Let me make it clear that I’m NOT having any problems with latency, but I’d like to keep it as low as possible…so I’m asking (after not finding a good answer using search)
If the network latency/overhead from using udpsend any greater than the latency between m4l devices using regular send?
I have my suspicions, but without a real measurement there’s not really any point.
If I wanted to set up some devices to measure this, (in general terms) how should I go about it?
the problem with send device is that latency is un-expected.
from what i gather something to do with multi-core processors
and different "levels" creating different latency’s.
but i cant tell you that i’m using the plugsend~ device
and that the latency is very minimal. i can scratch into a muted track
with plugsend~ on and monitor my scratching on the plugreceive~ channel
and it will still be responsive enough to scratch.
I understand what you mean, but I’m talking about the general send/receive and udpsend/receive. It’s not audio data, it’s midi.
I *am* happy your latency is low enough for scratching though. :)
I’ve measured the latency of send/receive and udpsend/udpreceive with [cpuclock].
In both cases it seems directly proportional to Live’s audio buffer size, ie.
512 samples: ~12ms
128 samples: ~3ms
Apparently the transmission is handled by buffering in sync with the audio buffer.
C74 RSS Feed | © Copyright Cycling '74