Midi clock stabiliser

[samplaire]

Hi,

It would be simple (really?) - you know, when you want to use your sequencer midi clock to synchronize, say, SFP arpeggiator device it always looses the beat. Argh! It's difficult to explain but I hope you get it. Everybody knows the midi clock isn't that stabile if you send it to any external gear (and SFP is something like it). You get more or less the desired tempo but it changes a bit and it's audiable (for example you have 120BPM and you get once 120.1 and once 119). The device I'd like to have would test the clock for a second or 2 and then stabilize it to the desired value.

[at0m]

Good to bring this one up again, Samplaire

Yes, I find that problem very annoying when playing any of the synths with built-in sequencer, or Lightwave for example, which has nice sync stuff going on. At least, untill you want to play along with this drumtrack and suddenly everything starts to drift away. We can't use Flexor's ramp for everything eh.

That's not even the bug where you can't slave any CW device to incoming clock, it's just the clock drifting on it's own.

Anyone working on 48khz can confirm these two problems happen there too, cos afaik all freq signals (here BPM) are primarily calculated for that samplerate.

[samplaire]

On 2005-02-13 11:46, at0m|c wrote:
Good to bring this one up again, Samplaire

Oh, my! Now I understand your words , at0m|c - my memory leaks

[Mehdi_T]

I have been able to make a Scope tempo display send midi CC messages to record in the sequencer. After, this track midi CC can be assigned to any scope tempo display.

However, it is so complicated in use, that I doubt anyone would ever use it. It require a bit of parametrization of the device before recording, and user intervention to have that midi CC track match the tempo change (when sequencer tempo changes, then midi CC track tempo changes, and the tempo is not changed at the same speed - so it may require quantizing/redrawing from the user.).

However, it doesn't use sequencer tempo, and allow all devs to be in internal tempo mode, it is stable.

So there is way to make all this work, but it does require a bit of work on the user side, and that's why i wonder if anyone would ever use this (but it works, that's sure).

May be i'll do a free device for this in a little while when i finish my ongoing updates.

well, only if people are interested... not sure it is very interesting (not sure i would use it myself so... ) it's the kind of things people ask for, and never use so, as long as there is no more interest...

So you better insist, or I keep it for myself

but it is not complicated, any sdk user can make it with a little bit if thinking.

see you..

<font size=-1>[ This Message was edited by: Mehdi_T on 2005-02-14 14:09 ]</font>

[samplaire]

It took me some time to get the idea, Mehdi, and I think it's a good alternative to the stabilise I suggested. My midi knowledge is limited (especially in sync/MTC/MMC/Midi clock area) and I thought the MCTtoClk module found in 'midi' devices folder would do the job but I couldn't connect it So, yes, I'd appreciate your device. But if I'm the only one to use it, don't bother me.

And what about my idea? Would it be difficult to prepare such stabiliser? What kind of knowledge would be required (besides the obvious knowledge about the CWA SDK)? What kind of calculation? Does it require any new dsp file?

[Mehdi_T]

hello,
about other solutions I really do not know. I think it could be a sequencer/clock generator problem in the first place, and would not know how to "smooth" this. I noticed a few other developpers have studied midi more than me (worlf and alfonso for ex). may be they can help better ?
the "tempo to cc to tempo" is not so complicated to do, but i don't know when it could be released.
I thought about it and it could be more useable than i thought in the first place, and can be an intersting workaround (expecially for tempo changes through internal clock of devices i think).
best
Mehdi

[samplaire]

Great, thanks

[symbiote]

This might be do-able with some sort of simple integrator, that would tend to stabilize around the average MIDI clock value. Might not be absolutely stable, but going to be much more stable than straight MIDI clock from sequencer to SFP. I'll fudge around with this a bit when I have time and provide coherent (yeah right!) and cohesive report/proof-of-concepts.

[alfonso]

What I can say is that I've never used slave midi in SFP.

I think that almost all the times there is not a real need for that, if incoming notes can retrigger an internal clock.

I think that the Ramp (FleXor's) concept of syncronizing sequences should be the real future, as it provides a start, an end, divisions and an audio-rate precision. There is a lot of work to do, but if you make a track of ramps in your sequencer(careful not to send it to your speakers), like at0mic suggested long time ago, you will have the tightest timing protocol you would ever dream of.
Then even gate signals could be rebuilt from those ramps to control the traditional CWA clock dividers, removing the Midi Clock module, but this is for the modular only....
Lot of work to do....

[valis]

You know what would probably make that accessible to a lot of people. I hate to say it, but some sort of vst or VSTi plugin that sends the ramp signal. Personally I have no issue with creating one as audio and using it that way, but given the complexity of our platform I think a plugin would serve a lot of people well while they come to grips with the less complex aspects of Scope.