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

Steve, the noise floor is really high (-45dB) so it isn't the best in terms of listening, but I can see things like looping and stretching. Should I proceed, or do you want to give it another go?

Ah, I think I've got the line in volume too high on the Mac, as I had to turn the volume on the P-80 quite low to stop any clipping. I'll swap them round, and try again!

I think I need to officially label the DPBSD test a "song" so people will stop saying this.

In the DPBSD "song" played on the CP1, I can clearly hear looping, stretching, and perhaps a velocity layer switch. With poor sympathetic resonance on top. If it can happen in my "song" it can possibly happen in other, real songs.

1990 called. They want their lame sample compression techniques back.

Ok hopefully this is a bit better:
https://docs.google.com/leaf?id=0Bw...DRiLTgxMDAtODNjMjUzMjllMWEz&hl=en_GB

Much better! Noise floor @ -57dB (lowest theoretical is approx. 16 bits x 6 dB/bit = -96dB). What's your setup, and which patch is this?

Thanks dewster for all your work with the analysis.

I do not see the problem with looping though. To me it seems to be imitating the decay of a real piano good enough. Can you make a comparable graph with your Young Chang playing the same note?

Excellent idea. Let's get some acoustic pianos into the comparison.

Great, that was with the default "Grand Piano", and now I realise it would also have had (my) default reverb setting of "Hall 2", which I should have turned off.

I'll do another one tonight, with no reverb!

I recorded it via the L/R RCA outputs to the line in on my MacBook, using Audacity, and its built in mp3 export (which uses Lame)

Dewster, thanks for your explanations, these layers and stretching are surprisingly very clear to notice.

And real grand piano would be great for comparison.

@dewster,

finally I've made it to record my KAWAI CA-63 (sound is also identical to KAWAI CA-93).

You'll find the file here:

http://www.mediafire.com/file/wnyzmd2hyqw/KAWAI_CA_63.mp3

Thanks in advance for you analysis!

Warning! Anyone who believes technical analysis of DP output is meaningless, or believes that I believe the DPBSD is the ultimate test of a DP, or believes the analysis files are fake or doctored in some way, or for whatever reason can't handle the truth, is strongly cautioned to avoid reading this post.


I don't have a laptop anymore so it's difficult at this point for me to record our piano upstairs. Fortunately, I do have the VintAudio sample of a real C7 handy, so I'll use that for comparison purposes here.

To make the spectral plots clearer, I applied a 10:1 compressor to the notes, then normalized the peak to -1 dB. Any "hash" you see towards the end of the note decay is due to degrading SNR (rise of the noise floor with decreasing signal due to the applied heavy compression). I do wish Audition had some way to make the spectral graphs clearer without doing this, as it would make my job of analysis much easier.


VintAudio spectral phase view of C4 note.


CP1 spectral phase view of C4 note. Cursor at looping point.


VintAudio spectral pan view of C5 note.


CP1 spectral pan view of C5 note. Cursor at looping point.

If you look closely at the time axis in the displays you will notice that they are very close to the same for both the VintAudio and CP1 note samples. So you can easily see that the CP1 decay is noticeably faster, though presumably you can adjust that. It would probably make the looping more audible though. As I've said before, I've come to see fast decay as a way to hide looping, and therefore see it as a negative thing. Unnaturally fast decay also sounds fake to my ears.

My we're getting touchy aren't we!? For the record I don't think your tests are meaningless I just don't think it's a good way to judge a musical instrument.

Interesting to see the differences between DPs and APs though!

Can you show the same thing for Pianoteq? I'd be fascinated to see how a fully modelled piano looks.

You can get slightly better results recording in Garage Band using 24-bit input and export modes. And then the MP3 encoding can use the Fraunhofer encoder which (as much as a like open sourse) is much better than Lame. Lame has to work around to many of Fraunhofer's patents but Apple includes the the Fraunhofer encoder so take advantage of it.

These tests are very interesting, Dewster, but not as compelling as Lawrence's and Dr.Popper's reviews of the CP-1.

So significant that I ordered a CP-1 this morning, as did one of my students.

I think Yamaha has a winner on it's hands with this instrument; I know I trust Lawrence's(Melodialworks) and Dr.Popper's real world, hands on experiences far more than putting any real confidence in a bunch of tests.

These guys are pros in every sense of the word.

All in all, it still comes down to what is important in an instrument, and it's much easier, and far safer, to take the word of respected and talented players, than it is to look at squiggly lines on a screen.

Soul can't be analysed with a computer.

Best regards,

Colleen

PS. My thanks to Lawrence and Dr.Popper, and to you as well, Dewster.

It's called "overcompensation"...and you're right, the tests aren't meaningless...a little tedious, maybe, but ultimately the instrument, and the DP is an instrument in it's own right, will have a hit or miss based on how it feels, sounds, and how it responds to player input.

I like Colleen's remark..."Soul can't be analysed with a computer."

That about sums it up for me too.

Snazzy

Colleen,

You've ordered a CP1 because Dr Popper and Lawrence like it? Would I be right to assume you haven't played one yourself? If so I really hope it works out for you...The CP1 does sound very very good from what little is available at the moment but even so, to buy just on the recommendation of others is a big leap of faith.

When do you expect to get it? Let us know what you think when it comes.

Best wishes,

Steve

I have uploaded a recording of Roland HP-307. The input level wasn't great, I hope normalization didn't screw anything.
http://www.mediafire.com/?2mwkmyhrzny

Thanks dewster!

Now I know the difference and understand a little more about your talk about loooping. The real piano seems to have a very complex decay compared to the digital one.

You have to be very clear and basic about the technical parts, because spectral analysis is not the usual occupation for most of us.

The more I read from several people like Colleen_500 here how deficient, inaccurate and even biased and manipulative this test appears to them (especially those who like the CP-1) this all evokes an uneasy feeling in me. There is no need for test bashing at all! We all have now understood your reservations. Let's move on! Colleen_500, it's great to hear that you are now ordering a CP-1 just because of the word of someone else you trust. I would never buy a DP without playing it for an extended period of time. Never.

I think we all have understood that you don't trust this test, so please... let us techie geeks just compare the test results to see what (maybe not completely accurate) technology, sampling length etc. lies behind the different models, and you can discuss the (never heard it but I trust you) great sounding quality of CP-1 and others at another thread. I would of course then join you there, without even mentioning this thread...

Hi Steve,

Yes I have ordered one; actually, the store is bringing in two anyway. I'm not committed to buy, and neither is my student.

I have correspondence with several other users who aren't on this forum, and they confirm what Lawrence and Dr. Popper have been saying as well.

These players are also professionals.

I haven't heard one bad comment; not one.

Remember, "Piano World" is not the end all be all for finding out an instrument's performance or potential.

So, it is not as much as a "leap of faith", but rather, I would consider it as a much more informed, and confident, move.

Have you played one yet?

The instruments are back-ordered; I suspect it's because they are quite popular.

Best regards,

Colleen