stereo sampling takes 2 notes of polyphony" thing?

2. I'm not quite sure how you're running the test, but if you played 64 C4 notes and then 1 G4 note, would the single G4 then pass? What about 63 G4 notes?

3. Last one--is it reasonable to conclude from this test that your Yamaha doesn't use any kind of advanced note stealing algorithm? It just drops anything new if the queue is already full, rather than find an old/decayed note to replace?

Thanks!
Posted By: Frédéric L

Re: 128 note polyphony question - 03/24/17 09:48 PM

1/

I suppose it is related to the stereo sampling. My 2002 digital piano has a basic AWM tone generator : no tone blending or whatever tricks which could eat polyphony voices. Then I suppose 128 is divided by 2 because of the stereo.

2/

The 64 C4 were sent by a DAW (Reaper) through a MIDI link. This can make some artifacts since not all the C4 are sent at the same time (around 1ms between each).

I have sent 64 G4 to be sure to get a volume comparable with the 64 C4. I think that if I don't hear any of the 64 G4, it would be the same with only 1 or 63.

3/

I am puzzled by the result. I would expect the G4 to stop C4. Perhaps, the first G4 started to decay a C4 note, the second one see that a voices is decaying and then do not start to decay a second one... And since all the G4 where sent to fast, none would wait the decaying C4 to be free. It is just a guess : I no nothing about what happended.
Posted By: Morodiene

Re: 128 note polyphony question - 03/25/17 01:37 AM

What about effects? Doesn't that add to polyphony as well?

I agree the behavior is odd, and perhaps this was before they figured out how to make the excess notes drop off from the earliest played, not the most recent.

The most important thing, of course, is can you hear it when playing an actual piece of music? If so, then it's not enough for you wink
Posted By: Charles Cohen

Re: 128 note polyphony question - 03/25/17 03:39 AM

Originally Posted by Morodiene
What about effects? Doesn't that add to polyphony as well?


Essentially, no. The "effects" are applied to the whole audio signal, not to each note individually. At least that's how traditional analog gear has always worked.

[/quote]
The most important thing, of course, is can you hear it when playing an actual piece of music? If so, then it's not enough for you wink [/quote]

+1.

Generating a MIDI stream, using software, that will overwhelm a sound generator isn't too hard. Actually _playing_ something that will overwhelm a sound generator -- that's a little trickier!

I've held down the pedal on my DP, and hit the keys with my forearms. No glitches out of Pianoteq.


Posted By: Frédéric L

Re: 128 note polyphony question - 03/25/17 05:58 AM

Charles Cohen, yes, you are right : on my comparison I wouldn't find the difference between a 64 notes and a 128 notes rendered polyphony, then I am happy with a 64 notes polyphony digital piano.

But what I dislike is that we can't trust the specifications which are meaningless. We should have a polyphony written in a number of "actual notes".
Posted By: 8 Octaves

Re: 128 note polyphony question - 03/25/17 06:44 AM

All specs from all manufacturers are to some extent lies. The specs are still useful if all the competing manufacturer lie the same way, then the number they publish is still meaningful for comparison. It's true of horsepower, mile per gallon, gigabytes, or polyphony.

When was the last time a 1 TB hard drive actually had 1 terabytes? Usually they have 1,000,000,000,000 bytes, when a terabyte is actually 1,099,511,627,776. That means a 1 TB hard drive is typically shorting you by some 10% or almost 100 GB. If every manufacturer does the same thing, then 1 TB from one brand is the same as 1 TB from another even if it is not 1 TB in the real world.

That's just one example. Everything in tech is like that if the company has a marketing department.
Posted By: Frédéric L

Re: 128 note polyphony question - 03/25/17 07:06 AM

The problem is that here we have a x2 factor... (and it would be interesting to see what mean a voice polyphony on a SuperNatural piano).

When I see "note" in a spec I understand "note", not a part of the sound which produce a note.

When we have a 115cv motor, we know it is the power of the engine... what would you think of an engine maker which would sold a so called 300cv engine but only measure the power dissipated by the combustion (and forget to tell you the actual power is 36% of it) ?

You forget to mention the speed of inkjet printers... We had in a time some really fancy figures as if the printers where measured to print blank pages. Now with ISO/IEC 24734, I think the situation is clearer.
Posted By: Jasper E.

Re: 128 note polyphony question - 03/25/17 08:19 AM

Originally Posted by 8 Octaves

When was the last time a 1 TB hard drive actually had 1 terabytes? Usually they have 1,000,000,000,000 bytes, when a terabyte is actually 1,099,511,627,776 bytes. That means a 1 TB hard drive is typically shorting you by some 10% or almost 100 GB.


Unfortunately they are correct. Terra prefix means factor of 1000^4, so 10^12, according to good old SI.

1,099,511,627,776 bytes = 1 tebibyte = 10 TiB.

If they would talk about 1TiB and provided 1TB they would be pretty eligible for a fine for misleading advertising...
Posted By: Jasper E.

Re: 128 note polyphony question - 03/25/17 08:21 AM

Again, I would be pretty interesting about measuring polyphony for classical four hands pieces.
Not the usual usage I agree, but I bet four hands need more polyphony.
Posted By: Frédéric L

Re: 128 note polyphony question - 03/25/17 08:28 AM

Have you got some pieces of music ?

If I can find the corresponding MIDI files, I can try.

I would suppose that each part of the piece is simpler than the Fantaisie Impromptu I have rendered, then you would not need twice the polyphony. The use of the pedal is important with the polyphony.
Posted By: Frédéric L

Re: 128 note polyphony question - 03/25/17 09:42 AM

I have seen that the CLP150 with 128 notes "Max" differs from other models from the same lineup (CLP120, CLP130) with 64 notes "Max".

They have not the same sound generator. (According to the CLP120 and CLP130 service manual), but the CLP150 has also 480 XG sounds.

When I try with the XG piano sound, I can reach 128 voices... But the "normal" way of playing my digital piano is the default "64 notes" piano.
Posted By: 8 Octaves

Re: 128 note polyphony question - 03/25/17 04:44 PM

Originally Posted by Jasper E.
Unfortunately they are correct. Terra prefix means factor of 1000^4, so 10^12, according to good old SI.


That is incorrect. Computers are not decimal and nor SI. All units are counted in binary. One thousand bytes (KB) = 1024 bytes in decimal or 2^10 bytes in binary. One MB is 2^20. One GB is 2^30. One TB is 2^40.

BTW, this is a recent practice they switched over to market USB flash storage, the practice of which bleed over to disc storage. I still remember buying hard drives when 120 MB drive had 2^27 bytes and the operating system reported exactly 131072 KB for a 120 MB drive. (Microsoft/IBM DOS reported all disc space in KB.)
Posted By: Frédéric L

Re: 128 note polyphony question - 03/25/17 05:51 PM

The problem is that a kB is near a KiB (2%), then one unit is often used instead of the other. If we get further (MB vs MiB, GB vs GIB, TB vs TiB), it is less true.

Then it seems me normal to use the TB for its actual meaning 1 000 000 000 000 B. k/M/G/T are SI prefixes.

When I read that a network is 100Mbit/s, this means 100 000 000 bit/s exactly. What an OS reports shouldn't make a "case law" about what units means what.
Posted By: 8 Octaves

Re: 128 note polyphony question - 03/25/17 07:31 PM

Originally Posted by Frédéric L
When I read that a network is 100Mbit/s, this means 100 000 000 bit/s exactly. What an OS reports shouldn't make a "case law" about what units means what.


Sorry this is simply wrong. The computer industry has never used SI. 64K has never been 64000. It has always been 65536, going back to the beginning of time.

The fundamental difference is, 1 kg = 1000 g but 1 Kb = 1024 bits. You simply have to accept it as the fundamental of computing.
Posted By: Frédéric L

Re: 128 note polyphony question - 03/25/17 10:15 PM

When I say that 100Mbit/s is 100 000 000 bit/s, I mean that the 100baseT data link has a 125MHz clock (The MHz from the SI unit system).... and since 4 bits are exchanged every 5 clock signals, we have 100 000 000 bit/s.

We have also 9.6kbaud for 9600baud and so on. The 64Kbit/s (ISDN) is also 64 000bit/s (which explains that the voice over ISDN is 8 bits at a 8kHz sampling frequency).

Then in data computing, we have some Mbit which are 10^6 bits and some other which are 1024*1024 bits. This explains the introduction of the Ki Mi Gi Ti prefix which makes things clearer. However, we still read about 4GB memory modules which are in fact 4GiB.

We still have both conventions.

We even have some weird convention : a 1.44MB floppy disk is twice a 720KiB floppy disk. Then the MB from 1.44MB is neither a real MB nor a MiB !!! Here it is 1000*1024 B.
Posted By: 8 Octaves

Re: 128 note polyphony question - 03/25/17 10:40 PM

I was actually there in 1998. There reason there is now the different terminology you so clearly pointed out is because of marketing, which is what I was saying. Some vendors that sat on standards bodies, my company included, wanted to be able to change the then accepted standard engineering terminology in order to market something smaller and call it bigger. That's why in 1998, gibibyte was made and GB was aligned to SI. I'm not saying the terminology is wrong even I objected to it back then, and still do. I'm saying consumers were robbed. The motivation was pure marketing, but we made it official for marketing purposes.

It's just like the number of speakers in a DP. It is now convention to count every single speaker cone as speakers even though we never did that before DP came along. How many speakers are here? http://revelspeakers.com/productdetail/~/product/salon2.html. If I was a DP maker, I would say 6 even though it has always been just 1. Just like some sales person on the Internet claiming how there is actually over 100 feet difference between a 5.5 foot and a 6 foot piano because there are 220 strings in a piano. I don't know how people get away with stuff but they will do whatever it takes to claim a higher number to the consumer.
Posted By: MacMacMac

Re: 128 note polyphony question - 03/25/17 11:20 PM

I, too, dislike disreputable specs ...
Originally Posted by Frédéric L
But what I dislike is that we can't trust the specifications which are meaningless. We should have a polyphony written in a number of "actual notes".
But MOST specs are "junk", so I find it best to ignore them.

And it's easy to do that. All you want from a piano is:
- Good feel. There's no spec for that. Let your fingers be the judge.
- Good sound. There's no spec for that. Let your ears be the judge.
- Some people care about size and weight. Let your eyes and your back be the judges.
- Appearance. Eyes, once again.

It seems to me that there is no need for ANY specs. So whatever specs are touted, I can just ignore them.
Posted By: 8 Octaves

Re: 128 note polyphony question - 03/25/17 11:40 PM

I tend to agree. As a musical instrument, specs are pointless. However, I think knowing size and weight is important. It was really helpful to know when I bought my piano that it weighs 704 pounds along with WxHxD, so I could provide an appropriate space with floor that's strong enough to accommodate it. Everything else is unimportant. The marketing literature could say whatever, but the proof is in the playing. From the specs, I would have never guessed that I would prefer the AvantGrand N2 to the N3, for instance.
Posted By: Frédéric L

Re: 128 note polyphony question - 03/26/17 03:16 AM

Some functions can be nice to have (record to WAV, AUX-IN...) and the specs can be useful to spot models which have them. But you are right, to judge a piano, we must consider only its actual use

But I am a little curious and like to know what is under the hood (I have also the service manual). But this doesn't interfere with my judgment.
Posted By: Jasper E.

Re: 128 note polyphony question - 03/26/17 07:55 AM

Originally Posted by 8 Octaves
Originally Posted by Frédéric L
When I read that a network is 100Mbit/s, this means 100 000 000 bit/s exactly. What an OS reports shouldn't make a "case law" about what units means what.


Sorry this is simply wrong. The computer industry has never used SI. 64K has never been 64000. It has always been 65536, going back to the beginning of time.

The fundamental difference is, 1 kg = 1000 g but 1 Kb = 1024 bits. You simply have to accept it as the fundamental of computing.


I have checked a bit further. As far as I see there is only one area where 1k = 1024, and that is semiconductor storage capacity -- for memory integrated circuits and microprocessors (L1/L2 caches). Also see JEDEC 100B.01 standard from 1997.

As hard disks are different technology, they were never compliant to JEDEC...

I wonder if SSD drive sizer are JEDEC or SI compliant... However, JEDEC only specified GB but not TB... and in their newest standard they also moved into the direction of using GiB and TiB, to avoid confusion.
Posted By: Jasper E.

Re: 128 note polyphony question - 03/26/17 07:57 AM

I partly disagree: specs are still somewhat useful.
Especially if you can know certain models of the same manufacturer share the same keyboard and/or sound engine or not smile
Posted By: Frédéric L

Re: 128 note polyphony question - 03/26/17 08:21 AM

We have a whole timeline of the use of SI vs IEC prefixes : https://en.wikipedia.org/wiki/Timeline_of_binary_prefixes
© 2020 Piano World Piano & Digital Piano Forums