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

I'm curious as to how we know whether long decays/sample lengths are by default better than shorter ones? What if that's all there is to it at the mid and high and long decays and sample lengths are not necessary or natural there? This is not specific to the CP1 or Yamaha or any other makes. Just wondering over all because if the pattern is there between mid-level or high-level DPs then maybe it's because that's all there is to it and 99% of what's important to hear is already captured, and it's not worth the diminishing return to try to overkill at the expense of more expensive resources to capture the remaining 1%.

One thing. The dp_bsd file is based on the default settings for the #1 AP patch. Perhaps tweaking can adjust the sympathetic resonance somewhat. I haven't worked with it enough at this point to know for sure.

The specs, as revealed to an extent by the dp_bsd test only tell part of the story.

For me, the bottom line is how does it sound, and how does it play? Does it inspire me as a musical instrument?

The advantage of DP's is the sound / action / gesture connection. The disadvantage is that there are compromises made to the sound itself.

The advantage of sample sets is that there are not (usually) compromises made - since size doesn't matter. The disadvantage is that the sound / action / gesture connection totally sucks.

So, with samples, you can get a better sound, but only with post recording editing.

I don't want to do the editing because (a) it isn't fun, and (b) you can easily kill the passion in the music. So, at this point I'm opting for the DP option. (Although God knows I've tried to make samples work, and have purchased so many of them, it is rather sickening.)

So, I'm quite happy with the CP1 sound and response, and not actually all that concerned about the specs! (Just as I'm happy with how my Honda Accord V6 rides and drives. I don't actually need to know or care about how the magic is achieved! Perhaps, though, if I was driving a Toyota, my opinion might be different).

Lawrence

Good question. One of the parameters that can be adjusted is DECAY. The default setting is 0, and the range is from -16 to + 16. I wonder if adjusting this would make Dewster happier? (Probably not!)

Interesting. So the CP1 is according to these tests entirely ordinary and with the acoustic piano sound Yamaha breaks no new ground at all. On a subjective level though it seems to have been very well received. I'm very impressed with the few bits I've heard.

Dewster's test does make you wonder what you are getting for four and a half grand (£) though!

Steve

The test is not taking into account aspects of the CP1 AP that are modelled, for example hammer hardness. Also, the approach to dynamics is new (according to Yamaha) and the proof is in the playing. It has a remarkable range and smoothness. Play it, and you'll be left wondering what the fuss is about, in terms of the DP BSD results. Adjusting the hammers on this instrument is amazing, and transforms the pianos! (And that is only one aspect that has been modelled).

The DP BSD is not the be all and end all - especially for an instrument that combines samples and modeling. (And also, let's not forget that the interpretation of the test results is to a certain extent subjective - and being made by one individual).

You have a point, long decays on very high notes are often unnatural - particularly if the decay is a poorly done loop! But short decay is almost always used to hide looping, so I have come to look upon it as a negative. I want a DP I can run through a compressor and not wince at the output.

BTW, the first test in the DPBSD is "calibrated" to the note decay times of Pianoteq, just to have a rough (though admittedly very relative) reference point.


That strikes me as something of an apology, particularly when you have PC samplers and modelers absolutely killing them in the sound department. I think it's more that all of the manufacturers keep doing what they've always done, with slight improvements now and then, and people keep buying them. Nothing inherently wrong with that, and I'm probably expecting too much from the market mechanism.

As Lawrence notes, the timing between the playing and the sound is very important, as is the feel of the keyboard, and I think they've been concentrating on those aspects more than the sound (for my tastes anyway).

While the sample length is fairly typical, the looping is very well done (which I think is 70% of it), the velocity blend is smooth (15%), and the stretching is done over only the low and high ends, with the mids completely sampled (15%).

I think they've done all they can with a fairly limited sample memory size (though I would improve the sympathetic resonance and include some key-up and pedal up/down noises).

My (eternal) question is this: why go to all that trouble to shoehorn the sound into such a small space when ROM is so inexpensive? Surely the effort they put into doing this must be expensive (unless it is largely automated, which is possible, but surely not 100%). Can $20 (retail, mind you) of Flash be that big a deal in a $5k instrument? Particularly when it could so profoundly improve it AND reduce engineering NRE? I must be missing some important factor somewhere...

And, don't get me wrong, the CP1 sound pretty fantastic in the recordings I've heard. I'm not here to crap on it, only to reveal some of the technical underpinnings of its sound.

For me, though, I need either a bigger sample or something modeled that doesn't sound too fake. The beat goes on.

People are saying the CP1 is much better sounding then the P155. I assume there is "next generation" level of improvement. But according to your analysis abut the only difference is that he CP1 uses less note stretching. I think there must be something more that is different. "Simple" things like how the model piano has mic'd or the size of the recording room or the brand of microphone could make a huge difference. So a question: What kind of test could capture this?

One other thing, You seem to not hear the key off samples. I know the P155 and other DPs tested have key off. (You should hear the key-off sample on the harpsichord voice, it is unmistakeable from across the room even with TV noise going on.) The P155 default setting it low but I can crank up the volume on the key off sample but that is not the way to go. I think, maybe your test is insensitive to key off.

Honestly, for me it's the most valuable topic on DP forum so far. Thanks for your work!! Great job.

The looping is much better, and I think that is the main difference, actually.


I'm more interested in what they're technically doing in DPs that correlates to what I hear. I'll be pickier about the actual sound when I know they're devoting enough memory / computational resources to it. Until then, for me anyway, it's kind of a cruel joke.


The key-off sample on the P-120/140/155 harpsichord is really wonderful - it completely adds to the realism, and my hat is sincerely off to Yamaha for doing that patch so well.

For the piano however, what I'm not hearing in the second phase of the DPBSD sympathetic resonance test is where 11 keys lift simultaneously. Pianoteq clearly makes a key-up sound there, as does a real piano, and several of the better PC samplers. Surely that can't be a fault of my test.

Yes and yes!

I've said from the start that is more of a go / no-go test for technical issues mainly associated with samplers, not a test-of-all-tests. That said, you can clearly hear many subjective things in the results if you listen for them.

I'm hoping others will listen to and analyze the MP3 files openly, so there can be more of a consensus that others can trust. I feel a bit naked being the only one who posts reviews here.

Thanks to Lawrence for providing the CP1 CFIII MP3 - we've all been waiting for that, I guess!

I'm with dewster here, I don't really understand how they can still be using stretching and looping. I mean, they seem to (or at least claim to) have gone the full mile with the Rhodes and DX7 sounds. And they claim they've used the "best available technology" for every sound type. Surely using stretching and looping is not the "best available technology", so I'm disappointed. Ok, so maybe they did the looping really well, and maybe it plays just fine. But I had really hoped the CP1/CP5 would be a revolutionary step up, where they've pulled out all the stops...

-------

Now I may be gready, but @Lawrence, could you also do the S6 and one of the (reportedly fully modeled) Rhodes? I'm aware that the Rhodes is out of topic, but I'd love to see what they really did with the Rhodes and how it affects the BSD results. Thanks much!

Yes, because that's the focus of the instrument.

I don't expect Yamaha's latest sound technology to be fully exploited for reproducing acoustic piano sounds until the next series of Clavinova models are launched...because that's the focus of those instruments.

Cheers,
James
x

James, this is an interesting point and it was brought up during the discussion about the apparently non-graded action of the CP1 and 5.

I suppose the potential buyer or any interested party can pick up a vibe from what the manufacturer gets out into the public domain. I'm not saying you're wrong but personally I don't get this idea that the acoustic pianos are intended to be secondary to the electro-acoustics and DX-7s. I did with the Nord Electro. I think Clavia were quite explicit, initially calling the AP patches a "bonus".

It's just my impression but I think Yamaha would want us to believe the APs are intrinsic to the purpose of the instrument. I feel the only change of focus is that they have now given EP sounds equal status, which is a first for a Yamaha DP. It would be useful to get an official steer from Yamaha on this. If you are right it would maybe make some sense given the action and the performance in the DPBS test.

Cheers,

Steve

Steve, you're right, this topic has been discussed at considerable length in the main CP1/CP5/CP50 thread - my apologies, it wasn't my intention to bring it up here.

However, I get the impression that the piano sounds in the CP1, while no doubt excellent, are intended for gigging musicians, not classical pianists.

We did have a couple of Yamaha folks posting here not so long ago, so I expect if they are following this thread (which they surely must be...), they will clarify this point.

Cheers,
James
x

Not sure if I agree with that:

- in the list of patches, APs come first, EPs afterwards
- every CP1/5 material I've seen mentions APs first

Maybe the CP1/5 development team had a different focus, and the marketing department didn't notice, or didn't agree? Of course that's possible. But even then, I still don't really understand the developers. They've reduced the amount of stretching in the CP1/5. Why reducing it? Why not dropping it altogether? Being a (software) developer myself, that doesn't make *ANY* sense to me whatsoever. Obviously they've consciously decided to keep using stretching, but why? Just to save a few pennies (or dollars)? Is that the right approach when talking about an instrument described as:

"The development of the new CP series started with the no-compromise CP1"

Looping and stretching are compromises, based on yesterday's hardware limitations which have long since gone.

I think there's a tendency on this forum to get overly analytical, how much stretching, looping and how many velocity layers are used are ultimately irrelevant. What really matters is how an instrument sounds and all reports I've heard say the Cp1/5/50 sound fantastic!

You can always "review" something based on technical facts and based on subjective impression. I believe both types of reviews are important and have their place.


"Sound fanstatic" compared to what? Compared to other DPs (all of which are also looping)?

The point is: Even if the CP1/5 "sound fantastic", they would sound even better if neither stretching nor looping was used. Nobody in his right mind could refute that. The only question is: How big would the difference be? Maybe it would be small enough to not matter. But who knows...

Is it not disappointing to you that the "no-compromise" advertized CP1 uses significant technical compromises for AP reproduction which could have been avoided rather easily?

madshi, keep it up! You are very eloquently saying everything I've been thinking.

PC pianos have run half-way around the world while DPs are putting their shoes on.

Our next DP may be a Roland. They claim the SuperNATURAL piano has no looping, stretching, or velocity stepping, and Roland is pretty up-front about the size of the bulk sample memory. Now if I could only get my hands on a DPBSD sample of it...