Good evening, everyone.
Mark, I was probably too casual in my description of these models, trying to provide general information without getting too specific.
It is true that all Disklaviers have a good bit of "head room" at the top of the note-on velocity scale. In my experience, reasonably expressive playing will live in the 30-95 range, a performance with a noticeably wide dynamic range will be in the 20-105 range (with an occasional spike at about 110), and a performance from the Minnesota International Piano-e-Competition may includes notes under 10 and may reach around 120 at the top (rarely, of course). And P-e-C performances may include some brushed notes that result in key movement without the hammer hitting the strings.
I am not sure that I have actually seen a note-on velocity of 120 in a Disklavier recording, but someone at Zenph with whom I spoke recently mentioned seeing 121 (if I recall the conversation correctly).
These numbers mean, of course,that the Disklavier really does not use the full spectrum of note-on velocities up to 127. Naturally this raises the question of "Why not?"
I have not heard a statement from Yamaha on this subject. As a general rule, Yamaha does not share this sort of technical information.
My own opinion is this: If you were designing a piano and decided that the loudest note that YOU could play was 127, one would have to wonder if that was the loudest note that any human can play. In other words, you play a note with all the force that you have and generate some measurable hammer velocity in feet per second. You decide to map that velocity to a MIDI value of 127. Then, someone else comes along and plays an even louder note. Does it also get 127?
It seems to me that it make sense to build in some headroom into the velocity scale so that no one ever exceeds the top MIDI value for velocity.
The Pro, unlike other Disklaviers, measures the MIDI velocity on a scale of 0-1023. This means, essentially, that a range of values is generated in between the normal MIDI values between 0 and 127. In other words, the Pro adds decimal places to the MIDI velocity concept. If I am wrong about this, someone from Zenph will need to speak up and correct me. Zenph has studied these issues in greater detail than just about any other entity in the world.
I think that you'll find that the best tools for normalizing MIDI performances for all Disklaviers will be coming from Zenph. Zenph's new RePerform editor is currently undergoing public beta testing. You can read about it here:
The program is currently focused on the editing of XP MIDI data. Odd as it may seem, there is no other sequencer in the world that makes it possible to edit XP MIDI data in any sort of convenient or practical way. XP data is comprised of a bundle of MIDI messages--many of which occur on the same clock tick--that are uniquely interpreted as high resolution data for note-on velocity (i.e. hammer velocity), key velocity, note-off velocity, and pedal data. The typical MIDI sequencer reorders some of these events in the MIDI data stream during file import, export, or both, thus resulting in corrupt XP data.
RePerform is a program that will grow in its feature set. Keep an eye open for it!
Regards,
PianoBench
On Mar 9, 2011, at 6:54 PM, Mark Fontana wrote:
George, it still remains the case that the effective "typical" velocity
range of a Disklavier (ALL past and current models) is compressed into
the MIDI velocity range of approximately 30-95, correct? Otherwise,
Yamaha would need to release different versions of their music diskettes
for different models.
The difference on newer models is that levels 95+ continue to get a
little louder and levels below 30 continue to get a little softer (to
some point), but the core range of 30-95 is intended to produce roughly
the same dynamic range in decibels across instruments. Right?
My understanding is that XP enhancement just adds more incremental steps
along this existing curve.
It would be great if Yamaha could develop some guidelines for content
creators (tips for ensuring consistent performance across Disklavier
models, particularly regarding note velocities and pedaling).
I'm not aware of any such documentation. It seems like Zenph,
WebOnlyPiano and others are having to figure these things out for
themselves. Is there anything official?
Mark Fontana
On Wed, Mar 09, 2011 at 05:17:17PM -0500, George Frederick Litterst wrote:
> The early Disklaviers compressed the dynamic range on playback,
> boosting the low MIDI velocities and capping the high MIDI velocities.
>
> Starting with the Mark IIXG Pro, Yamaha achieved a full spectrum of
> playback possibilities on the Disklavier. The original Pro and
> subsequent Pro models have been unsurpassed in this regard and offer
> more resolution in between the traditional 0-127 velocity
> measurements. The Mark III, current E3, and Mark IVs non-Pros also
> have a much wider dynamic range compared to the early model
> instruments.
>
> Regards,
> PianoBench