Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards The DPBSD Project!

Hey ChrisA, don't guess if you don't know. I made the recording, and I boosted the recording I think by more than 12dB because the recording was relatively low, because the recording needs some headroom to be sure that no distortion occurs. It's not 12-bits.

ChrisA, what's the math behind that? (apart from the obvious but strange to me "multiply the bits by 6 to obtain noise floor in dB" )

The exact number is a bit more involved than that but 6 dB per bit is pretty much what it comes down to.

If you think about it, 6 dB is a doubling of the range, and each bit is a doubling of the range as well.

It's an entirely digital process:

Tone generator (digital) --> MP3/WAV codec (digital) --> USB memory (digital).

Cheers,
James
x

KAWAI James, thanks for the info! The price of CA63 from the Bulgarian distributor is excellent and that piano is so tempting to me! I am about to sell my RD-700SX this evening and if everything goes well, I will order the CA63 pretty soon. The DPBSD project had a great impact on that decision, as well as demo files made by Kawaian in few threads. The only problem is that I missed the opportunity to try the model myself since they sold the demo model a week ago And they can only order one for me, but I have already played MP8 II once and I remember I loved the wooden action so I expect no surprises at all.

In the p-155 result test is missing the last 2 stretching groups. Total = 30

The Key-off sample is not a "clunk" sound or percussive. Instead they put a sample that recreates a subtle kind of harmonic sound. And it's quite good!

That key-off noises they put in the DPs software is often quite exaggerated.

You're right, the upper group of 10 is actually 4,3,3 when you zoom up on it.


There is a fairly nice mute sound at the end of the unpedaled notes in the partial pedaling test. It's not buzzy like felt hitting a vibrating string, but it isn't badly done. I was just noting that there were no explicit knocks or "loom of strings" sounds associated with key-up and pedal up/down.

I've updated the review in the text file. Thanks!

dewster,

I actually just discovered this thread, 'DPBSD' must've triggered some "This can't be interesting" opinion in my mind

This is fantastic work. I haven't had a chance to look over all of it yet, but what I've seen is incredible.

Can I help with my Roland HP-203? I have Access to a Mac to record Audio, and I do have a USB stick that I store my songs on, no idea if the piano does it in standard MIDI or proprietary though...

Important note!!

The P155 has the sustain resonance function and the damper resonance function. The damper resonance is the old “echoy” effect referred in the BSD_DP test and is the same that exists in the old P80 / P120. Fortunately the P155 has the sweat option for turning it off! On the other hand there is the "sustain resonance" effect for simulating the "all strings damped" resonating to the soundboard.

The p155 and my old p80 have similar results in this test but there was something in the P-80 that I couldn’t live with: - the absence of the sound of hammer striking in the mid and upper notes! The P155 nails this remarkably! The decay is much better also.

Yes absolutely!


From the HP-203 manual, it looks like it will play SMF or Standard MIDI Files, so that should work from the USB stick. It doesn't look like it will record WAV or MP3 to the stick though, so you will have to hook the line-out to your Mac. Let me know if I can help in any way, there are instructions in the readme file at the Share Point.

Hello Dewster,

here is the sample of Bechstein 280 from EWQL Piano. It's recorded in Cubase 5 and converted to MP3 in Wavelab 6. EWQL Play engine is last 1.2.5.3 and it's full master patch without dynamics (linear velocity curve). All settings are default only ambience is turned off. Also all level settings are default so it could be maybe too week (wavelab shows level about 50%).

So please let me know if this setup is fine or I should do some changes. When it will be O.K. I will make samples of other EWQL pianos (Bosendorfer 290, Steinway D and Yamaha C7).

http://www.mediafire.com/?qzzzgrjzy2d

cheers, Petr

Thanks! That's the East West Quantum Leap pianos I assume? Quite a bit of sampling from what they say on that page!

The piano output level could be set higher by about 9 dB to get the peaks near clipping and provide for more dynamic range. I think they keep the default level rather low to keep you from clipping when playing power chords and such.

I'm not hearing key up or pedal noises (perhaps you could turn these up to make them audible?) and there seems to be no response to partial pedaling. I'm seeing no stretching, and 8 velocity layers with very abrupt and very audible switching.

I'm hearing some really odd rumbling, which I first thought was trucks rolling by outside our house. I first noticed it in the decay of C3, the second note in the looping test - here is a sample (highly compressed to make it more easily heard). Any idea on what that is? I wonder if it's in the samples themselves (air conditioning noises) or was caused by the MP3 conversion, etc.

Thanks for commnets, I'll make some changes and send another sample soon. Levels of articulations are adjustable so I will increase release articulation (it could be something like the release sample in Ivory). There is no artiuculation or switch for pedal noise.
There si no stretching in Full patches. In Lite patches there should be stretching over 3 keys. Layer switching is realy horrible and I hope it will be improved in Play 2.0 engine. There is no smooth cross fading but sometimes realy big jumps in timbre.
About the rumbling - it sounds to me like some kind of air condition in background. There are also very audible timbre differences between keys played on the same velocity level but maybe it could be issue of real instrument.
I'm realy looking forward to Ivory II

So here is the new version:

http://www.mediafire.com/?zn1z04nyzzz

Release articulation is now +6 dB (see picture). Output level should be fine too (these few tones in the beginning are great for adjusting).

Now I've used different MP3 codec but I think it's something in their samples. When I compare to Ivory then EWQL has warmer samples with thicker middle. Ivory samples are more neutral to me.

Yes, the layer switching leaves a lot to be desired in the EWQLP.
The specs say "10 to 18" velocity levels, but it just doesn't seem to live up to that. In the Steinway, there's one note with just 6 samples in it's respective sample folder, so I assume that note only has 6 velocity layers. I really do wonder about their specs. If they've supplied all the layers they actually sampled, I guess it will be a big job to fix this properly - they'd have to do some more sampling. If they're not going to fix it, I'm toying with the idea of using the AET morphing in Kontakt 4 to create new samples. :P :P (I don't have Kontakt 4 yet and I have no idea how well this would work. I have seen positive comments about the AET though) This would of course require EW to convert the samples into the Play format, or to give me a utility to do this. I suspect/hope this will be possible in the Pro version in any case.

And yes, I've heard various artifacts - aside from the rumbling, I've heard a faint "crackle". I like to think it's a crackling valve in their exotic valve mic pre-amps.

I still think it's a very good product despite it's flaws. When I play it normally, at normal levels, I rarely hear any artifacts. The velocity switching is a real problem though. I also think the bass notes are often too quiet - maybe EQ will help but I think the volume just needs to be increased a bit in the playback engine.

Greg.

(re: fixing the layer switching, in all likelihood I'd probably just do a few notes as an experiment - there's no way I'd waste months fixing entire instruments this way)

Greg.

Yes Greg, you are right, some fixings and improvements of EWQLP are necessary. But I'm sure they'll do that to stay competitive. I also like their pianos due they warm full sound. On the other hand after while they make me tired so for longer practising I preffer Ivory's Italian Grand.

After lot of experiments, tests and internet searches I found that there is no "one size fits all" product. Software pianos satisfy my sound tastes (which are variable) better so I don't need to replace whole instrument but only download different software. With low latency sound setup it has same instrument response like DPs. But of course for live musicans it's better to have "one box" instrument.

So I realy like this Dewster thread because it's look behind marketing hypes to technical reality. DPs are computer technology so it's great that there is somebody who's got unemotional engineer point of view. One picture from spectral analyzer says more than tons of subjective feelings. But of course no analyze or review will tell you whether you'll like this instrument or not - everybody must test it first (for a while). But it's generaly known...

Petr

Petr, I sent you a PM.

EastWest Quantum Leap Bechstein D-280 Concert Grand Review

Petr (pesk) has kindly submitted a DPBSD MP3 file for the EastWest Quantum Leap Bechstein D-280 Concert Grand Piano for our listening and analyzing pleasure - thanks Petr!

Being a fully sampled instrument, there is no looping or stretching, just nice long decays, particularly for the lower notes, along with beautiful pedal down sympathetic resonance. While it has many velocity layers (I count 9 @ velocity 63, pedal up) layer transitions are unfortunately quite audibly abrupt and pronounced.

There seems to be quite a bit of timbre variation between adjacent notes - listening to the stretch test (chromatic walk up the keys) it almost sounds like the notes are picked from random recording positions, some are bright, some dull, with no particular pattern to it.

There are extraneous noises going on in the recordings, I could hear a rumbling air conditioner kick in, a police siren, and some pops and clicks from things moving around in the recording space. Granted these things are going on near the noise floor, but it is unusual to hear these things in a PC sampler, and if nothing else they tend to raise the noise floor.

I once owned the EW PMI Bosendorfer 290 and had all kinds of trouble recording with it until I finally just gave up. The noise floor was so high it noticeably pumped around during playback, and one sample had a loud extraneous noise in it. The velocity layer switching was so abrupt I had to avoid very low and very high velocities in any MIDI files I fed it. This one doesn't seem much better. EastWest could really use better quality control over their recording, and they need to post process their samples for noise.

The MP3 and more pics are up at the share point.


Spectral frequency view of the layer test - layer transitions are quite visible.

-------------------------------------------------------
- EastWest Quantum Leap Bechstein D-280 Concert Grand -
-------------------------------------------------------
FILE & SETUP:
- dbpsd_v1.6_ewql_bechstein_d-280.mp3
- Recorded in Cubase 5 and converted to MP3 in Wavelab 6, EWQL Play engine 1.2.5.3
- All settings default except ambiance off, release articulation +6 dB.
- Recorded by "pesk".
PROS:
- Passes the pedal down sympathetic resonance test.
- Realistic "buzzy" string damp sound.
- Beautiful long note decay.
- No looping.
- No stretching.
- 9 velocity layers visible and audible.
- Velocity switch @ vel=10,22,38,56,74,92,104,116
- Nice sympathetic resonance.
- Large dynamic range (47dB, vel=1:127).
CONS:
- Fails the key down sympathetic resonance test.
- Fails the pedal down silent replay test @ pedal up.
- Fails the brief pedal partial damping test.
- Fails the partial pedaling test - 75% & 88% pedaled note resonance ends abruptly!
- High notes decay rather quickly.
- Many samples have low frequency noise (air conditioner?).
- C3 & C6 samples have a "clack" sound.
- C4 sample has a police siren.
- C4 key-up sounds like a faint note playing.
- No pedal up/down sounds.
- Variation in sound (muffled/bright) between adjacent notes is unusually pronounced.
- Velocity layer switches quite visible and unusually audible.
- No key-up sound if damper pedal down?
OTHER:
- MP3 levels: peak @ -0.11dB, noise floor @ -90dB.
- Date reviewed: 2010-04-17.