2017 was our 20th year online!

Welcome to the Piano World Piano Forums
Over 3 million posts about pianos, digital pianos, and all types of keyboard instruments
Join the World's Largest Community of Piano Lovers (it's free)
It's Fun to Play the Piano ... Please Pass It On!

Shop our online store for music lovers
SEARCH
Piano Forums & Piano World
(ad)
Wessell Nickel & Gross
PianoForAll
(ad)
Best of Piano Buyer
 Best of Piano Buyer
(ad)
Faust Harrison Pianos
Faust Harrison 100+ Steinway pianos
Who's Online Now
43 members (David B, Chrispy, 1957, Beowulf, dannylux, almo82, anotherscott, CoogerTown, 5 invisible), 423 guests, and 420 robots.
Key: Admin, Global Mod, Mod
(ad)
Estonia Pianos
Estonia Pianos
Previous Thread
Next Thread
Print Thread
Hop To
Page 4 of 5 1 2 3 4 5
Re: How to Measure Latency?
newbert #2514754 02/25/16 02:25 PM
Joined: Feb 2010
Posts: 5,531
W
5000 Post Club Member
Offline
5000 Post Club Member
W
Joined: Feb 2010
Posts: 5,531
I measured the latency in my piano, it seems in the order of 4ms.

Left mic right under the keybed, under the key where I press.

Right mic right next to the string that is being hit.

I'm just pressing with moderate force, like playing mf. Not banging at all, to make as little key noise, only a bump when bottoming.

Gives me this

[Linked Image]

You can see noise in the RIGHT channel (where the hammer is) already BEFORE the key thump in the left channel is reported. That is to understand, as the action is already making noises before I hit the key bed.

Then the latency between hitting the bottom and the first big peak in sound near the string is only 4 ms.

Last edited by wouter79; 02/25/16 02:28 PM.

[Linked Image][Linked Image][Linked Image][Linked Image]
Re: How to Measure Latency?
wouter79 #2514766 02/25/16 03:06 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
Originally Posted by wouter79
I measured the latency in my piano, it seems in the order of 4ms.

Left mic right under the keybed, under the key where I press.

Right mic right next to the string that is being hit.

I'm just pressing with moderate force, like playing mf. Not banging at all, to make as little key noise, only a bump when bottoming.

Gives me this

[Linked Image]

You can see noise in the RIGHT channel (where the hammer is) already BEFORE the key thump in the left channel is reported. That is to understand, as the action is already making noises before I hit the key bed.

Then the latency between hitting the bottom and the first big peak in sound near the string is only 4 ms.


I am not convinced that you have found the beginning of the tone of the note played. To find it, I had to measure the period of the peaks and compare that to the expected period of the note I was playing, then I chose the first rise of the peak across the zero reference line. This ensures that the noise of the key moving through letoff and the sound of the key hitting the keybed is not mistaken for the note. I think the area from .112 on in your example above is not a tone, but more noise. See my example above.

Re: How to Measure Latency?
newbert #2514785 02/25/16 04:02 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
Here is my AP showing the latency difference between a forte and a piano strike.

[Linked Image]

The struck note is A3 at about 219.12Hz. All energy between the yellow lines is noise from the finger/key/whippen/shank/hammer/keybed.

Last edited by prout; 02/25/16 04:06 PM.
Re: How to Measure Latency?
newbert #2514804 02/25/16 04:57 PM
Joined: Feb 2010
Posts: 5,531
W
5000 Post Club Member
Offline
5000 Post Club Member
W
Joined: Feb 2010
Posts: 5,531
Originally Posted by prout
I am not convinced that you have found the beginning of the tone of the note played. To find it, I had to measure the period of the peaks and compare that to the expected period of the note I was playing, then I chose the first rise of the peak across the zero reference line. This ensures that the noise of the key moving through letoff and the sound of the key hitting the keybed is not mistaken for the note. I think the area from .112 on in your example above is not a tone, but more noise. See my example above.


It seems to me that the volume at the right of the interval [in my picture] is way beyond normal mechanics noise. The peak on the right is almost the full volume of the sound. Maybe the first ms is just noise of the hammer hitting the string, But whether it's white noise still or a already clear pitch, it does not matter for the latency.

I have never seen weird waves like your recording coming from my piano. It looks like your mic is clipping.

Last edited by wouter79; 02/25/16 05:00 PM.

[Linked Image][Linked Image][Linked Image][Linked Image]
Re: How to Measure Latency?
newbert #2514806 02/25/16 04:58 PM
Joined: Feb 2010
Posts: 5,531
W
5000 Post Club Member
Offline
5000 Post Club Member
W
Joined: Feb 2010
Posts: 5,531
>All energy between the yellow lines is noise from the finger/key/whippen/shank/hammer/keybed.

So you say that your mechanics noise is just as loud as the sound of the piano at its loudest position?


[Linked Image][Linked Image][Linked Image][Linked Image]
Re: How to Measure Latency?
newbert #2514815 02/25/16 05:17 PM
Joined: Jul 2009
Posts: 3,314
S
3000 Post Club Member
Offline
3000 Post Club Member
S
Joined: Jul 2009
Posts: 3,314
I think it would be more precise to take the measurement from the moment the hammer strikes the strings, in which case the signal may well not be pitched at that point - it's just a transient "thud".

Greg.
p.s [irrelevant nonsense] Dire_tonic: I like that "muck in" phrase - a very hearty British phrase. I've been racking my brain trying to figure our where I had heard it recently - turns out I had remembered hearing it in Rowan Atkinson's brilliant "Fatal Beatings" skit.

Re: How to Measure Latency?
wouter79 #2514833 02/25/16 05:52 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
Originally Posted by wouter79
>All energy between the yellow lines is noise from the finger/key/whippen/shank/hammer/keybed.

So you say that your mechanics noise is just as loud as the sound of the piano at its loudest position?


Remember that I placed the mic 10 cm from the key. The lid of the piano was closed and a heavy quilted string cover was laid over the strings. So in this case the noise was definitely as loud as the string tone. I listened to the recordings and can confirm that fact.

Could you post either a .wav file or another picture of what you posted above, but much longer, say 0.5 seconds? That way, we can easily see the point of the first complete tonal waveform emerge from the noise.

Thanks,

prout

Edit: It's hard to see from the waveform but the max noise is about -31db and the max tone is about -14db.

Last edited by prout; 02/25/16 06:51 PM.
Re: How to Measure Latency?
sullivang #2514844 02/25/16 06:42 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
Originally Posted by sullivang
I think it would be more precise to take the measurement from the moment the hammer strikes the strings, in which case the signal may well not be pitched at that point - it's just a transient "thud".


If we can positively identify that moment in both a DP and an AP, sure, that makes sense, but my way gives a positive relativistic measurement. By dampening most of the upper partials in the tone, I was able to clearly pick out the first occurrence of the tonal waveform from the preceding noise.

Presumably the 'thud' is also modeled or sampled and forms part of the sound of the note, (though I would argue one can't 'sample' that 'thud' without including in it the noise of the key bottoming out, and then, you get both the real key and the sample bottoming out around the same time, creating an inaccurate reproduction of piano tone/noise) but, for measurement purposes, they both reach a 'first complete cycle' point in the tone creation, and that is, for me, the easiest point to measure accurately.

Last edited by prout; 02/25/16 06:44 PM.
Re: How to Measure Latency?
newbert #2514886 02/25/16 10:28 PM
Joined: Jul 2009
Posts: 3,314
S
3000 Post Club Member
Offline
3000 Post Club Member
S
Joined: Jul 2009
Posts: 3,314
prout: I just feel a little uneasy about how consistent the delay is between the very first signal (instant of hammer contact), and the emergence of the pitch. Anyway, it's this kind of problem that led me to suggest using a simple, high pitched tone (earlier in this thread), if one wants to measure the raw capability of the DP hardware, separately to the piano. (ideally, a simple square wave with a zero attack time,if the DP happens to have any synth tones). Of course, with the DP, we can simply record the note with no key noise whatsoever, and then see exactly what the delay is from hammer strike to the pitched portion, for any given note & velocity.

Greg.

Re: How to Measure Latency?
sullivang #2514901 02/25/16 11:01 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
Originally Posted by sullivang
prout: I just feel a little uneasy about how consistent the delay is between the very first signal (instant of hammer contact), and the emergence of the pitch. Anyway, it's this kind of problem that led me to suggest using a simple, high pitched tone (earlier in this thread), if one wants to measure the raw capability of the DP hardware, separately to the piano. (ideally, a simple square wave with a zero attack time,if the DP happens to have any synth tones). Of course, with the DP, we can simply record the note with no key noise whatsoever, and then see exactly what the delay is from hammer strike to the pitched portion, for any given note & velocity.

Greg.


You are quite correct that the delay described above is inconsistant, both with regard to hammer velocity and thus string displacement, but also with string mass and hammer mass, as well as hammer resilience.


Re: How to Measure Latency?
newbert #2514961 02/26/16 03:46 AM
Joined: Jul 2011
Posts: 3,134
D
3000 Post Club Member
Offline
3000 Post Club Member
D
Joined: Jul 2011
Posts: 3,134
lest we forget the OP...

Originally Posted by newbert
Although I'm pretty happy with the performance of my gear, I do wonder what my "numbers" are with regard to latency.

Originally Posted by prout

Is speaks when you expect it to speak. Therefore, from the standpoint of your brain and making music, it is instantaneous.

Rest on your laurels and rejoice.

Re: How to Measure Latency?
sullivang #2514967 02/26/16 04:02 AM
Joined: Jul 2011
Posts: 3,134
D
3000 Post Club Member
Offline
3000 Post Club Member
D
Joined: Jul 2011
Posts: 3,134
Originally Posted by sullivang
I like that "muck in" phrase - a very hearty British phrase. I've been racking my brain trying to figure our where I had heard it recently - turns out I had remembered hearing it in Rowan Atkinson's brilliant "Fatal Beatings" skit.

I might have to sit through another repeat of the Blackadder seasons soon. Good medicine.

Re: How to Measure Latency?
prout #2515154 02/26/16 05:12 PM
Joined: Feb 2010
Posts: 5,531
W
5000 Post Club Member
Offline
5000 Post Club Member
W
Joined: Feb 2010
Posts: 5,531
Originally Posted by prout
Originally Posted by wouter79
>All energy between the yellow lines is noise from the finger/key/whippen/shank/hammer/keybed.

So you say that your mechanics noise is just as loud as the sound of the piano at its loudest position?


Remember that I placed the mic 10 cm from the key. The lid of the piano was closed and a heavy quilted string cover was laid over the strings. So in this case the noise was definitely as loud as the string tone. I listened to the recordings and can confirm that fact.



You just want to measure the time the thud of the hammer hitting the string happens? But if the string is damped, why are you looking for "tonal waveform"? I expect just a thud with no good waveform?

Originally Posted by prout

Could you post either a .wav file or another picture of what you posted above, but much longer, say 0.5 seconds? That way, we can easily see the point of the first complete tonal waveform emerge from the noise.

Thanks,

prout

Edit: It's hard to see from the waveform but the max noise is about -31db and the max tone is about -14db.


I can post a bit longer but I threw away the file. Maybe I can record again tomorrow.

Again, the point where the "first complete tonal waveform emerge" seems irrelevant. The tone probably starts with the hammer crashing into the string and I expect mostly NOISE at that moment, not a "tonal waveform". But maybe it depends on how hard you strike.


[Linked Image][Linked Image][Linked Image][Linked Image]
Re: How to Measure Latency?
wouter79 #2515160 02/26/16 05:37 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
Originally Posted by wouter79
Originally Posted by prout
Originally Posted by wouter79
>All energy between the yellow lines is noise from the finger/key/whippen/shank/hammer/keybed.

So you say that your mechanics noise is just as loud as the sound of the piano at its loudest position?


Remember that I placed the mic 10 cm from the key. The lid of the piano was closed and a heavy quilted string cover was laid over the strings. So in this case the noise was definitely as loud as the string tone. I listened to the recordings and can confirm that fact.



You just want to measure the time the thud of the hammer hitting the string happens? But if the string is damped, why are you looking for "tonal waveform"? I expect just a thud with no good waveform?

Originally Posted by prout

Could you post either a .wav file or another picture of what you posted above, but much longer, say 0.5 seconds? That way, we can easily see the point of the first complete tonal waveform emerge from the noise.

Thanks,

prout

Edit: It's hard to see from the waveform but the max noise is about -31db and the max tone is about -14db.


I can post a bit longer but I threw away the file. Maybe I can record again tomorrow.

Again, the point where the "first complete tonal waveform emerge" seems irrelevant. The tone probably starts with the hammer crashing into the string and I expect mostly NOISE at that moment, not a "tonal waveform". But maybe it depends on how hard you strike.


I am interested in the methodology of measuring events. In the examples I posted above, I can positively identify the start of the keystrike by my fingernail and the zero crossing point of the first complete tonal waveform. It would be much more difficult to aurally measure the point of hammer/string contact. The problem with hammers is that, if properly voiced, they remain in contact with the string (and dampen it) when played softly and bounce off much more quickly when played loudly. This causes a dampening of the upper partials for soft playing and increased partial structure and richness for loud playing.

I use the word 'tone' to mean a measurable pitch and that pitched based tone cannot be perceived until the string has completed one complete cycle of at least one partial in the waveform. That is why I damped my A3 so that only the first and second partials would be obvious.

If you can provide me with a different type of sound based sample I would love to analyse it.

Thanks.

Edit: I was reading today about AP tone production and timing. When played very loudly, the hammer hits the string long (+20ms) before the keybed is hit by the key. This means that there is, relative to the keybed, a negative latency. The DP cannot reproduce this effect, unless some company has found a way to predict future events.

Last edited by prout; 02/26/16 05:40 PM.
Re: How to Measure Latency?
prout #2515171 02/26/16 06:11 PM
Joined: Jul 2009
Posts: 3,314
S
3000 Post Club Member
Offline
3000 Post Club Member
S
Joined: Jul 2009
Posts: 3,314
Originally Posted by prout


Edit: I was reading today about AP tone production and timing. When played very loudly, the hammer hits the string long (+20ms) before the keybed is hit by the key. This means that there is, relative to the keybed, a negative latency. The DP cannot reproduce this effect, unless some company has found a way to predict future events.


That's interesting - I thought it was just the opposite, with the reason being that the shank flex delays the hammer. Are you sure about this? Do you have a reference?

Here's my reference smile http://www.forum-pianoteq.com/viewtopic.php?pid=8226#p8226
Quote:
Quote
When one attempts to play fortissimo, the hammer of a real grand piano typically contacts the string AFTER the key front has bottomed out in the keybed, because the high force of the action upon the hammer shank has caused the shank to flex (e.g., the hammer itself remains stationary by its own inertia for a minute fraction of a second AFTER the hammer shank has started to move). In addition, bass note hammers have more inertial mass than upper octave hammers. Typically, this difference in timing -- delay, if you will -- is on the order of about 10ms difference (one-one hundredth of a second).


Greg.

Last edited by sullivang; 02/26/16 06:13 PM.
Re: How to Measure Latency?
sullivang #2515178 02/26/16 06:19 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
I dumped the source, but it is on my iPad History. I'll try to find it.

While I am trying to find it, remember that the hammer acceleration is related to the action ratio which is around 5.4:1 on modern pianos, so that last 1.5mm motion of the key at letoff to keybed and the 1.6mm hammer to string travel occur at different speeds in that ratio. The hammer flex will have an effect on that timing.

OK, I found, I think, the source and I was wrong. I got it reversed. It states that the hammer hits the string before the key hits the keybed in soft playing.

Last edited by prout; 02/26/16 06:44 PM.
Re: How to Measure Latency?
prout #2515183 02/26/16 06:48 PM
Joined: Jul 2009
Posts: 3,314
S
3000 Post Club Member
Offline
3000 Post Club Member
S
Joined: Jul 2009
Posts: 3,314
Originally Posted by prout

OK, I found, I think, the source and I was wrong. I got it reversed. It states that the hammer hits the string before the key hits the keybed in soft playing.


Ah, good. That certainly tallies with that Pianoteq forum post I referred to. Thanks for checking.

Greg.

Re: How to Measure Latency?
newbert #2515204 02/26/16 07:35 PM
Joined: Nov 2013
Posts: 4,468
P
4000 Post Club Member
Offline
4000 Post Club Member
P
Joined: Nov 2013
Posts: 4,468
You might want to read this paper. "Touch and temporal behaviour of grand piano actions"

https://iwk.mdw.ac.at/goebl/papers/Goebl-Bresin-Galembo_JASA2005_PianoAction.pdf

[Linked Image]

For pianists, this shows that the struck touch has a much faster response time than the pressed touch, but with similar sound levels. Interesting. Note that the struck touch hammer hits the string and the key hits the keybed at the same time, whereas the pressed key hammer strike occurs some 3ms before the hammer strike.





Re: How to Measure Latency?
prout #2515321 02/27/16 06:59 AM
Joined: Nov 2010
Posts: 5,879
A
5000 Post Club Member
Offline
5000 Post Club Member
A
Joined: Nov 2010
Posts: 5,879
Originally Posted by prout

Edit: I was reading today about AP tone production and timing. When played very loudly, the hammer hits the string long (+20ms) before the keybed is hit by the key. This means that there is, relative to the keybed, a negative latency. The DP cannot reproduce this effect, unless some company has found a way to predict future events.


Actually there are a few DP actions that have escapement in the action which allows the hammer and the key to be moving in different directions. It's not all key sensors out there. I certainly agree that there needs to be a hammer sensor to have it behave anything like an AP. I believe the Kawai GFI and GFII, the Yamaha AG and the Casio Grand Hybrid do it this way - and a few others but I can't remember which ones so I won't name them here.

Re: How to Measure Latency?
ando #2515408 02/27/16 02:50 PM
Joined: Nov 2009
Posts: 3,608
S
3000 Post Club Member
Offline
3000 Post Club Member
S
Joined: Nov 2009
Posts: 3,608
Originally Posted by ando
Originally Posted by prout

Edit: I was reading today about AP tone production and timing. When played very loudly, the hammer hits the string long (+20ms) before the keybed is hit by the key. This means that there is, relative to the keybed, a negative latency. The DP cannot reproduce this effect, unless some company has found a way to predict future events.


Actually there are a few DP actions that have escapement in the action which allows the hammer and the key to be moving in different directions. It's not all key sensors out there. I certainly agree that there needs to be a hammer sensor to have it behave anything like an AP. I believe the Kawai GFI and GFII, the Yamaha AG and the Casio Grand Hybrid do it this way - and a few others but I can't remember which ones so I won't name them here.

I know we obsess over these details on this forum, but don't believe what you say about the Kawai GF actions is correct.

Page 4 of 5 1 2 3 4 5

Moderated by  Piano World 

Link Copied to Clipboard
(ad)
Pianoteq
PianoTeq Karsten Collection
(ad)
PianoDisc

PianoDisc
(ad)
Piano Life Saver - Dampp Chaser
Dampp Chaser Piano Life Saver
(ad)
Mason & Hamlin Pianos
What's Hot!!
News from the Piano World
Our October 2020 Free Piano Newsletter is Here!
---------------------
3,000,000+!
------------------
Posting Pictures on the Forums
-------------------
Forums RULES & HELP
-------------------
ADVERTISE on Piano World
New Topics - Multiple Forums
Piano Safari Christmas collection
by ebonykawai - 10/19/20 08:59 PM
Another "beginner" piece w/twist
by MilesAbbott - 10/19/20 08:58 PM
Sostenuto Pedal....in an upright?
by Duaner - 10/19/20 06:38 PM
Piano vs keyboard
by GHN - 10/19/20 06:22 PM
Download Sheet Music
Virtual Sheet Music - Classical Sheet Music Downloads
Forum Statistics
Forums41
Topics202,314
Posts3,015,317
Members98,949
Most Online15,252
Mar 21st, 2010
Please Support Our Advertisers


Faust Harrison 100+ Steinways

Dampp Chaser Piano Life Saver

 Best of Piano Buyer

PianoTeq Bechstein
Visit our online store for gifts for music lovers

Virtual Sheet Music - Classical Sheet Music Downloads



 
Help keep the forums up and running with a donation, any amount is appreciated!
Or by becoming a Subscribing member! Thank-you.
Donate   Subscribe
 
Our Piano Related Classified Ads
| Dealers | Tuners | Lessons | Movers | Restorations | Pianos For Sale | Sell Your Piano |

Advertise on Piano World
| Subscribe | Piano World | PianoSupplies.com | Advertise on Piano World |
| |Contact | Privacy | Legal | About Us | Site Map | Free Newsletter |


copyright 1997 - 2020 Piano World ® all rights reserved
No part of this site may be reproduced without prior written permission
Powered by UBB.threads™ PHP Forum Software 7.7.4