80.133.237.179
In Reply to: RE: USB audio interface modes posted by John Swenson on May 27, 2007 at 19:47:50
Hi folks.
As John said, the standard adaptive USB receiver is not able to cope with jittery incoming signals. "Rubbish in = Rubbish out". Even though you'd use precision clocks on the receiver chip ( I use a PCM2707 with Tent clock). Many have tried to recover or reclock the signal afterwards and they all had to realize, once the signal is messed up you won't be able to revover it. (That's valid also for SPDIF/Firewire and so on)
The USB jitter depends heavily on the PC Timer clock (1000Hz) and the IRQ handling.
As I posted in some other threads over here, it is worth to improve the upstream situation first, before you try to swap HW.
Of course under Windows you're very limited with the tweak potential.
Here the usb-audio.com driver is your first choice to get the distortions down. You can further configure your application lowest priority within the task manager. This gives the background-processes, such as the usb-communication, higher priorities.
You'll hear the difference.
You should also try your different USB ports. On my quality T60P they all
sound different.
Here your tweak potential under Windows stops.
Under Linux you're entering a different arena.
Just some tweaks you might apply to improve the USB situation:
1. Realtime kernel (Con Kolivas)
2. Higher timer frequency (e.g. 10000Hz)
3. Tweak of USB Driver snd-usb-audio
4. Look for the best USB port
5. Lots of other stuff further upstream (application and buffering)
With above you limiting the jitter on the USB heavily.
I do not feel I'd have to switch HW any longer.
Windows is not coming close to my Linux setup.
I wonder if the EMU would solve all the issues on the PC side.
I mean -- even if the EMU is in sync with the USB-controler, it won't resolve issues generated further upstream. The EMU would be an
interesting try if I could tap off an I2S signal.
Cheers
Klaus
Follow Ups: