Hi

There was a previous spost about 'power of 2 in buffer size' but I deleted it....

What is the holy grail of setting the buffer size in Pd? I'm used to setting ^2 sizes like 64,128,256 etc but this doesn't seen to translate into anything meaningful in Pd.

I've seen posts saying stick to 50,25,15 ms etc but this doesn't translate to ^2 buffer sizes. Can I just put the buffer size in as samples? Ms just seems a little arbitrary to me.

Also, I seem to get much better performance when running in Jack mode. Should the buffer size in Pd be the same as Jack?

thanks S

I guess Pd gives you this way of doing it because the block size is set per patch. Matching the buffer sizes is good practice but not that much of an overhead for either side to split or join different blocks, it's just concatenation. You can choose any ms value and it will round to the nearest ^2 without you having to worry.

I'm not an expert on Jack, but I believe the client-server connection involves a negotiaton so that
we get matching/optimal buffers. What I do is just tell Pd to use Jack and don't set anything in Pd, then go to qjackctl and set up the business there - Pd then follows along on the next restart of jackd

Anyone noticed how this works in more detail?

AHhh peace of mind!

I've noticed that when Live is using Jack for its audio driver, the buffer size box in its prefs panel is greyed out and adopts to whatever is set in Jack prefs. So this does concur with what you just said.

thanks again