[somestranger26]

Do you have file buffering on in foobar and have you set the playback thread priority to 7? Changing the players' process priority just changes the UI and decoding priorities. Even with the playback set to realtime priority, the default driver for the TE7022 doesn't run in realtime so that's one thing out of your control.
 
What sort of results does LatencyMon give you when you're getting dropouts?

 

[drez]

Strange thing is that letency monitor doesn't show any activity that correlates to dropouts - ie when it shows hard pagefults (mostly from its own operation) there are no dropouts, and when there are dropouts from the DI there are no exceptional hard pagefaults or latency spikes (or spikes in processor or I/O activity for that matter)
 
This makes me think that its either to do with the driver mis-managing some kind of buffer, or some other electrical fault (eg. grounding).
 
When the DI is set to upsampling ie. it reclocks the data with the onboard DSP circuitry, there are no dropouts (but this occurs after the USB receiver chip does it's work)
 
I could try turning grounding off on the USB controller (has a jumper for this) and see if that works but then the DI might not work at all.  I'm out of my depth here and purely speculating - so if anyone can comment this would be helpful.
 
EDIT: OK so it runs without ground or power, and with ground but no power.  I will test both of these without upsampling when I recover the nerve to deal with dropouts again.  Initial impressions - with no ground and no power soundstage seems more intimate and hazy, and imaging seems a little more haphazzard (relative to no power and with ground).  No power and with ground (compared to no power and no ground) soundstage has more depth and sense of space.  Imaging is similar but more subjectively stable due to the better sense of space.  Will compare power power to ground only next.

 

[kr0gg]

Quote:

 
I could try turning grounding off on the USB controller (has a jumper for this)

 
nice feature. do you use a separate controller?

 

[drez]

yes I use this overpriced pci USB controller.  Not sure if any others can modify these settings.
 
To conclude my comparison I seem to find the ground setting to sound the best - best soundstage and instrument placing/imaging.  Strangely one of the USB ports detects the DI as "TE7022" while the other just shows "USB audio device."  Not sure how this might relate to the driver installed.
 
EDIT: unfortunately this setting does not allow me to test the OMILEX isolator with the DI (OMILEX needs to draw power from USB port).  One last variable to test is whether the computer's main PSU sounds better than the external PSU I am using which has no published ripple settings, and I don't have an oscilloscope, or the knowledge to use one.

 

[kr0gg]

Quote:

yes I use this overpriced pci USB controller.  Not sure if any others can modify these settings.
 
To conclude my comparison I seem to find the ground setting to sound the best - best soundstage and instrument placing/imaging.  Strangely one of the USB ports detects the DI as "TE7022" while the other just shows "USB audio device."  Not sure how this might relate to the driver installed.
 
EDIT: unfortunately this setting does not allow me to test the OMILEX isolator with the DI (OMILEX needs to draw power from USB port).  One last variable to test is whether the computer's main PSU sounds better than the external PSU I am using which has no published ripple settings, and I don't have an oscilloscope, or the knowledge to use one.

WOW. i've seen it before and i'm pretty interested in it. can you share your experience with this controller?
 
as about the OLIMEX and power - i actually got best SQ out of Digital Interface when i connected DI's PSU to the OLIMEX dongle, not to the DI.
that way it will feed BOTH of them (isolator+DI) with clean power.
you should try this too
 

 

[leeperry]

as about the OLIMEX and power - i actually got best SQ out of Digital Interface when i connected DI's PSU to the OLIMEX dongle, not to the DI.
that way it will feed BOTH of them (isolator+DI) with clean power.
you should try this too

I've spoken w/ the designer of the Olimex dongle: whatever PSU you'll input on the external connector will always go through a switched mode regulator(BF9001F). Isolation doesn't come for free ripple-wise, it never will. He advised to tap the PSU directly onto the device USB connector pins.
 
I don't recall who feeds the 4160 chip, though...but I'm quite sure it's the computer, you may wanna ask Olimex for confirmation

 

[drez]

erm I just compared the tx-usb to the motherboards USB 2.0 port.  Motherboard was more focussed imaging - better treble tonality, less bass bloom.  
 
wow don't I feel stupid now.  I now have a $300 usb controller which sounds worse than the onboard controllers.
 
wait ignore the above - the tx-usb sounds great - want to buy it?

 

[kr0gg]

Quote:

I've spoken w/ the designer of the Olimex dongle: whatever PSU you'll input on the external connector will always go through a switched mode regulator(BF9001F). Isolation doesn't come for free ripple-wise, it never will. He advised to tap the PSU directly onto the device USB connector pins.
 
I don't recall who feeds the 4160 chip, though...but I'm quite sure it's the computer, you may wanna ask Olimex for confirmation

ok. i've just sent him a letter.
i'm pretty sure that with a PSU connected the isolator is fed too (at least, my ears tell me so )
but we'll see

 

[kr0gg]

Quote:

erm I just compared the tx-usb to the motherboards USB 2.0 port.  Motherboard was more focussed imaging - better treble tonality, less bass bloom.  
 
wow don't I feel stupid now.  I now have a $300 usb controller which sounds worse than the onboard controllers.
 
wait ignore the above - the tx-usb sounds great - want to buy it?

never expected _such_ an answer
i remember leeperry dreaming about one of those cards...
 

 

[drez]

yeah really disappointed here - I kind of didn't bother to compare it to onboard before.
 
 I just did quick comparison PSU into DI vs into OMILEX - kind of prefer into DI but that impression might be a bit premature.  Will compare vs straight USB later.  Critical listening is tiring.
 
BTW is anyone using the Ploytec drivers?

 

[kr0gg]

Quote:

yeah really disappointed here - I kind of didn't bother to compare it to onboard before.
 
 I just did quick comparison PSU into DI vs into OMILEX - kind of prefer into DI but that impression might be a bit premature.  Will compare vs straight USB later.  Critical listening is tiring.
 
BTW is anyone using the Ploytec drivers?

i am using Ploytec drivers.
they _seem_ to sound better (especially if set to minimal latency, i.e. 6ms)
but they simply "hang" the DI if you transmit anything at 48khz and they also glitch if you start playing a second file with different frequency without stopping the first one.
i'm using windows7 and currently i've managed a nice thing: when i plug the DI in one usb-controller it automatically uses Ploytec drivers, when i plug it in another one - it automatically uses default Windows drivers.
but...due to the long process of turning the DI on and off (at least mine can initialize only after a few minutes after being turned off) it's impossible to really A/B those drivers.
 
 

 

[Gilst]

Quote:

Have you tried USB -> Isolator -> DI?

 
No, the 40-50 bucks shipping just doesn't feel right to me.

 

[sfogar]

Hi,
 
I received the 'high bandwidth' firmware for my Buffalo II Dac, installed it on the dac and tried the usual tests.
 
Well, the dropouts are still there, no change, probably a bit worse...
 
So I tried to use the upsampling to 96khz and the dropoouts are gone !
 
So definetly there's a problem with Sabre Dacs with no upsampling which is 'cured' by the upsampling, I don't know how ...
 
This is not a complete solution cause I didn't want to upsample but ...
 
AF

 

[kr0gg]

Quote:

 
No, the 40-50 bucks shipping just doesn't feel right to me.

Olimex isolator shipping is pretty cheap actually

 

[leeperry]

So definetly there's a problem with Sabre Dacs with no upsampling which is 'cured' by the upsampling

Kingwa confirmed clock synchronization problems on their 9018 chip indeed: http://www.head-fi.org/forum/thread/516886/audio-gd-nfb-11/570#post_7173071
 
Mandatory ASRC looked dubious anyway.