HD Midi

Anything about hardware musical instruments.
Post Reply New Topic
RELATED
PRODUCTS

Post

I wonder whats the status of it regarding hardware controllers that support it. Or are there any similar technologies. I would love to have high resolution controllers instead of the dreaded 7bits that are still used in MIDI for centuries now. It seams that only a few controllers out there support high resolutions MIDI messages, I recon Moog and Numark have 14bit controllers but that usually is not supported by many DAWs if any at all.

I'm really tiered of the stepping I hear (like on filter sweeps) on almost every digital or VA synth out there just because of MIDI. They try to compensate by some delay/lag but still thats not like the instantaneous continuous real analog sweeps.

Post

S0lo wrote:I wonder whats the status of it regarding hardware controllers that support it. Or are there any similar technologies. I would love to have high resolution controllers instead of the dreaded 7bits that are still used in MIDI for centuries now. It seams that only a few controllers out there support high resolutions MIDI messages, I recon Moog and Numark have 14bit controllers but that usually is not supported by many DAWs if any at all.

I'm really tiered of the stepping I hear (like on filter sweeps) on almost every digital or VA synth out there just because of MIDI. They try to compensate by some delay/lag but still thats not like the instantaneous continuous real analog sweeps.
Use inc/dec messages. Problem solved.
www.mkdr.net

MophoEd - the BEST DSI Mopho Editor VSTi

Post

Or "abuse" the pitch wheel to record parameter tweaks. That one has a precision of 14 bits :-P
We are the KVR collective. Resistance is futile. You will be assimilated. Image
My MusicCalc is served over https!!

Post

mkdr wrote:Use inc/dec messages. Problem solved.
Excuse my lack of knowledge, can you program a KNOB to do that? or does it have to be by button? I have the Axiom 49
BertKoor wrote:Or "abuse" the pitch wheel to record parameter tweaks. That one has a precision of 14 bits :-P
Tried that on FLstudio, yes the values sent are thousands above 128, but the increments between the values are still 20, so still steps!!. Are you referring to some thing else? Am I lost?

Post

Sounds like a hardware limitation. The pitch wheel has a value of 128 * 128 = 16384 possible values. Your keyboard's pitch wheen apparently goes through that in steps of 20, leaving just short of 820 values.

Used as a pure pitch wheel in a range of +2 / -2 semitones, that's 163 steps per semitone: 0,6 cents per step. With a range of a whole octave that's 33 steps per semitone: 3 cents per step. For pitch modulation that seems a reasonable resolution. If you set the range for modulating the filter narrower, this should be workable.

Maybe it's also an issue with the synth plugin. Some do smooth parameter changes, some do not. :shrug:
We are the KVR collective. Resistance is futile. You will be assimilated. Image
My MusicCalc is served over https!!

Post

Inc CC 96
Dec CC 97
Very simple.


But to be more helpful, what are you trying to achieve? If we knew the problem we could address it better. If you are indeed recording midi messages down to your sequencer and then using the data to control softwaresynths, you only need to learn to use your sequencer. Record the data into "automation data", then your seq will interpolate the data(draw lines between the dots your midi controller does) and playback should be lot better.

Hardware VA's don't step their controller data btw. The softsynths you are using are clearly designed to be worked with automation control, not with midi CC control.

If your Axiom sucks and you need a new controller there are great solutions for softsynth control. Automap for example. Maybe look into that.
www.mkdr.net

MophoEd - the BEST DSI Mopho Editor VSTi

Post

BertKoor wrote:Sounds like a hardware limitation. The pitch wheel has a value of 128 * 128 = 16384 possible values. Your keyboard's pitch wheen apparently goes through that in steps of 20, leaving just short of 820 values.

Used as a pure pitch wheel in a range of +2 / -2 semitones, that's 163 steps per semitone: 0,6 cents per step. With a range of a whole octave that's 33 steps per semitone: 3 cents per step. For pitch modulation that seems a reasonable resolution. If you set the range for modulating the filter narrower, this should be workable.

Maybe it's also an issue with the synth plugin. Some do smooth parameter changes, some do not. :shrug:
Ok, I just found out it's even worse than that, my pitch bend sweeps between the minimum and maximum in exactly 128 steps. In Flstudio, the minimum is 5827 and maximum is 9557. In Reaper the minimum is 0 and maximum 16383. But both do it in 128 steps.

The thing is, even if there is a way around it, or using another controller, why do we have to be stuck with a "pitch bend" wheel. Why not have all controllers 14bit or more
mkdr wrote:But to be more helpful, what are you trying to achieve? If we knew the problem we could address it better.


All I want to do is have ALL my controller knobs have 14bit or more resolution. Why do we have to do some indirect ways or edit some internal parameters to do such a simple task. Shouldn't this be out of the box by now
mkdr wrote:Hardware VA's don't step their controller data btw. The softsynths you are using are clearly designed to be worked with automation control, not with midi CC control.


Sorry for not being clear, yes I meant softsynths for the most part. Although I've seen hardware that steps, believe it or not, The prophet 8 steps when tweaking pitch and filter cutoff!!! knowing that it's DCO, DCF (not VCO, VCF) I could imagine that it's MIDI controlled. The 128 values are clearly there, make no mistake about it. For VA hardware, the nord modular I have is very smooth, but I've seen reviews of VAs that step, like the Yamaha MOX.

You see talks about this and that plugin not being 64bit (while it has nothing to do with sound quality most of the time, just more memory). You see long debates about how 96Khz "sounds better" than 44khz while that matter is highly subjective and depends. Very few times some one talks about the 7bits of MIDI, for god sake, 7bits!!.

Post

S0lo wrote: You see talks about this and that plugin not being 64bit (while it has nothing to do with sound quality most of the time, just more memory). You see long debates about how 96Khz "sounds better" than 44khz while that matter is highly subjective and depends. Very few times some one talks about the 7bits of MIDI, for god sake, 7bits!!.
Haha LOL!
That is so true :D



Midi is ancient, that's true. But most things can still be made with it. 128 steps is not a big restriction, if it's done right. Knobs on softsynths/effects usually have less steps in their GUI's, although the parameter itself can have a 32 bit resolution.

Most people HAVE found ways around the midi's 7bit restriction by now.


Btw. Prophet 8 has 164 steps in its filter. The frequencies are in semitones. Same as the Mopho... which i dare to claim to know very well :)

Btw2. The reason synths don't use 14bit controllers, or atleast the full resolution available, is because the midi-port would be overwhelmed with information when a knob is turned. On the Mopho a quick flick of the cutoff knob can rob all available bandwidth easily. If something were to be made to the Midi standard, i'd rather be concerned about the bandwidth. Make it usb 2.0 480mbps or something and be done with it.

Btw3. Midi is just a serial port. The way software uses it(or doesnt) makes it what it is.
www.mkdr.net

MophoEd - the BEST DSI Mopho Editor VSTi

Post

mkdr wrote: Most people HAVE found ways around the midi's 7bit restriction by now.
Really? Please enlighten me, because I haven't found any that are satisfactory. I know about the trick of mapping things to the pitch wheel, but this isn't very satisfying considering I have 16 knobs and sliders on my controller that are basically useless and steppy as hell. You call that a solution?
"The Juno 60 was often incorrectly referred to as a synth. It is, in fact, a chorus unit with a synth attached." -PAK

Post

Technically the specs allow for 32 controller knobs/faders to send a value in 14 bits (although not very efficiently.) See http://www.midi.org/techspecs/midimessages.php#3
Controllers #0-31 represent the MSB, #32-63 the associated LSB. Using NRPNs can yield the same results (even less efficient in terms of message length compared to the real payload.)

You are at the mercy of the manufacturer to implement this though.
We are the KVR collective. Resistance is futile. You will be assimilated. Image
My MusicCalc is served over https!!

Post

BertKoor wrote:Technically the specs allow for 32 controller knobs/faders to send a value in 14 bits (although not very efficiently.) See http://www.midi.org/techspecs/midimessages.php#3
Controllers #0-31 represent the MSB, #32-63 the associated LSB. Using NRPNs can yield the same results (even less efficient in terms of message length compared to the real payload.)

You are at the mercy of the manufacturer to implement this though.
That's all fine and good but what DAWs actually support this feature? If I get a controller that sends 14-bit midi data, will Ableton Live or Reaper or any other DAW be able to actually interpret it?

And if it is possible for the DAW to use 14-bit control data, what hardware controllers have this feature?

I'd gladly make the investment if this is actually possible. All I've heard so far is a lot of hypotheticals. Who actually is using this stuff successfully?
"The Juno 60 was often incorrectly referred to as a synth. It is, in fact, a chorus unit with a synth attached." -PAK

Post

joe:

1: calm down, breathe.
2: do some research, there *are* controllers that use RPNs and NRPNs to send 14-bit data. Behringer, Novation, and others make them.
3: look on youtube, you can find videos of people using Novation Remotes (or other gear) to control things like the Kawai K3 and the Oberheim Matrix 6... that is how its working, usually, through use of NRPNs. You can also find video of people using a Novation or Behringer to edit a Alesis Micron, which is ALL NRPNS.

Post

mkdr wrote:if it's done right. Knobs on softsynths/effects usually have less steps in their GUI's, although the parameter itself can have a 32 bit resolution.

Most people HAVE found ways around the midi's 7bit restriction by now.
I doubt, sounds more like most people are using their mouse which has a much better resolution (usually bound to screen height, like 1024). But still even that number is nothing compared to 16384 of 14bit, and that in turn is nothing compared to the almost infinite resolution of real analog control voltage. I agree with you that softsynths usually internally have a much higher resolution, thats why LFOs and envelops are usually very smooth, but once it comes to external control, that is no use. I would say that most poeple (if they notice the stepping) would probably blame the synth or the controller, few would realize it's a midi limitation to begin with. They would be lost trying to fix the problem by other means, and I don't blame them for that, most of the time you really don't realize the stepiness until you experience the smoothness of real analog gear (In my case a eurorack modular).
mkdr wrote:Btw. Prophet 8 has 164 steps in its filter. The frequencies are in semitones. Same as the Mopho... which i dare to claim to know very well :)
Thanks good to know, still very low though, for an analog piece of gear. I like every thing about my prophet except this part.
mkdr wrote:Btw2. The reason synths don't use 14bit controllers, or atleast the full resolution available, is because the midi-port would be overwhelmed with information when a knob is turned. On the Mopho a quick flick of the cutoff knob can rob all available bandwidth easily. If something were to be made to the Midi standard, i'd rather be concerned about the bandwidth. Make it usb 2.0 480mbps or something and be done with it.:)
Thats a good point, but lets calculate it. Say that a good keyboard player or synthesist would do a full filter sweep in 1/4 (0.25) of a second. So for 14bit, he will pass through all 16384 values in 0.25 of a second. And each value get sent by the USB port contains 14bits (obviously). So the rate would be 16384 X 14 / 0.25 = 917504 bit/s. Convert to Mbit/s, 917504/(1000 X 1000) = 0.917 Mbit/s which is nothing but a fraction of 480 Mbit/s that USB 2 can provide you. Now say that the player would turn two knobs at the same time with both hands and tweak a mod wheel with his little finger and press a pedal with his foot all at the same time, this is 4 X 0.917 = 3.67 Mbit/s still a fraction of what USB 2 can do.

Some synths like Diva, Saurus (I think) use some laging to compensate for that, so the values would fade slowly between the two steps in a fraction of a second that is so small you usually can't notice, but this will fail as you try to move the knob so quickly because the lag needs this small time to reach it's target value.

It goes further than this, it's not just about the stepping (and I wish some one would prove me wrong here). When the pitch of a waveform steps between two far values, a hard switch happens resulting in a sharp edge in the waveform at the step time, this sharp edge can introduce sub harmonics that are not in tune with the overall pitch. This is very subtle and brief. But apply this to a whole lot of stepping controllers, you would hear some thing wrong but just don't know what it is. Some might say it's aliasing while it's really caused by midi. I wish some one more experienced than me would say something about this.

Post

The problem isn't MIDI. The problem is hardware and cost.
NRPNs and other 14bit messages are great for transmitting high resolution messages from one application to another but controllers use 8-10bit ADCs to read analog pots and faders and 2 of those bits are usually eaten up by noise. To get 14bit resolution form a pot or a fader you would need a 16bit ADC for every control element or a complex method of muxing to get similar results using a few 24bit ADCs.

I would be willing to bet that very few would pay the price for a controller that used higher resolution ADCs. I may be wrong though.

The best solution at the moment is to use encoders which can increment or decrement a value (rather than using absolute values). However, high resolution encoders are also expensive. Most encoders in consumer level devices have 24-32 step of resolution. This means that without acceleration curves you can't inc/dec more than 32 values in a single rotation. Most people don't notice this because of the use of acceleration curves but some don't find it satisfactory.

Post

justin3am wrote:The problem isn't MIDI. The problem is hardware and cost.
NRPNs and other 14bit messages are great for transmitting high resolution messages from one application to another but controllers use 8-10bit ADCs to read analog pots and faders and 2 of those bits are usually eaten up by noise. To get 14bit resolution form a pot or a fader you would need a 16bit ADC for every control element or a complex method of muxing to get similar results using a few 24bit ADCs.

I would be willing to bet that very few would pay the price for a controller that used higher resolution ADCs. I may be wrong though.

The best solution at the moment is to use encoders which can increment or decrement a value (rather than using absolute values). However, high resolution encoders are also expensive. Most encoders in consumer level devices have 24-32 step of resolution. This means that without acceleration curves you can't inc/dec more than 32 values in a single rotation. Most people don't notice this because of the use of acceleration curves but some don't find it satisfactory.
Good point, but: http://www.neobits.com/texas_instrument ... ml?atc=gbs

Isn't this a 32bit ADC that sells for just $10. I'm not knowledgeable in the field so correct me if I'm wrong.

Even if the link is not relevant, from history we've known that cost of any electronics can be heavily brought down by mass production.

Post Reply

Return to “Hardware (Instruments and Effects)”