INTERACT FORUM

Please login or register.

Login with username, password and session length
Advanced search  
Pages: [1]   Go Down

Author Topic: USB Audio and high CPU usage?  (Read 7506 times)

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
USB Audio and high CPU usage?
« on: May 16, 2013, 08:15:43 am »

I don't have any experience with USB audio, but I'm looking to buy an external DAC for my computer.

However something I have noticed is that, particularly when Media Center is analyzing multichannel DSD tracks, my CPU usage hits 100% and the mouse cursor starts skipping around. (this doesn't always happen at 100%, but does when analyzing DSD files)

Firstly - has anyone else noticed this behavior?
Secondly - if you have noticed this happening, does it mean that USB audio will start skipping?

Even though the mouse cursor starts skipping around, audio playback is not interrupted in my current setup - but I'm not using USB.
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5234
  • "Linux Merit Badge" Recipient
Re: USB Audio and high CPU usage?
« Reply #1 on: May 16, 2013, 08:45:19 am »

I don't have any experience with USB audio, but I'm looking to buy an external DAC for my computer.

However something I have noticed is that, particularly when Media Center is analyzing multichannel DSD tracks, my CPU usage hits 100% and the mouse cursor starts skipping around. (this doesn't always happen at 100%, but does when analyzing DSD files)

Firstly - has anyone else noticed this behavior?
Secondly - if you have noticed this happening, does it mean that USB audio will start skipping?

Even though the mouse cursor starts skipping around, audio playback is not interrupted in my current setup - but I'm not using USB.

How many channels of DAC output do you need?  If you only need a stereo USB DAC, you have plentiful and surprisingly cheap options (of which I can recommend several).  If you need multichannel, check out this thread: http://yabb.jriver.com/interact/index.php?topic=55607.0. or this one: http://yabb.jriver.com/interact/index.php?topic=77508.0 . The bottom line is that there are fewer options (although there are still several) and costs are much higher.

I've never had USB audio skip with any of the DACs I've used (once I'd set the buffers right), even under fairly heavy processor load.  A humorous anecdote: I once had a system lock up that froze the OS for about five minutes (no response to Ctrl Alt Delete, etc.), and then there was a BSOD.  Right up until the BSOD, although the OS was completely non-responsive, the USB audio continued playing.  Your mileage may vary, and different DACs have different buffering structures, but I've never experienced USB-audio skipping that wasn't related to a too small buffer setting.
Logged

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
Re: USB Audio and high CPU usage?
« Reply #2 on: May 16, 2013, 10:09:10 am »

Thanks for the links, I'm looking for a stereo DAC which also has a good headphone output. (<1 ohms, good channel tracking at low volumes)
I'm unsure about whether I care for things like sample rates above 24/96, and if I need native DSD support or not.
I've read articles which show that 24/96 should be better than DSD - but people on forums with DSD DACs generally seem to report that DSD sounds better.  ?

Have you also experienced things like the cursor skipping around while playing audio via USB though? High CPU usage is one thing, but CPU usage high enough to cause that is another.
I've just checked though, and in 18.0.184 I'm able to analyze multichannel DSD without the cursor skipping everywhere, so maybe it's not an issue any more.
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5234
  • "Linux Merit Badge" Recipient
Re: USB Audio and high CPU usage?
« Reply #3 on: May 16, 2013, 11:01:53 am »

Thanks for the links, I'm looking for a stereo DAC which also has a good headphone output. (<1 ohms, good channel tracking at low volumes)
I'm unsure about whether I care for things like sample rates above 24/96, and if I need native DSD support or not.
I've read articles which show that 24/96 should be better than DSD - but people on forums with DSD DACs generally seem to report that DSD sounds better.  ?

Have you also experienced things like the cursor skipping around while playing audio via USB though? High CPU usage is one thing, but CPU usage high enough to cause that is another.
I've just checked though, and in 18.0.184 I'm able to analyze multichannel DSD without the cursor skipping everywhere, so maybe it's not an issue any more.

The best low price option for a stereo DAC w/ headphone amp, in my opinion is the Fiio E17.  There's a thread about it over here http://yabb.jriver.com/interact/index.php?topic=78637.0.  I've had one for about a year and like it quite a lot.  It has low output impedance, a reasonable amount of power, and very low noise.  It also can be fitted with an $8 attachment that will turn it into a line source.  

The best performance amp/dac combo I've encountered (that still doesn't cost a mint) is the O2/ODAC combo, which has some pretty incredible specs and is under $300:  http://www.jdslabs.com/item.php?fetchitem=48.  You can also get a standalone ODAC for $150 (no headphone amp).  I run my headphones off of an O2, and demoed an ODAC a while back.  I really wish there was something similar to the ODAC in the eight channel realm (comparatively cheap, low noise, and USB operated), but there really isn't.  The cheapest 8-channel DAC I've seen with comparable performance is about $750.  But that's my problem, not yours  ;D

I don't think either of those DACs natively support DSD, but I don't have much experience using DSD.  I believe that both support 24/96.

I haven't ever had a CPU condition that caused cursor skipping while using USB audio, so I can't comment on that.  I don't use USB audio for my speakers, I just use it with headphones for music, or for video games.  I've never had non-buffer induced stutter for music.  Now that I think of it, I've had very occasional sound stuttering in video games, but it was typically when the game itself was also choking (graphically), and I have drop outs with my PCI soundcard as well under those conditions (i.e. it's likely that the game just stops feeding the buffer).  So my experience has been that USB audio is (in my PC) not any less reliable than in-PC audio.
Logged

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
Re: USB Audio and high CPU usage?
« Reply #4 on: May 16, 2013, 11:11:16 am »

Thanks again. The O2+ODAC is one of the main options I had been considering... I'm put off by the fact that it's a DIY device and there's a mess of cables coming out the front of it though.
I can't seem to find anything that measures better until you are looking at almost 10x the price though, so maybe I should just buy one.

I'm not sure why the cursor skips. It happens with both my Razer mice (one at 500Hz, the other at 1000Hz) and my Logitech Anywhere MX. (which I guess is just polled at 125Hz)
Happened on Windows 7 and Windows 8. In fact, I can't think of a PC I've used where it doesn't happen when something is really stressing the CPU. I figured it was just a Windows quirk. (a very annoying one at that)
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5234
  • "Linux Merit Badge" Recipient
Re: USB Audio and high CPU usage?
« Reply #5 on: May 16, 2013, 11:44:49 am »

Thanks again. The O2+ODAC is one of the main options I had been considering... I'm put off by the fact that it's a DIY device and there's a mess of cables coming out the front of it though.
I can't seem to find anything that measures better until you are looking at almost 10x the price though, so maybe I should just buy one.

That's where I wound up.  I can tell you (based on extended use) that the O2 I have (assembled by JDSLabs) is rock solid.  Ordering it pre-assembled gets you the benefit of a warranty. My interaction with the ODAC was also very positive, but briefer.  There are copious measurements available of both over at the designers blog (nwavguy), along with measurements of other affordably priced DACs.  I'm not a fan of the front cables either, but the device is the size of a paperback book, so it's easy to conceal (I just turn it around).  

And I don't mean to pooh pooh the e17 either.   It's currently feeding my O2 and sounds great.  

Quote
I'm not sure why the cursor skips. It happens with both my Razer mice (one at 500Hz, the other at 1000Hz) and my Logitech Anywhere MX. (which I guess is just polled at 125Hz)
Happened on Windows 7 and Windows 8. In fact, I can't think of a PC I've used where it doesn't happen when something is really stressing the CPU. I figured it was just a Windows quirk. (a very annoying one at that)

I occasionally get cursor skipping with wireless mice when my mouse battery is dying or there's something physically obstructing the wireless pathway.  I've never had mouse skipping with a wired USB mouse, and the wireless mouse skipping is so infrequent that it's never come up when I happened to be listening to headphones.  I don't really do much that maxes out the CPU load (when I can still see a mouse cursor) on my current PC though (I've never observed the CPU load get much above 50% except during a fault condition, although it may get higher than that when it's hard to have a look at, i.e. during video games), so maybe my mice would skip under those circs, and it just hasn't come up yet.
Logged

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
Re: USB Audio and high CPU usage?
« Reply #6 on: May 16, 2013, 12:18:39 pm »

That's where I wound up.  I can tell you (based on extended use) that the O2 I have (assembled by JDSLabs) is rock solid.  Ordering it pre-assembled gets you the benefit of a warranty. My interaction with the ODAC was also very positive, but briefer.  There are copious measurements available of both over at the designers blog (nwavguy), along with measurements of other affordably priced DACs.  I'm not a fan of the front cables either, but the device is the size of a paperback book, so it's easy to conceal (I just turn it around).  
The issue for me is that ReplayGain doesn't work well enough that I can leave volume alone 100% of the time (I think R128 analysis might) and I was unable to convince Matt to either allow ReplayGain/R128 to process videos, or at least provide an option that levels the volume between Stereo and 5.1/7.1 downmixing. (the more channels you have, the quieter the downmix needs to be)

So I would need to have the O2 within reach - and the ODAC input is on the rear of the unit anyway, so you can't just turn it around to hide the cables.

This is actually my biggest issue with using a DAC. With an AVR, it uses HDMI CEC so that I can control everything through my TV remote as I do now, and HDMI supports automatic lip-sync correction to adjust for delays. But an AVR is a huge unit, and the headphone outputs suck.

With an external DAC that is connected to the PC, I lose the ability to control volume via the TV remote (unless they add support for libCEC) and I lose the automatic lip-sync correction. (and I was unable to convince Matt to add the option to adjust the delay depending on whether a video is being played fullscreen or windowed - as my TV has a variable delay)

I occasionally get cursor skipping with wireless mice when my mouse battery is dying or there's something physically obstructing the wireless pathway.  I've never had mouse skipping with a wired USB mouse, and the wireless mouse skipping is so infrequent that it's never come up when I happened to be listening to headphones.
It's not a wireless issue - only the Logitech mouse is wireless, and it's linked to CPU usage rather than range/signal/battery life.
I do think it must either be Windows, or coincidentally an issue with all the mice I have used across multiple computers. (I can't rule it out)
I don't really do much that maxes out the CPU load (when I can still see a mouse cursor) on my current PC though (I've never observed the CPU load get much above 50% except during a fault condition, although it may get higher than that when it's hard to have a look at, i.e. during video games), so maybe my mice would skip under those circs, and it just hasn't come up yet.
The main thing that was causing it to happen for me was when Media Center was analyzing multichannel DSD files, or multiple DSD files at once. (with auto-import, it seems to do one at a time, manually analyzing I do four at once)
It also happens when using sacd_extract to convert DST compressed DFF files to DSF. (if I didn't realize the disc used DST compression)
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5234
  • "Linux Merit Badge" Recipient
Re: USB Audio and high CPU usage?
« Reply #7 on: May 16, 2013, 12:22:10 pm »

The issue for me is that ReplayGain doesn't work well enough that I can leave volume alone 100% of the time (I think R128 analysis might) and I was unable to convince Matt to either allow ReplayGain/R128 to process videos, or at least provide an option that levels the volume between Stereo and 5.1/7.1 downmixing. (the more channels you have, the quieter the downmix needs to be)

So I would need to have the O2 within reach - and the ODAC input is on the rear of the unit anyway, so you can't just turn it around to hide the cables.

This is actually my biggest issue with using a DAC. With an AVR, it uses HDMI CEC so that I can control everything through my TV remote as I do now, and HDMI supports automatic lip-sync correction to adjust for delays. But an AVR is a huge unit, and the headphone outputs suck.

With an external DAC that is connected to the PC, I lose the ability to control volume via the TV remote (unless they add support for libCEC) and I lose the automatic lip-sync correction. (and I was unable to convince Matt to add the option to adjust the delay depending on whether a video is being played fullscreen or windowed - as my TV has a variable delay)


With my speakers, I have my DAC hooked directly to block amplifiers that have no volume control.  I just use JRiver's internal volume, which works with a mouse or a remote (my JRiver remote works on my TV too for the most part).  I haven't generally missed having a hardware volume control.  I presume you currently have internal volume disabled, but is there some reason that wouldn't work in your application? (i.e. you could set the O2 hardware control to your maximum listening volume, and then use internal volume to attenuate)

[I don't know enough about lip sync correction to offer a suggestion there]

EDIT: Also, I just noticed that the combo O2/ODAC doesn't appear to have a line out (which the separate ODAC does have), so that's something to think about before buying the combo if you plan on using the DAC for other than headphone listening.
Logged

gtgray

  • Galactic Citizen
  • ****
  • Posts: 261
Re: USB Audio and high CPU usage?
« Reply #8 on: August 23, 2013, 05:26:09 pm »

The issue for me is that ReplayGain doesn't work well enough that I can leave volume alone 100% of the time (I think R128 analysis might) and I was unable to convince Matt to either allow ReplayGain/R128 to process videos, or at least provide an option that levels the volume between Stereo and 5.1/7.1 downmixing. (the more channels you have, the quieter the downmix needs to be)

So I would need to have the O2 within reach - and the ODAC input is on the rear of the unit anyway, so you can't just turn it around to hide the cables.

This is actually my biggest issue with using a DAC. With an AVR, it uses HDMI CEC so that I can control everything through my TV remote as I do now, and HDMI supports automatic lip-sync correction to adjust for delays. But an AVR is a huge unit, and the headphone outputs suck.

With an external DAC that is connected to the PC, I lose the ability to control volume via the TV remote (unless they add support for libCEC) and I lose the automatic lip-sync correction. (and I was unable to convince Matt to add the option to adjust the delay depending on whether a video is being played fullscreen or windowed - as my TV has a variable delay)
It's not a wireless issue - only the Logitech mouse is wireless, and it's linked to CPU usage rather than range/signal/battery life.
I do think it must either be Windows, or coincidentally an issue with all the mice I have used across multiple computers. (I can't rule it out)The main thing that was causing it to happen for me was when Media Center was analyzing multichannel DSD files, or multiple DSD files at once. (with auto-import, it seems to do one at a time, manually analyzing I do four at once)
It also happens when using sacd_extract to convert DST compressed DFF files to DSF. (if I didn't realize the disc used DST compression)
actually you can control volume with E17 with JRiver and a Windows Media Center Remote. I have recently setup a dedicated headphone listening  station. It is a  23" HP Touchsmart, sitting on an end table immediately to the left of my seat on the end of the couch. So I can control the volume at the DAC OSD or via the touch screen using the Theater View Now Playing Volume UP/Down buttons or via a standard WMC remote.
Logged
Pages: [1]   Go Up