Re: [LAD] RME madi latency

It turns out I am doing something very wrong, and i *can* get much lower latency’s even with the large buffersize. I could not get jack_delay itself to work (it wouldn’t connect to any ports no matter what i tried), but i did use jack rather than my own alsa implementation and with a period size of 32 samples @ 48KHz I measured a roundtrip delay of around 169 samples (3.5 ms), which is very reasonable. The method I used was perhaps a bit crude, so it could be off by a few samples. But it seems accurate enough. (I used another soundcard on a different machine, and plugged one output straight back in, and the other went through the converter /madi fx / jack and then back in. The difference in timing should be the rtd). So now i am a but confused how to get the same kind of latency in my own code (using alsa directly rather than through jack) Basically what happens is when i have a buffersize of 8192 and a period of 32 is that snd_pcm_avail_update(capture_handle), will keep returning 0 until i have played back the full 8192 samples. After that I will start receiving input samples, but obviously the latency is now more than 8182 samples… How can i make alsa not wait until the entire buffer is full? best, Fokke