• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Computer Activity Can Impact DAC Performance

Opus111

Addicted to Fun and Learning
Joined
Mar 2, 2016
Messages
666
Likes
38
Location
Zhejiang
Fact check - the adjustment's 3 times per second, measurement is 30 times per second.
 

jkorten

Member
Joined
May 22, 2017
Messages
24
Likes
12
Fascinating post amirm! I've been listening to different computer OS's and different players etc. I did finally pay for (minimum) and download audiophile linux (http://www.ap-linux.com/) and it does not make a small difference. It makes a HUGE difference! And the less powerful your computer, the more impressive the change is.

Loading in AP linux is picky (you must follow ALL instructions and in detail) but in the end it isn't that hard.

Maybe you can post the difference between using plain OS-X, Windows, Linux and AP-Linux? Or Windows and Linux?

Is your ADC running on the computer's clock too? Or is it running on it's own clock?

I think the weak link is in the feeding of DACs. Unbelievable to me that these DACs don't seem to buffer and reclock the data once it is transmitted, but this seems to be the only explanation for the kind of frequency spread and jitter you are reporting.
 

Fitzcaraldo215

Major Contributor
Joined
Mar 4, 2016
Messages
1,440
Likes
632
Careful. Amir was referring to only one particularly schiity DAC, though some others may be like that. Many other DACS are not as schiity, though they may have other issues of isolation. Still other USB DACS are well isolated from computer activity. I hear absolutely none on my galvanically isolated USB Exasound DAC when using a straightforwarward, unmodified PC with standard USB cables, for example.

I believe that, if there is a problem, it is an engineering issue at the DAC end. Modifying, replacing the computer, its innards or OS is a fool's errand, IMHO, except perhaps for quieting its acoustic noise output from fans, etc.
 

Johnny2Bad

Member
Joined
Dec 22, 2017
Messages
42
Likes
28
I recently came across an issue that previously I was unaware of. Apparently, RF emissions from motherboards are sometimes an issue with compliance body certification. A solution employed by the MB mfr is to adjust the clock frequency (typically in a +/- 5% range) which changes the emissions frequency, and can lead to compliance certification. The setting can be adjusted in BIOS. If the MB was shipped with the adjusted clock, it makes clock sync for audio problematic, and adds jitter (or at least that is the claim).

There is a Wikipedia page on the subject (not particularly helpful, but some). I ran across it at the AQXOX website (manufacturer of audio components, Germany)

" ...
Optimize PC´s BIOS for AUDIO-Streaming - switch off the CLOCK-spreading / clock spread option / spread spectrum - less Jitter

If you enter the PC´s BIOS you will find somewhere in the voltage/clock options the CLOCK Spreading or Clock Spread Spectrum or SPREAD Spectrum option. This function is actually meant for EMI (electromagnetic radiation) test during the CE-test of the Mainboards. If the Motherboards during the CE - certification-tests for high radiations in a certain frequency range produce to high distortions, caused by possible overlay of frequencies (also harmonic waves) and thus a reinforcement (constructional interference) of the radiated signal. This radiation behavior can be changed by letting the systemclock not longer precisely work on a specific frequency but changing its frequency very fast. The clock-frequency fluctuates and thus flattens the spikes because SPREADING it on a broader frequency band. The BIOS adjustment possibilities can look like: 0.25%; 0.5%; 1.5%; Enabled and Disabled. Disabled means the Spreading is switched off. This is the best option, however you should restart it, if radio or TV are disturbed. Disabled improves also the PC performance. Sonically effects: What the CLOCK Spreading now to do with the sound? Very simple, an instable/varying system clock produces JITTERS, all the same whether the audio data are transmitted via USB or over an internal or external sound-port or soundcard by SPdif or TOSlink.

See what wikipedia says about one of the effects of this spreading - the clock skew
http://en.wikipedia.org/wiki/Clock_skew
..."

From: http://www.aqvox.de/tech.html

If present, this would be a model-by-model variation, so would require a check to determine if it's an issue for your specific hardware.

Note: First Post
I didn't know of this forum previously, was pointed to it by a UK reseller, Keith at purite audio (~co.uk)
 
Last edited:

Purité Audio

Master Contributor
Industry Insider
Barrowmaster
Forum Donor
Joined
Feb 29, 2016
Messages
9,051
Likes
12,150
Location
London
Welcome Gordon, I hope you thoroughly enjoy ASR.
Keith
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,393
Location
Seattle Area
My welcome also Johnny2Bad.

Spread spectrum techniques are quite effective for passing emissions standards where "jitter" is put to good use. Let's say we have a peak at 500 Mhz created by a clock at the same frequency. If we apply random jitter of +- 1 Mhz, it will take that peak and spread across a 2 Mhz region now. Since the total energy remains the same, the peak by definition goes down. Of course jitter created must be within the range that doesn't disrupt the timing of the circuit and cause it to fail. This technique is used more often in switchmode power supplies where there is no consequence in timing being different by that manner.

Here is a nice picture that shows the effect (it is not always this effective) from Maxim (chip company):

3512Fig02.gif
 
Last edited:

Johnny2Bad

Member
Joined
Dec 22, 2017
Messages
42
Likes
28
Makes perfect sense. The articles I read said that +/- 5% was the problem area. However from your graph I wouldn't want to see anything too near those 91 and 93.7 Mhz values. Also that you could remove the spreading in some deep BIOS settings, recommending you check for interference with nearby devices after you do so.
 

FrivolsListener

Active Member
Joined
Feb 14, 2018
Messages
114
Likes
19
Definitely not a problem with WMP. My team at Microsoft wrote it :).

Um. That's not a great claim to fame. If you like, we can discuss specific issues, but given the date of your post, I expect that it's water under the bridge at this point.

ObOnTopic: No, that doesn't speak well of that DAC. But the graphs are gone, so I'd have to see them again.

Feel free to PM me on either subject. Cheers!
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,393
Location
Seattle Area
Um. That's not a great claim to fame. If you like, we can discuss specific issues, but given the date of your post, I expect that it's water under the bridge at this point.
Didn't say it as a claim to fame but rather, having specific knowledge of how it works.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I just installed Volumio on a Raspberry Pi 3 B+ and connected USB to SPDIF bridge to it. I'm using Linn player Kinsky to stream via UPnP to it and it sounds fine. The same thing when using BubbleUPnP from Android phone. I contacted Volumio developer and he told me that it should work with practically all USB DACs on the market.

Tried Foobar as well but its UPnP plugin (which is still in beta) cannot send 24 bit output, when you set it to 24 bit it sends 32 bit instead which my bridge doesn't support. I was also not able to get it to send SACD (DSD64) correctly. Foobar was also not able to see my Denon HEOS system in the other room as a renderer while Kinsky and BubbleUPnP recognised it and are playing to it without problems.

All 3 players were able to fetch music from my Synology NAS running UPnP Media Server.

Unfortunately I can't measure how well it performs, but I put more trust in Raspberry Pi 3 B+ running only Volumio than my notebook which does all sort of things while playing to USB.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Here is what made me do this: I mostly listen to the music in the living room, but my notebook is quite far from my audio equipment, more than 5m of USB cable would be needed, which would impact playing some HiRes material, not to mention make my wife to throw the vble and me out of the flat. :D

So obviously, another solution was needed. I tried using USB to Ethernet server Silex DS-510 and it worked, but really flawlessly. Every time when player on my notebook (any player, they were all behaving the same) was changing bitrate via ASIO or WASAPI, as next song was having different bitrate than the one playing, there was a period of 5-10 seconds of sounds as if player was fastforwarding audio tape - at first I found it was amusing but later it started to annoy me. It seems that Silex does this with all Singxer USB DACs/bridges probably due to a mismatch between driver and Silex firmware. I tried using the generic Windows 10 driver which didn't manifest that behavior but I was experiencing random dropouts, say 1 sec of dropout each 5-10 minutes. Here I would also like to mention that my WiFi network is pretty good . I can, for example watch movies with bitrate higher than 40Mbps without any issues.

So, I start to Google for alternative solution and found this monster: https://www.linn.co.uk/hifi-separates/network-music-players/klimax#klimax-ds

I got cooled down the moment I saw the price (15800 GBP in UK, roughly $21400) . Yep, no typo here, I checked the figure twice. Measurements done by Stereophile can be found here for anybody who think they can justify the price.
https://www.stereophile.com/content/linn-klimax-ds-network-da-processor-measurements

As I don't have $21400 to spend on DAC I turned my head to my Singxer USB bridge and I came to conclusion that if I spend $50 on Raspberry Pi 3 B+ (power supply included) I can make two of them work together as an UPnP renderer or, if you want, Network Media Player. Pursuing the even greater irony, I bravely downloaded Linn music player that goes with the Klimax DS, et voila - it works perfectly! :D
 

saturnaal

Member
Joined
Jul 24, 2019
Messages
46
Likes
70
Location
United States of America
For anyone wondering, this is still an issue.

The on-board audio in my motherboard (touted as "premium audio" by the motherboard manufacturer), is remarkably susceptible to interference. GPU load, USB peripheral activity (mouse, removable drives), and drive R/W activity could all be heard rather loudly through it and repeatedly confirmed with testing. With some applications, especially ones which place a high load on the GPU (PC consuming up to 450 watts), the noise level was so high it made any listening impossible to enjoy. It would drown out ambient sounds in games, mask the sound of footsteps, and overwhelm quiet passages in music.

I had originally assumed that adding an external DAC between the PC and amp would address the issue, but after getting a popular USB DAC I found the noise to still be present, just at somewhat reduced levels. So the electrical noise must be carried over the USB connection and still making it into the signal. It wasn't until I switched to S/PDIF that I was finally rid of the interference. Now, the measurements on this site demonstrated that the DAC I've been using produces potentially audible levels of distortion, so I've got a new one on the way. It will be interesting to see if their USB implementation actually eliminates the problem or if I'll be required to continue using S/PDIF.
 

MediumRare

Major Contributor
Forum Donor
Joined
Sep 17, 2019
Messages
1,949
Likes
2,275
Location
Chicago
Any experience if using BT to send data to a DAC is any different in this regards than USB? Is there an easy way to test this ourselves such as with REW?
 

Bathrone

Member
Joined
Jun 28, 2022
Messages
17
Likes
5
Location
Australia
Schiit have been making noise about their proprietary development and implementation of their own USB kit. It wouldnt be super hard to come up with specific "problem child" stream inputs in a test harness to test this sort of thing. I think in reviews, there should be more "fuzz" testing and more tests focused on known real world issues. Im yet to see a review establishing through testing that Schiits in house USB solution, isnt shit
 
Top Bottom