need for larger velocity scales, 0-128 to small

[peter_schwartz]

@ Kayemef,

It is never my intention to quote out of context. I may abbreviate a quoted passage, but I respond to what I believe are complete thoughts/concepts from the posts I quote. Anyway, I didn't mean to butcher what you wrote. Barring any further discussing about velocity and stuff, I'm going to suggest that you learn this one new bit of English, as in:

"Academic" -- not of practical relevance; of only theoretical interest: "the debate has been largely academic."

...and I'm going to try and learn a bit more French than just "merde".

@ ozy, "right on brutha!"

[Kayemef]

hypothetical or theoretical and not expected to produce an immediate or practical result

or

not of practical relevance; of only theoretical interest: "the debate has been largely academic."

...

?

[jemkeys25]

just remember midi v1.0 was writen back in 1983, 1983, almost thirty years ago, everything else with keyboards have evolved and advanced, except that midi code, back when keyboards had one sound at a time, where touch sensative was a big thing, when 32 mb of memory cost over $1000 dollars. It's time to update that code it's been long enough. I don't want 3 dimentional sound, I know the sound is coming from a speaker, but even a recording of a 9' concert grand piano sounds better than a recording of a rompler piano. Did you read the thread on HD MIDI from the MMA, even they know it's time to update, that it will allow for more exspressive instruments, by allowing greater controll over the nuances of sound, people sound like they have a vested intrest in midi v1.0. if they can make these instruments more exspressive, then I want it.

Why not?

sound can be broken up into more than 128 parts per human range of hearing.

[peter_schwartz]

just remember midi v1.0 was writen back in 1983, 1983, almost thirty years ago, everything else with keyboards have evolved and advanced, except that midi code, back when keyboards had one sound at a time, where touch sensative was a big thing, when 32 mb of memory cost over $1000 dollars. It's time to update that code it's been long enough.

The code is constantly being amended to accommodate new features. But the problem isn't necessarily with the MIDI spec, even with the new HD velocity resolution spec. See below...

I don't want 3 dimentional sound, I know the sound is coming from a speaker, but even a recording of a 9' concert grand piano sounds better than a recording of a rompler piano.

Of course it does.

Did you read the thread on HD MIDI from the MMA

Yes, I posted the link.

...even they know it's time to update, that it will allow for more exspressive instruments, by allowing greater controll over the nuances of sound, people sound like they have a vested intrest in midi v1.0. if they can make these instruments more exspressive, then I want it.

I don't think it's a matter of having a vested interest. Here's why... Manufacturers have had 30 years to make improvements to provide for better control of nuance using aspects of the original MIDI spec, like 14-bit pitch bend messages. That was part of the original MIDI spec. But 30 years later, many MIDI instruments and plugins still do not respond to the LSB (the "fine tuning" part of the message).

People still regularly complain about zipper noise when modulating things like filter cutoff. Why is that? Because 30 years later, manufacturers of various keyboard and other controllers aren't designed to respond to 14-bit messages. Then again, how many controllers output 14-bit messages? So the degree of nuance you can achieve isn't just about the MIDI spec. It's about how devices are programmed to respond to it. And in the case of MIDI controllers, it's about the kinds of messages they can transmit to allow for that kind of fine control.

And that's been the point I've been trying to make. I'm not arguing against more control. All I'm saying is that the situation isn't as black and white as having more than 127 degrees of velocity.

[jemkeys25]

i posted the "HD MIDI" link, thank you, that mentions the idea of" more exspressive instruments" by changing the midi v1.0 to hd midi, they said it ,not me, the mma, i trust that the midi manufacturers know what they're talking about, I don't claim to know anything about it, if they feel they can improve a keyboards exspressive response, then I believe them, and we should all believe them, instead of trying to find reasons not to believe them, or reasons why it can't or won't work.

[StephenKay]

i posted the "HD MIDI" link, thank you, that mentions the idea of" more exspressive instruments" by changing the midi v1.0 to hd midi, they said it ,not me, the mma, i trust that the midi manufacturers know what they're talking about, I don't claim to know anything about it, if they feel they can improve a keyboards exspressive response, then I believe them, and we should all believe them, instead of trying to find reasons not to believe them, or reasons why it can't or won't work.

Not to be argumentative, but Peter posted a link earlier in the thread (page 3, near the top).

[peter_schwartz]

i posted the "HD MIDI" link, thank you

HELLO?



jem,

The MIDI Manufacturer's Association have not changed the MIDI v1.0 spec to "HD MIDI". Then have amended the spec to include HD MIDI velocity response. It is not a whole new system. There's a big difference.

BTW, there's a slight (but very slight) downside to adding CC#88 to the data stream... the inclusion of that message will actually negatively affect the timing of MIDI note data, increasing MIDI latency by about two-thirds of a millisecond, worst case.

It's always a tradeoff.

Anyway, it's not the MMA that feels they can improve a keyboard's expressive response. Their job is to come up with a protocol to accommodate improvements & changes to the MIDI spec. It's up to the manufacturers of instruments and plugin developers to adopt and incorporate the changes. Each company will decide on their own whether or not they're going to incorporate HD velocity into their instruments, and, whether to make any of their older systems backwards compatible.

[synthjoe]

Manufacturers have had 30 years to make improvements to provide for better control of nuance using aspects of the original MIDI spec, like 14-bit pitch bend messages. That was part of the original MIDI spec. But 30 years later, many MIDI instruments and plugins still do not respond to the LSB (the "fine tuning" part of the message).

So true... And hence my complaint is not that MIDI should be replaced altogether, but there are elements to improve. Just as they proposed with HD MIDI velocity - I'm quite happy with that, provided controllers use it corretly and engines respond appropriately. And yes, the baud rate is another one that feels a bit outdated. Shall see.

I wish I could develop something that will be in use fundamentally unchanged for a mere 30 years by a large community...

[StephenKay]

I don't think it's a matter of having a vested interest. Here's why... Manufacturers have had 30 years to make improvements to provide for better control of nuance using aspects of the original MIDI spec, like 14-bit pitch bend messages. That was part of the original MIDI spec. But 30 years later, many MIDI instruments and plugins still do not respond to the LSB (the "fine tuning" part of the message).

That may be true for any 14-bit controllers. But most manufacturers do, I believe, handle 14-bit pitch bend (not, technically, a controller). Korg does, Yamaha does... the effects of 7-bit pitch bend are easy to hear if you set the pitch bend to an octave or more and then bend it with the wheel/joystick: unacceptable stepping/quantization.

I know for a fact that Korg has always supported (and uses) 14-bit pitch bend, because I had to implement it in my KARMA software for those keyboards. But you are correct that 14-bit controllers have been available ever since MIDI 1.0 and nobody in general has supported them. Pitch bend stepping is a bit more obvious than stepping on the filter cutoff, but not much when the resonance is high. It's all about cost of implementing these things and engineering resources....

[synthjoe]

Sorry, I've accidentally deleted this post - had to recompose it, please let me know if you remember something that is not in here...

Keybed --> Velocity curves ---> Velocity signal ---> see below

Excuse me Dan, I know it might not be quite appropriate for me to correct your words, but I think 'Velocity value' would be more accurate a term instead of 'Velocity signal' here. The difference is significant enough to highlight, I feel (matter of where the quantization takes place).

No, it's one interval, not two. Key down to key contact. The key down point is the zero point in the timeline. The time it takes to make contact is the first point. That's one interval, not two.

Basically, if so many musical instrument manufacturers have gotten it so wrong, no electronic keyboard instrument would feel right to play.

The technical implementations I've seen so far were all based on two contacts (break/make or make/make), key down is usually defined as the state alteration of the first contact rather than the touching of a key - so 'zero time' is actually a bit further ahead into the key travel. Similar 'two (or three) point' time measurements in opto and other sensors.

As to the second part of the quote, if Ford has gotten it all wrong with his first car, we'd not be riding cars today. If he'd gotten it all right then we'd still ride Model T today. My point is that us riding cars today does not warrant that Ford had nothing wrong with their most popular car at the time.

However, rather than having few (two or more) discrete sensors and measuring time, it would be interesting to see an inductive or opto coupled sensor that would track the complete key travel in an analogue way and derive the HD velocity value from the resulting signal curve.

[peter_schwartz]

Synthjoe,

You're absolutely correct. I had in my mind an image of the key contacts for my old A80 controller and only saw the gray blobs, totally forgetting that there are two contacts in there. Still, I would maintain that how quickly those contacts make/make are a perfectly sufficient way to measure "touch".

To your analogy... fundamentally, a car is a car is a car. Yes, we have greater speed, better safety protections, and fewer emissions. But those don't equate to an fundamental change in the function or definition of what a car is. A major change in the way we transport ourselves would need to be much more radical, like teleportation, or astrally projecting ourselves to work every day.


Stephen,

The degree to which 14-bit pitch bend is implemented seems to vary quite a bit. For example, testing out alterations of pitch bend LSB response using the sampler in Logic (EXS-24), only one value of the LSB (with respect to the MSB) cause a change in pitch. In other words, regardless of the fact that we have 128 bits of resolution between MSB's, only one actually causes any change at all in that sampler.

[synthjoe]

OK, let's leave the rest of my previous post.

This is, what I'm really intrigued by:

...it would be interesting to see an inductive or opto coupled sensor that would track the complete key travel in an analogue way and derive the HD velocity value from the resulting signal curve.

[mm-pro]

My beef is not the lack of resolution in MIDI data, but the inability of MIDI to deliver an expressive stream of data without smearing it in time. I stopped using MIDI sequencing around 1997 because of this fact. As I used more controller info to make my performances more expressive, MIDI couldn't handle the density of data in a reliable fashion.

Until this bottleneck is engineered out of the spec, I think any perceived lack of resolution is kind of moot, as MIDI can't reliably deliver the resolution we already have.

Regards,

MM

[jemkeys25]

I was talking about the" HD MIDI" thread on the main topics page, Peter posted the mma website, i posted the page on hd midi protocol.


p.s. is the kronos out yet?

[synthjoe]

Thinking of it, the hammer hitting a piezo element outputting the velocity signal should come pretty close to the real thing on a hammer action keyboard. Anyone aware of a keyboard that uses this instead of the time difference principle?