Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards Roland FP10 Hi-Res Midi & Release Vel ??

Hi, I'm looking at both the Casio PX1000 and the Roland FP10 digital pianos.

I know the Casio PX1000 transmits both Hi-Res Midi and Release Velocity, (that
was stated by Mike Martin)

Does the Roland FP10 also transmit these? I can't find any info. I know the FP30

DOES transmits these.

Does anybody have any solid info on this?

Thanks for any help!

BK

According to the Roland FP-30 spec, this DP supports note-off/release velocity but it does not support hi-res MIDI since the velocity data uses 7 bits (0-127). There is no such information available for the FP-10 at the moment. But if the FP range (including the FP-90) do not support hi-res MIDI it is likely that the FP-10 will not support it as well.

I am not aware of the MIDI capabilities of the new PX-S1000 model since Casio seems not to have yet released its documentation. Note that hi-res MIDI and note-off velocity were only supported by the PX-360, 560 and 5-S. The entry model PX-160, which is now being replaced by the PX-S1000, did not support these two features.

Anyway, may I ask why do you consider hi-res MIDI to be a relevant feature?

I consider them a relevant feature, but I doubt HiRes MIDI is useful at all with Casio's plastic actions. I wouldn't be surprised if the lower 7 bits of those 14 bits are just white noise.

One consideration is that PianoTeq might be the only piano VI that uses high resolution MIDI. There may be some obscure piano VIs but I can not recall any.

PianoTeq has a free trial version so you can see if you like the sound and playability. If so, maybe you would consider that a digital piano with high resolution MIDI out capabilities might be worth trying. Quite a few people here use PianoTeq but very few have tried high resolution MIDI as there are not many controllers that offer that feature.

All that said, internal sounds on the digital pianos are not necessarily limited to the basic 0-127 velocity range and may well be using higher resolution data from the keystrokes but I don't see any evidence for or against that.

In theory it makes sense to have additional resolution, especially with controllers such as mod wheels. But I wonder if there are any practical advantages of hi-res note-on/off messages. I have tested a Casio PX-560 with PianoTeq with and without hi-res MIDI and the result was basically the same...

Over the years, I have read posts from about a half-dozen people who also tried this scheme and felt the same. I searched quite a bit but haven't uncovered any posts from Piano Players noting any difference. There may be cases but I haven't seen them.

High resolution MIDI sure seems like a smart idea. Perhaps virtual instruments could take advantage of extra velocity levels with some "intelligent" randomising factors; OTOH that might be done just as well with the standard 127 MIDI velocity levels.

I have yet to see an explanation (or a use case) that explains why anyone could make use of high-res MIDI.
We have a span of 127 with regular MIDI ... roughly 1% increments.
Is that not enough? What could be achieved with the finer precision of high-res MIDI?

Maybe some specialized controllers like the Roli can take advantage of a larger sensitivity range. The same may apply to synth-like keyboards with aftertouch/pressure detection or controllers like wheels or knobs. However, I also fail to understand how high-res MIDI would provide a tangible benefit with hammer-based keyboards. The quantization would be lower, but the low-res MIDI range already covers roughly 1% increments as you say...

Thanks guys for your info and opinions on this matter.

I was mainly interested in release velocity but HiRes Midi is an intriguing
concept, although maybe at this time, not a noticable improvment.
I guess I'd rather have it in a DP than not.

Release velocity could/would have more tangible impact on both Modeled and
sampled pianos.


BK

The internal sound engine of the CA78 and 98 makes good use of the release velocity. The Rolands with the modelled pianos also use it. Another example is Pianoteq. But another question is if the 16K levels of resolution have any practical advantage over the 127 levels. Anyway, all velocities (note on/off, pressure, aftertouch, controllers,...) in the new MIDI 2.0 spec have 16K resolution. So, after almost 30 years with the current MIDI spec, we may have a new generation of digital instruments supporting the MIDI 2 spec in the near future...

The only time I've heard of hi resolution MIDI being implemented was that VAX thing that was a flash in the pan, and hmm, I think the Roland RD-2000 has it for Roland Cloud use, or something similar to it. I don't know that it works for anything besides their cloud.

Despite it being a spec, it doesn't really seem to be implemented and taken advantage of yet? I'd be glad to informed otherwise. I believe some instruments might take advantage of greater than 0-127 internally of course, but when it comes to sending midi data, I really haven't heard anything about hi resolution midi actually being sent by hardware instruments.

The DAW I use, Studio One, translates standard MIDI to 32bit high resolution MIDI for internal use -- but then back again to normal sending it out.

Hi-res MIDI is just the MIDI data transmission protocol. A keyboard or software like a DAW or VSTi may internally use higher resolutions.

The hi-res MIDI specification (aka CC#88) was not part of the MIDI 1.0 spec. It was specified around 2010 as an addendum and it will be part of the MIDI 2 spec. Currently there is a limited numbers of keyboards supporting it and sending hi-res data. These include the Lachnit/FL Keys, VAX, Casio PX-360/560/5-S and Roland RD-2000. All of these send hi-res velocities and at least PianoTeq will use the information. Other controllers like the Roli also support hi-res velocities. I understand the use case for hi-res velocities on "continuous" controllers like the Roli but not on a DP with hammer-based action...

Has anyone ever tested the precision of high res midi on a hammer action keyboard? Like using a key pressing robot? It would be interesting to know if the same velocity can be struck reliably, or if it's all jitter...

And ... can anyone actually strike keys at all of those thousands of points ?
And just how does a note at velocity 19,384 sound different from one at velocity 19,385 ?

Who cares - even if it responds precisely to the robot, I still don't think a human is precise enough for it to have any benefit.

Greg.

To me there are two issues. Whether the sensors and the action can repeatedly strike at a given rate (the old "measure with a micrometer, cut with an axe" thing), and whether a human player or listener can tell the difference between 128 and "more than 128" levels of velocity (which includes volume).

The robot would be a good test of the former to tell if it even makes sense to transmit those values, but I honestly think that asking whether a pianist can strike the precise velocity requested is nothing more than a red herring. I doubt most concert pianists can reliably play a virtuosic piece and strike even 10 discrete layers perfectly, but that doesn't mean the audience can't hear or feel the nuance of playing more than 10 layers.

If the mechanism can't reach the resolution but people CAN hear the difference, then I think it would actually be beneficial for VSTs to just jitter the input, and play a velocity of 60 randomly between 59.5 and 60.4, for instance. It's be comparable to accepting the input from a mechaniclsm that doesn't have the timing resolution in the first place

Why ten?
When you choose ten ... you may be right. Ten may not be enough. MIDI needs more than ten discrete values.

But how can the case be made for tens of thousands of values?

As computing becomes way more powerful and as data transport schemes become way faster and able to transmit bigger more complex messages, the issue is why not? In general, computational messaging has continually moved to "no limitations". While we may not be able to appreciate one particular aspect of more detail, the overall architecture of "more detail" may produce big advantages in something else one day. Just something to think about.

Peace
Bruce in Philly

In other words "more is better"? Even if it isn't?

The point I'm trying to explore is that it seems a strawman to claim "pianists cannot precisely strike a requested velocity on demand, therefore there is no value to having that many velocities." If a pro can't even do it in 10 limited layers, but we also agree 10 layers isn't enough for nuance, that whole strawman falls apart. Being able to do something on command is very different from being able to perceive a difference between that and "not that".

Honestly, I feel 128 velocities is probably more than enough for piano, but technically it's also fairly "cheap" to output at the reliable limit of the mechanism...why not capture in as much detail if doing so is essential "free?" Clearly there are other things that should take precedence (speakers, sample length, storage size) and make a bigger difference, but free is cheap.

And even if 128 is not enough, we shouldn't immediately jump to "16k is too much!" Because if 128 happens to be deficient, the existing 16k hi-res standard is probably enough to capture whatever nuance is missing from 128, be it 192, 256, or 512 layers. We should be measuring against the minimum imho.