A curious discussion about sample rates in Windows vs Soundcard CPL

  • Thread starter Thread starter mrk
  • Start date Start date

mrk

mrk

Man of Honour
Joined
18 Oct 2002
Posts
105,116
Location
South Coast
Okay so the past few days I've been casually browsing various audio forums in part trying to wet my lips on more Senn HD800 info but also Xonar info and i cam across a few posts relating to the best usage of the sample rate settings.

Now, we know that there are 2 places you can set the sample rate conversion to be output, in the Advanced tab of the Speakers properties dialogue and of course in the soundcard's own control panel which tells the card's onboard processors to use "x" sampling rate.


The former is software driven so if you set it to 96000Hz then all sounds will be upsampled by Windows to that rate then sent to the soundcard which will again do some conversion.

Now the discussion was that because the sampling rate converters on the soundcard are of much higher quality than the obvious software ones built into Windows Vista and 7 (XP need not apply because it does not operate at such a high level) that it would be logical and best practice to just leave the Windows sampling rate output to 24bit 44100Hz (the default being 16bit 44100Hz) and then set the soundcard's control panel output to a higher sampling rate such as 96000Hz (96KHz).

Does anyone else follow this train of thought? it makes sense, you'd want the samplling rate to be upconverted at a hardware level if you've got something like an XFi or a Xonar and they say setting the sampling rate to 96KHz in Windows could cause jitter and/or quality loss in extreme cases.

This may explain why in some (not all) movies with AC3 or DTS audio I sometimes heard some crackling on low vocal tones (observed using a variety of decoder filters so not a codec issue) and since setting Windows sampling rate to 24bit 44100Hz and keeping the Xonar CPL to 96KHz as usual has resulted in no such crackles on the same media.

Anyone wish to share their thoughts on this?
 
Last edited:
I have my W7 default set to 44.1KHz at 24bit, then the Xonar STX set to 192Khz, I have no artifacts or crackling in the sound at all. It all sounds great. There is something in the drivers for the STX that messes with the STX output when you raise the window default higher that 24bit 44.1Khz, not that is matters because the soundcard hardware is far superior anyway.
 
Yeah that goes in line with what the topic is about, I did notice what you say too, if you set Windows rate higher than 44.1 then load Xonar's control panel notice the equaliser bars don't become active when you play sounds so Windows is stopping the Xonar's driver from doing what it needs to.

And so we return to the "let the hardware do it" :D
 
Yeah that goes in line with what the topic is about, I did notice what you say too, if you set Windows rate higher than 44.1 then load Xonar's control panel notice the equaliser bars don't become active when you play sounds so Windows is stopping the Xonar's driver from doing what it needs to.

And so we return to the "let the hardware do it" :D


Yeah I noticed that too, definitely Hardware all the way.
 
Not sure setting 44100Hz as the base in Windows is a great idea if you watch DVD content on your PC, seeing as the audio is 48000Hz. You'd be down-sampling the content on the software mixer, reducing the quality in the process, then up-sampling that to 96000Hz on the sound card for no gain whatsoever.

I'd also dispute as to whether up-sampling on the sound card itself provides any real quality benefit anyway, seeing as Vista/7's mixer is 32-bit floating point with little to no distortion at all. Perhaps it'll save some CPU resources if you use higher levels of up-sampling on the card itself as opposed to software, but even then it's a bit of a non-issue these days.
 
Not sure setting 44100Hz as the base in Windows is a great idea if you watch DVD content on your PC, seeing as the audio is 48000Hz. You'd be down-sampling the content on the software mixer, reducing the quality in the process, then up-sampling that to 96000Hz on the sound card for no gain whatsoever.

I'd also dispute as to whether up-sampling on the sound card itself provides any real quality benefit anyway, seeing as Vista/7's mixer is 32-bit floating point with little to no distortion at all. Perhaps it'll save some CPU resources if you use higher levels of up-sampling on the card itself as opposed to software, but even then it's a bit of a non-issue these days.


Dont really see what you are implying here? I have W7 set to 24bit 44.1Khz, my STX set to 24bit 192Khz and the sound output is 24bit 192Khz to my Amp, I checked the amp and it confirmed it, Doesnt matter whether I play CD, DVD, or iTunes etc, sample rate is upsampled to 192Khz regardless, and the quality is exceptional. You can certainly hear the difference when I change to another sample rate.
 
Unless you're sending audio directly to the sound card (via ASIO or OpenAL), then all audio passes through the Windows audio mixer first. If you set said mixer to 44100Hz, then it looks like this when playing DVD content:

DVD (48000Hz) > Windows audio mixer (down-sampled to 44100Hz) > Xonar (up-sampled to 96000Hz or whatever)

Whenever you down-sample, audio data is lost and quality reduced. It isn't restored by up-sampling. For that reason, you want the baseline sample rate in Windows to be at least equal to the sample rate of the content you're playing.
 
Unless you're sending audio directly to the sound card (via ASIO or OpenAL), then all audio passes through the Windows audio mixer first. If you set said mixer to 44100Hz, then it looks like this when playing DVD content:

DVD (48000Hz) > Windows audio mixer (down-sampled to 44100Hz) > Xonar (up-sampled to 96000Hz or whatever)

Whenever you down-sample, audio data is lost and quality reduced. It isn't restored by up-sampling. For that reason, you want the baseline sample rate in Windows to be at least equal to the sample rate of the content you're playing.


Which at 24bit 48Khz it is, I see what you mean. Its my fault for typing 44.1khz. I have W7 setting at 24bit 48000hz and my STX at 192000hz so all should be ok?

How would I go about bypassing the W7 audio mixer and send everything directly to the soundcard.
 
Last edited:
Whenever you down-sample, audio data is lost and quality reduced. It isn't restored by up-sampling. For that reason, you want the baseline sample rate in Windows to be at least equal to the sample rate of the content you're playing.

I have a lot of respect for your wording - you just explained in one small paragraph what took me over 10 posts to say haha!

The only thing I would say is that I personally notice a difference if the 'baseline' sampe rate in Windows is anything lower than 96kHz / 24-bit (irrespective of what sampe rate is set in the audio driver panel of your interface). As I do all my DAW work at this samplerate anyway, it is a no brainer however I will hold my hands up in the air and say I can't tell the difference between 96kHz and 192kHz....well not at least anymore. I think my hearing has finally dipped off as I was working in the studio yesterday and can't hear above 11kHz (I used to be able to hear up to 16kHz). This is a sad reflection of the line of work I am in (or just the natural leveling off of the frequency range which all adults experience).
 
Ah yes of course I never thought about DVDs for example so indeed re-visiting this it looks like 24bit 48KHz is the baseline that should be set in Windows.

Worth noting though that setting the Windows sample rate above 48KHz stops the soundcard's control panel from taking over as observed by some of us above where the EQ bars don't become active any more so looks like 48 is the max.

Interesting stuff!

Also ASIO, I wasn't aware that ASIO completely worked in Windows Vista/7 ? I would not mind trying a compatible ASIO plugin for Winamp 5.5 if anyone knows of one?

Edit*
I remember now why I stopped using ASIO out back in the day.

I tried MPC Audio Renderer and when watching a movie the bypass takes over the soundcard so no other sounds in other apps will work other than in that instance of MPC-HC.

The same went for ASIO when I used to have it, it seems to take exclusive control over the soundcard and only let other apps play sound when you stop movie playback.

Edit 2*
I fixed it by disabling exclusive mode control here:

modecontrol.jpg


Edit 3*
Okay this is interesting, Settings exclusive mode control works but when you play a video in MPC-HC and right click > filters to see what filters are being used you can notice something interesting:

1: With Exclusive Mode Control unticked in Windows the active audio filter being used is Default DirectSound Device even though MPC Audio Renderer is enabled in the output settings page. This allows other sounds in Windows and apps to play during media playback in MPC-HC.

2: With Exclusive Mode Control ticked the active filter is MPC Audio Renderer but no other sounds in any other app will play until you stop media playback in MPC.
 
Last edited:
Worth noting though that setting the Windows sample rate above 48KHz stops the soundcard's control panel from taking over as observed by some of us above where the EQ bars don't become active any more so looks like 48 is the max.

Thing is though I would argue the difference in sonic quality between 48kHz and 96kHz (irrespective of the word depth) is more substantial than not having access to the Xonar's DSP settings and EQ. However, I am using a hi-fi seperates system so access to hardware EQ on the preamp is a given. Likewise, I thought the DSP settings sounded awful.

I notice a stellar difference between the lower samplerates and higher ones once set in Winblows although I spend most of my day recording, mixing or practising my production skills so what might sound substantial to me, could sound insignifcant to you guys. I liken it to the difference between a 192kbps MP3 and a 320 one - its the artifacting of the higher range of the frequency spectrum (presumably due to aliasing).

Do any of you lot notice a difference? Please help I literally think my heads going to explode!!!!
 
I'm not saying it's probelm accessing the EQ "settings" on the Xonar's CPL, I'm merely saying that in order to see if the Xonar control panel is passing through audio or not you have to watch the EQ bars to see if they are moving - if they are moving you know the Xonar environment is passing audio through but if the bars are not moving then Windows is sending audio directly to the speakers bypassing the Xonar control panel which in turn would mean the XOnar settings and hardware rate conversion is not being applied.
 
I dont use the Xonars' EQ setting anyway, they never sound quite right. It maybe because I use a separate amp, I don't know. I'm of the same opinion as Hughy, My head is going to explode as well!!


Edit: I have listened to each way of doing this and its definitely superior when the W7 setting is 48000Hz and the Xonar is 192000Hz, at least through my system. The EQ setting still sound rubbish so I just leave it on default and the quality is excellent. I'm outputting through Kef 3005se for multichannel and a Pair of Dali Lektor 3's for stereo, so pretty decent speakers.
 
Last edited:
That's exactly what I am saying, you don't need to use the EQ settings, the bars will move regardless during audio playback and I don't use the EQ settings either.

Read my post again, and again if needed :p What I am saying is you can use the EQ bars as an indicator as to whether the Xonar control panel is passing audio from Windows or if Windows is sending the sound directly to the card so in effect bypassing hardware rate conversion and outputting at whatever is set in windows via software.

EG:

eq.jpg


Notice the bars activity?

Hi-Fi mode is tone defeat and disables all EQ effects which is what you'd want for an original sound.
 
Last edited:
That's exactly what I am saying, you don't need to use the EQ settings, the bars will move regardless during audio playback and I don't use the EQ settings either.

Read my post again, and again if needed :p What I am saying is you can use the EQ bars as an indicator as to whether the Xonar control panel is passing audio from Windows or if Windows is sending the sound directly to the card so in effect bypassing hardware rate conversion and outputting at whatever is set in windows via software.

EG:

eq.jpg


Notice the bars activity?

Hi-Fi mode is tone defeat and disables all EQ effects which is what you'd want for an original sound.

No need to be sarcastic mate:p, I was agreeing with you after all, you dont need to enable Hi-Fi if you are not using the EQ anyway, its makes no difference.

Do you know why the EQ sounds so poor, also why the updated driver 1777 hasnt tackled the issue of the W7 smaple rate setting.
 
Wasn't trying to be! but yeah I guess hifi mode isn't needed if you're not using them anyway but I guess for me anyway it's just peace of mind.

The 1777 drivers have been out a bit now, can't remember when I installed them but it was posted in a thread here some time ago - I remember reading someone posted that they could not tackle the windows sample rate issue because of the way windows 7/vista work with the sound and how its output - could be rusty memory though!
 
Yer think that was me mrk.

I'm sorry I didnt read your first post properly I assumed you were referring to the lack of DSP modes when enabling higher sample rates but I understand now you were talking about the Xonar's hardware being totally bypassed.

I am having difficulty with this - when I had a Xonar (remember we are going back about 6 months now), on my system, setting a higher sample rate in windows + Xonar's driver panel indeed meant that the EQ bars werent moving BUT all of my audio sounded miles better. Are we 100% that the Xonar was being bypassed? Maybe a driver issue? I am not convinced but again, its been a while and I am always willing to be proven wrong! :D
 
Yer think that was me mrk.

I'm sorry I didnt read your first post properly I assumed you were referring to the lack of DSP modes when enabling higher sample rates but I understand now you were talking about the Xonar's hardware being totally bypassed.

I am having difficulty with this - when I had a Xonar (remember we are going back about 6 months now), on my system, setting a higher sample rate in windows + Xonar's driver panel indeed meant that the EQ bars werent moving BUT all of my audio sounded miles better. Are we 100% that the Xonar was being bypassed? Maybe a driver issue? I am not convinced but again, its been a while and I am always willing to be proven wrong! :D

No it is not bypassing the xonar.

My amp only changes what it is displaying when I alter the setting on the Xonar control panel, it doesnt matter what I set the Windows setting to. It must just be a glitch in the driver.

For example windows setting is 96000Hz, Xonar setting 96000Hz, amp recieves 96000Hz, but eq bars not working. Set xonar to 192000Hz, windows still 96000Hz, EQ bars not working, amp recieves 19200Hz. Drop the windows to 48000Hz, EQ bars start working, Amp still recives 192000Hz.
 
Back
Top Bottom