[HaVoC-28]

Quote:

I recently bought a DI to work with my NFB-10. But I find that the sound sometimes stopped for a short time (about 0.5s), especially when playing games.
 
Did anyone else have such issue?

drop out ? hum allready seen in this thread i think , Are you using your DI  without resampling ?

 

[somestranger26]

Quote:

I recently bought a DI to work with my NFB-10. But I find that the sound sometimes stopped for a short time (about 0.5s), especially when playing games.
 
Did anyone else have such issue?

 
Try searching. It's about the #1 most discussed thing in this thread. You will never eliminate dropouts in games your computer struggles to run already. Older ones should not have a problem.

 

[HaVoC-28]

 
Quote:

 
Try searching. It's about the #1 most discussed thing in this thread. You will never eliminate dropouts in games your computer struggles to run already. Older ones should not have a problem.

and with another usb/spdif converter no prob ? 

 

[drez]

Upsampling seems to be necessary - also improves imaging, but the DI demands lots of PC tweaking to get the best sound.

 

[melancholic]

Oh, I failed to realize that the problem is named "dropout" in English.
 
When I'm merely listening to music, the dropouts happen every 10 minutes on average. In lucky days there are no dropouts. But when playing games, the gap between dropouts is less than 20 seconds. And I am just playing old games such as Starcraft.
 
Seems upsampling can fix it? I'll have a try.

 

[somestranger26]

I've gotten dropouts with upsampling before.
 
After much optimization to my computer, I only get one every day or two. You probably have a lot of background processes and software running in the background causing the dropouts to be so frequent. Kingwa posted a guide to help with them not far back in this thread, but there's almost endless optimizations you can make. Some changes, such as setting the USB polling rate to 5ms can even improve the sound quality slightly.
 
When playing intensive games like Deux Ex:HR and Battlefield BC2, I get them every few minutes but that's just the CPU trying to juggle rendering the graphics and handling game mechanics while outputting audio.

 

[drez]

Hey somestranger26 can you point me to a good method of changing polling rate?

 

[somestranger26]

http://www.cicsmemoryplayer.com/index.php?n=CMP.07Optimisations All the way at the bottom. I've done most of these optimizations (and gone considerably further) to get dropouts under control. Bonus effect of improving sound quality.

 

[melancholic]

It seems that the dropout problem is almost solved by upsampling to 96KHz.

 

[petemac110]

/img/forum/go_quote.gif
 

Quote:

I'm not sure if this has been discussed within the thread, but has anyone tried the no-change setting without the clock in the socket?
 
I was testing between the stock clock, a Vanguard TCXO and a Tentlabs XO for the upsampling. In the end I preferred without upsampling. Since Kingwa said the clock is not used in the no change setting, I took out the clock. Surprisingly there seems to be some difference in the sonic and I wonder if there's any influence the clock has in the socket and without. I tested all three clock in the socket in the no-change setting. There were perceptible changes in the sound with each clock having a subtle impact on the sound. In the end, I found that no clock in the socket turns out to provide the most transparent and clean output signal to the DAC. Details and ambient information are a lot more noticeable without any clock in the socket. I thought Kingwa said the clock isn't used when not upsampling but it seems to have an effect on the signal.

I tried removing the standard clock today, and noticed that it affected the imaging in my system. Note that I don't have a headphone rig - I have a restored Sansui AU-517 dual mono amp, Squeezebox Touch, audio-gd DI, audio-gd NFB-3 and Dynaudio speakers. When I removed the clock, rather than having a strong centre image for vocals, the vocals were spread out/smeared across the sound stage, and it was easier to locate the speakers. They sounded 'boxier' - with the clock in place, the speakers did a better job of disappearing.
 
I'm not sure what is the cause of this, but for the time being, the standard clock will be left in place in my system. Still tossing up whether to get a different TCXO - Valab, Vanguard (either 1ppm or 0.3ppm gold) or audio-gd. I've also got a  few hex inverters arriving late next week for experimentation.
 

 

[somestranger26]

Wow. I guess removing the clock does make a difference (improvement from what I'm hearing). Seems pretty strange that it changes the sound when the clock "isn't used" when not upsampling.

 

[petemac110]

 
Quote:

Wow. I guess removing the clock does make a difference (improvement from what I'm hearing). Seems pretty strange that it changes the sound when the clock "isn't used" when not upsampling.

 
Yeah, tell me about it!
 
I should say that's not a massive difference - to my ears, the overall sound quality and definition was virtually the same with the clock installed vs the clock removed. The change in the imaging/sound stage was the really noticeable part.
 
 

 

[Currawong]

There are measurements showing the NFB-12 with a lower noise floor and more dynamic range with 96k up-sampled music, so it might be an idea to get the better clock and leave the DI set to up-sample to 96k.

 

[Baird GoW]

I personally like the clock more... I feel that it takes away from the impact of the bass on the HE-6. I can tell without the clock there is more precision and more detail but to me the trade off is not worth it imo.

 

[petemac110]

Quote:

. Still tossing up whether to get a different TCXO - Valab, Vanguard (either 1ppm or 0.3ppm gold) or audio-gd. I've also got a  few hex inverters arriving late next week for experimentation.
 

Further to my above comment, I note that the DI provides 3.3V for the standard clock. However, most of the TCXOs seem to be 5V (Vanguard 1ppm and 0.3ppm, Valab 1ppm etc).  The information on the website for the tentlabs XO suggests a power supply of 5V for the XO, but mentions a range of 3.3v to 5v in the specs.
 
Do the Vanguard or Valab TCXOs definitely work OK with the DI? It would seem that a few people have tried them, but there's not a lot of follow-up comment indicating that they work fine aftert plenty of use.
 
So... if you're using a 5V specc'd TCXO in your DI, please post with an update!
 
Worst case scenario I guess I could always just grab an upgrade 3.3V TCXO direct from audio-gd.