Dave Mollen

I've been trying to change the grain duration at audio-rate, but can't figure out how to do it.

I walk through the grains with a phasor and I want to be able to change the speed of the phasor only at the end/start of a cycle.

In the patch below you can see what I've tried so far.

Now I know this won't work, because it's a feedback loop.

But I don't want to have the delay of one vector size either.

I've also tried to use two phasors, firing one after the other has finished a cycle,

but I can't figure out how to do this right as well.

I've been trying to change the grain duration at audio-rate, but can't figure out how to do it.
I walk through the grains with a phasor and I want to be able to change the speed of the phasor only at the end/start of a cycle.

In the patch below you can see what I've tried so far.
Now I know this won't work,  because it's a feedback loop.
But I don't want to have the delay of one vector size either.

I've also tried to use two phasors, firing one after the other has finished a cycle,
but I can't figure out how to do this right as well.

granular-at-audio-rate

i am no expert. i muddle along with all this stuff and get reasonable results. you need to introduce the secret MSP 1 vector delay:

you should search this forum as there are many excellent patch examples of exactly this sort of thing, going back many years. basically, sah~ is your friend, small use of vector delay is your friend, and you will be interested in the awesome volker bohm external [vb.phasor0~] and the andrew benson signal externals, especially [wait~].

p.s. - you only want 1 phasor~ for EVERYTHING. use the rate~ object lots to split up your syncs and timings. also, forgot to mention:

oops! that first exaample should have been audio rate send/receives of course!

But all your examples use a delay of 1 signal vector.

But actually I'd rather have my patch to be absolutely sample-accurate.

So I wondered if there was any other way to do this.

Maybe some construction which doesn't involve feedback.

Thanks for the advice.
But all your examples use a delay of 1 signal vector.
Maybe this isn't that bad.
But actually I'd rather have my patch to be absolutely sample-accurate.

So I wondered if there was any other way to do this.
Maybe some construction which doesn't involve feedback.
Then you wouldn't need the vector delay.


I wouldn't call myself a digital-audio/DSP guru, but I've been researching this exact topic for several months and all the above comments and recommendations hold true in my own implementation of a granulator in Max.

There are numerous threads here, going back several years recommending different approaches and the 'central' [phasor~] to [play~] plus [sah~] algorithm works best (for me);

+1 for vbphasor~; it does exactly what is expected - I just built my own though, for more specific control.

This related thread may take you someplace helpful...

https://cycling74.com/forums/help-with-an-exercise-about-filters

This related thread may take you someplace helpful...
https://cycling74.com/forums/help-with-an-exercise-about-filters

I've tried both the tapin~ tapout~ and the send~ receive~.

But this actually doesn't work.
I've tried both the tapin~ tapout~ and the send~ receive~.


Later this weekend I will be posting a detailed video description of one approach to audio rate granulation, on my blog if you wish to see it:

I'll check out your video, maybe it'll help me with some other difficulties.

Thanks.
I'll use this then for now.
I'll check out your video, maybe it'll help me with some other difficulties.


Granular at audio-rate