I recorded a test file on the DC3X and then opened the midi file in Pianoteq. Enspire pro claims to record 1024 velocity values, and I'm not sure what that would look like. The thing I noticed in PT is that the values are now decimal. Is that their implementation of 1024 values? The decimals seem they could take on the values of .125, .250, .375, etc. So 128*8 is indeed 1024.

Anyway, I'm wondering if PT actually does anything differently? I can't help but to think everything would be the same if you simply rounded the number to an integer.

There is also a damper velocity value??

Pianoteq doesn't sound too bad when piped back into the piano and output via the monitor speakers beneath piano. Although acoustic is obviously better sounding.

I think they should change the model name to DVC3X, or "Darth Vader" C3X. I couldn't help to think that when I looked at the underside of the piano.

"Thanks to the physical model that Pianoteq is built on, it has the unique capacity of reproducing all the 16,384 velocities by the Hi-Res CC#88 format, and all the 1,023 velocities by the XP MIDI format."

This Japanese document shows how the 10-bit data is calculated for XP MIDI (as you suspected). See pages 14-15

https://www.slideshare.net/knoike/midi-25801187

The only Japanese you might want (but don't need)
整数 = integer
小数 = decimal

So you’re saying Pianoteq is the best?

As you noted, data is sent differently for cc88

What is that picture?

I have never managed to find an inaccuracy of my 127 levels virtual pianos (excepted badly sampled ones... EWQL Bechstein, an Imperfect Samples...). Then I can say 127 levels is enough to me.... and the advantage of Pianoteq is not enough too make me prefer it than other sampled ones.

?? No, not at all. I hope I didn't imply that somehow.

@newer player : Thanks for that, although I really can't read it ;0

The DC3X doesn't transmit it though, only records. And I'm not sure I'd ever by able to tell the difference compared to 127 levels.

pianoteq is becoming something like politics that people are afraid to express their preferences

Basically it says a XP MIDI velocity signal is sent not by using cc88 (used for high resolution MIDI) but by combining note on and cc16 to get 10 bit.

Looking at the second picture above we see a XP MIDI velocity of 73.25 (out of 128) for PianoTeq (last row). That is from note on (73) + cc16 (0.25).

Also, the old RePerform software displayed this XP MIDI data as 586 (out of 1024). That software seems to be discontinued and the page goes to Steinway Spirio.