Quote:
Originally Posted by JamesPeters
Lol. I forgot to check if I was using a lowlatency kernel.  I was using the generic one. (MX's kernel has the lowlatency timing by default, if I recall correctly.)
Now that I switched to the latest linux-lowlatency kernel, I can set the buffer to 5 blocks of 64 samples and it's every bit as solid/stable as MX was. This system is a little better overall but also I didn't use RTIRQ in MX. I don't know if that's having any effect on my system performance though since my audio card is PCIe and it was probably fine already. |
Very cool. With a card inside the machine, I would expect pretty low latency. I could run all day set to 64 samples and 2 periods with my M-Audio PCI cards as long as I didn't use many bridged Windows plugins. Same is true with the USB interface I use now but I have to use 3 periods instead of 2.
I run so much other stuff while I'm writing though, I just keep it set 128/4 where nothing can knock it off the track. Latency is livable enough for me set like that and I never have to change anything. I used rtirq with my PCI cards, and have it set now for both the one remaining PCI card (for computer audio now) and the USB port for the new audio interface.