FYI, the East West Quantum Leap Pianos software already has repetition samples (multiple samples of the same note played rapidly). I realise this is not a DP though. Also, as I've said before, just about every digital piano will overlap voices (to the best of my knowledge) for repeated notes with the sustain pedal down, producing quite a realistic subtle variation in tone for each strike. The Lounge Lizard electric piano software does NOT use overlapping voices for this, it seems, and early versions had a deficiency where sometimes repeated notes with the sustain pedal down would be extremely loud. I.e - it seems that they actually model the interaction of the hammer with tines & reeds, taking into account the phase of the reeds/tines at the instant the hammer strikes. Very cool. (well, it is now that it's working better ;^)
EDIT: Major oops - I've just fired up Lounge Lizard and it seems that they have now completely disabled the repeated note with sustain pedal variation in timbre - every strike sounds EXACTLY the same! (sorry for this disgression)

Greg.

I really hope things will go in a more modeling / sampling combined direction. So far I haven't seen anyone really do it. Yamaha's "Spectral Component Modeling" appears to be the regulation tiny sample set, with blended velocity samples and well done looping samples (for small loop samples that is), modeled in-line effects, and some stretching thrown in - i.e. a slightly newish spin (for DPs - modeled electric guitar effects have been around forever) on really ancient stuff.

In the spectral domain it should be fairly easy to blend in both the velocity dimension (which everyone seems to be doing) and frequency dimension (which no one seems to be doing) and gain quite a bit of sample compression without all the attendant audible issues in the schemes used to date (looping, stretching, layering).

Separate out the strings from the soundboard, then convolve in the soundboard of your choice at playback to pick your piano type. I wish they would take all the labor they currently expend shoehorning what once was a beautiful piano sample into a lame sounding 10MB ROM and put it into something like that instead.

Personally, I'd rather have a sound that is somehow based on a real piano (e.g. impulse responses) than one that is 100% modeled from scratch.

The Mystery Stretch Game - RESULTS!

Grand prize goes to bkmz who correctly identified the stretch groups in all samples, including the extra credit DP. The comment "1 & 2 sounds similar.." was also spot on (see below).

Second prize goes to Volusiano who correctly identified the extra credit DP stretch groups first.

Honorable mention goes to JMMEC for getting some of the stretch groups right, and for ranking DPs #1 through #3 the way I tend to think of them.

Let's have a round of applause!

What's that? What do you win you ask? Well, er, um, the satisfaction of a job well done!


Here are the results:

Stretch Groups:
1) 4,4,2,4,3
2) 4,4,2,4,3
3) 3,3,3,3,3,3
4) 4,4,4,4

DP Make & Model:
1) Yamaha CLP-330
2) Yamaha P-155
3) Casio PX-330
4) Kawai MP5

Pix (1:4):
What's fascinating about the spectral phase pictures here is they reveal the fact that Yamaha CLP-330 and the Yamaha P-155 share the same sample set at MIDI velocity 63!

Could you tell me what headphone are you using or what will you recommand for this AvantGrand ? is the headphone Amp "powerfull" enough for headphone with low efficiency or high impedence ?

Thanks

My Yamaha dealer told me about current AvantGrand models being replaced (by new ones) "early 2011". Does someone has infos ?

... didn't I just respond to this very same question?