View Single Post
Old 01-18-2008, 06:26 AM   #27
axeman
Human being with feelings
 
axeman's Avatar
 
Join Date: Dec 2007
Location: New Jersey, USA
Posts: 244
Default

Hi again,

Quote:
Originally Posted by pipelineaudio View Post
See what DPC Latency checker says, my dell's give me the same grief if I run USB audio
Thanks a lot for reminding me about this tool. Will give it a shot and report my findings.

Quote:
Originally Posted by Tuukka View Post
Thanks for your reply.
I was referring to Cubase LE4. I think it was not intuitive enough.

The 1641 manual says that you should not have any kind of mass storage connected. It will eat the same bandwidth..

You can see the bandwidth statuses from Device manager -> USB bus controllers -> "USB 2.0. controller" -> Additional settings. You can try also to disable the USB 1.1 controllers.

Does someone know why system is reserving 20% of the bandwidth all the time?

Why is it a bad thing to record to system drive? Would it be ok to record to a different partition instead?

Tuukka
My pleasure. Actually, I haven't even powered up the external USB hard drive while testing with the US-1641 (yet). Am trying to see how fast/stable I can get things prior to injecting any other variables into the equation.

Appreciate the tips on dealing with the USB controllers. Will probably dig through there, after running the DPC Latency Checker tool (as Pipeline recommended).

I think I can answer your final question: It's bad to record to the system drive because the system routinely has to access it. Therefore, you are sacrificing disk/disk-access performance by interleaving your digital audio with system-related access. I don't think different partitions would help (with the exception of reducing fragmentation on the audio partition), as you're still stuck with the same disk access bandwidth and spindle speed.

By the way, I finally had a chance to try adding a few more parts to my US-1641 test recording last night. Everything seemed to be going OK for about 15 minutes or so, but then the audio output started chopping up quite a bit. Ended up changing the only remaining options I could think of, which I believe included clicking on REAPER's "Pre-zero output buffers, useful on some hardware (higher CPU use)" and "Ignore ASIO reset messages (needed for some buggy drivers)" settings--as well as forcing the US-1641's clock source to internal (instead of auto). (I kept the US-1641's latency setting at "normal" for the entire session.) After enabling these changes, I did not detect any further dropouts/choppiness for the 30 minutes or so before I shut down.

One interesting thing: It seems like all the problems are associated with playback (don't even have to be recording anything or presenting any audio signal to a single input when the dropouts start up), although I'm not sure if the US-1641 actually sends USB data for the audio on all 16 channels, even if nothing is attached (presumably all-zeros)--or if it's smart enough to set a threshold under which not to bother sending data toward the PC (so that unused inputs do not waste USB bandwidth). If I have time, I may hook up as many microphones as I can find and just try recording them all at once, just to see if and how that affects performance.

Take care and thanks again,


Alan
axeman is offline   Reply With Quote