Does ASIO or WASAPI really optimize sound quality?

[Marcos Mazur]

Does ASIO or WASAPI really optimize sound quality? I'm not an audio expert but I've noticed that these ''bit-perfect'' drivers improve sound in one way and worse in another, definition and accuracy improve (low latency) but sound vibe (so to speak) is a bit tiresome and ''forced''', Direct Sound being more limited sounds slightly congested but at the same time provides a less artificial sound, oddly enough. Searching forums I didn't see anyone commenting on it, would it be my ears, personal taste or is there something real about it? many DAC's like mine come automatically with ASIO support, I've used some others before and I realize the same thing.

 

[Propheticus]

Something with more latency is not less accurate, it's just 'late'. No issue for most domestic music listening. Low latency is needed for people producing/recording music (eg using midi devices to play and capture while also listening to a loopback on a monitor/headset) or of course live music.

Upsides are the availability of exclusive modes (Asio or wasapi are not synonymous to exclusive) and automatic format matching to prevent resampling.
Still that does not guarantee bit perfectness, which depends on playback software, DSPs enabled, EQ applied, system wide APOs.

Will not make a generic statement because it always depends, but audibility wise Direct Sound can be just fine and as good as other options.

 

[aandres_gm]

It's not that they improve/optimize anything, but more a case of not letting Windows touch the signal on its way out, which may degrade the signal.

In my experience, unless you know what you're looking for, there won't be noticeable differences.

 

[Offler]

Does ASIO or WASAPI really optimize sound quality? I'm not an audio expert but I've noticed that these ''bit-perfect'' drivers improve sound in one way and worse in another, definition and accuracy improve (low latency) but sound vibe (so to speak) is a bit tiresome and ''forced''', Direct Sound being more limited sounds slightly congested but at the same time provides a less artificial sound, oddly enough. Searching forums I didn't see anyone commenting on it, would it be my ears, personal taste or is there something real about it? many DAC's like mine come automatically with ASIO support, I've used some others before and I realize the same thing.

I did some tests using Windows PC, Realtek ALC1220 soundcard and NAD D3020 v2 with integrated DAC.

a) Wasapi Exclusive Mode
Sound Settings > Properties > Advanced > "Exclusive Mode" Enable both settings.
Then you have to enable this mode in you music/Media player. Its available in Foobar via plugin.

This mode means that data are sent from the PC without any modification. In my case i sent music data from PC to amplifier over Optical SPDIF to separate amplifier from EM noise which is in PC.

This is the best possible quality to play music/movies. on Windows PC.

Data are simply sent to DAC inside my amplifier, separated from the noise inside my PC, while OS is not allowed to alter the sound in any way.

b) Tweaking DirectSound
I also play games, so its important for me to fix their sound as well.

I used measuring software called AudioTester V3.0 to confirm that Windows distorts sound most of the time by using 'CAudioLimiter' filter. There is a workaround for this. Install Equalizer APO and set gain to -4.00dB. That will prevent triggering CAudioLimiter.

In my case it reduced THD+N (Total harmonic distortion) from -60dB to -83dB (which is figure given by manufacturer of my amplifier).

c) I cannot help with ASIO. My soundcard supports it very poorly.

 

[Vincent Kars]

would it be my ears, personal taste or is there something real about it?

Most likely you have establish this in a sighted test. This is what is called "listening with the eyes" and that is not the right organ.
Try a unsighted test.

Technically, what they call
Direct Sound = WASAPI in shared mode
WASAPI= WASAPI (of course) in exclusive mode.

Obvious, the protocol (WASAPI) remains the same.
As Offler already pointed out, the difference is using the Win audio stack or not.
If you do,
and resample,
and the audio contains signal close to or up to 0 dBFS,
there is a measurable degradation.

In all other circumstances, there isn't.
This assuming you use a 24 bit (or 32) data path to the DAC.

 

[Marcos Mazur]

Something with more latency is not less accurate, it's just 'late'. No issue for most domestic music listening. Low latency is needed for people producing/recording music (eg using midi devices to play and capture while also listening to a loopback on a monitor/headset) or of course live music.

Upsides are the availability of exclusive modes (Asio or wasapi are not synonymous to exclusive) and automatic format matching to prevent resampling.
Still that does not guarantee bit perfectness, which depends on playback software, DSPs enabled, EQ applied, system wide APOs.

Will not make a generic statement because it always depends, but audibility wise Direct Sound can be just fine and as good as other options.

Something with more latency is not less accurate, it's just 'late'. No issue for most domestic music listening. Low latency is needed for people producing/recording music (eg using midi devices to play and capture while also listening to a loopback on a monitor/headset) or of course live music.

Upsides are the availability of exclusive modes (Asio or wasapi are not synonymous to exclusive) and automatic format matching to prevent resampling.
Still that does not guarantee bit perfectness, which depends on playback software, DSPs enabled, EQ applied, system wide APOs.

Will not make a generic statement because it always depends, but audibility wise Direct Sound can be just fine and as good as other options.

''Something with more latency is not less accurate, it's just 'late''
Truth.

'Upsides are the availability of exclusive modes (Asio or wasapi are not synonymous to exclusive) and automatic format matching to prevent resampling.''
I don't know if I understand exactly, it means that in direct sound some resampling can occur?

I use Direct Sound and it seems to me ''normal'' sound, it also questions whether the sound could be optimized in this way.

 

[Berwhale]

I don't know if I understand exactly, it means that in direct sound some resampling can occur?

Yes, all sources are mixed and output at a fixed bit depth and sample rate. For example, my D50s is connected via TosLink and is set to 24bit, 96Khz

If you want to learn more about resampling and if it's an audible concern, there's a lot of information in this thread: Ending the Windows Audio Quality Debate | Audio Science Review (ASR) Forum

 

[Lupin]

I don't know if I understand exactly, it means that in direct sound some resampling can occur?

Yes.
In direct sound mode or WASAPI shared mode everything will be resampled by windows to whatever is set in the property window.
If that is actually an audible issue.. hundreds of pages are already written about that just on this forum alone, I'm not going down that rabbit hole

 

[Ron Texas]

I find the main value of WASAPI or ASIO for music playback is if a computer is being used for playback and say a browser is open the exclusive mode blocks all audio from the browser. News sites are the worst culprits. The other advantage is if you are using a convolver with FIR filters this will bypass Equalizer APO which may be there for other sources like video playback.

 

[Offler]

Just started to test ASIO plugins for Foobar2000 and MPC HC on Realtek 1220... Surprisingly both works without any issue.

Playback acts as with WASAPI Exclusive mode - other sounds are supressed.

Subjectively... The sound is very different from WASAPI both Exclusive and Shared.

Shared Mode could benefit from EQ APO, Exclusive bypassed windows audio stack... ASIO sounds like having benefits of both. Mainly treble sound and percussions like chinas are much "real" and dont sound like hiss.

Some of the background noise became clear and audible.

 

[Lupin]

Just started to test ASIO plugins for Foobar2000 and MPC HC on Realtek 1220... Surprisingly both works without any issue.

Playback acts as with WASAPI Exclusive mode - other sounds are supressed.

Subjectively... The sound is very different from WASAPI both Exclusive and Shared.

Shared Mode could benefit from EQ APO, Exclusive bypassed windows audio stack... ASIO sounds like having benefits of both. Mainly treble sound and percussions like chinas are much "real" and dont sound like hiss.

Some of the background noise became clear and audible.

I think you really need to read the note on the Foobar2000 ASIO plugin page...

Please note that this component is meant for systems where ASIO is the only available output method. It is highly recommended to use the default output modes instead of ASIO. Contrary to popular "audiophile" claims, there are NO benefits from using ASIO as far as music playback quality is concerned, while bugs in ASIO drivers may severely degrade the performance.

 

[Propheticus]

Strange... do not expect a difference, let alone a big one. Possibly one of them is 'broken' as implemented on the Realtek codec (my bet would be Asio), or the volume differs...or you're imagining things.

 

[Offler]

I think you really need to read the note on the Foobar2000 ASIO plugin page...

I cannot measure anything on audio tracks of course.

What was measurable was higher count of ISR and DPC calls compared to WASAPI, and audio driver could take up to 60 microseconds which is 5x more than with WASAPI or default. It simply indicates that audio is taking much more time, but also higher priority to process.

I was definitely not expecting anything audible, only in cases where I could confirm presence of resampling or anything related.

There are few instances where sound which is definitely on the track became more audible - like there is a track where drummer hits the sticks by mistake and bassist hits the strings 3x almost inaudibly during intro...

Strange... do not expect a difference, let alone a big one. Possibly one of them is 'broken' as implemented on the Realtek codec (my bet would be Asio), or the volume differs...or you're imagining things.

I am surprised, it makes sound, not even that it sounds good. I use Realtek only for its digital output, so the processing should be minimal anyway.


Also regarding MPC HC...

There is ASIO audio renderer - MBSE Multichannel Asio Renderer - it claims it renders audio in 48KHz 32bit FLOAT format, which does not make much sense... Output is Optical SPDIF and DAC is capable just 24bit sound.

There was a tool which allowed to display each object in an active sound chain from file, through application, all filters and driver... I dont remember the name...

A presence of an APO or transform filter in one or another setup would explain it.


edit:
Did some listening tests with Foobar ASIO plugin VS Wasapi and Default.

Wasapi Exclusive and ASIO sound the same. WASAPI plugin requires to specify output bitrate. If its not matching the source, there might be transform filter involved...
ASIO sets output bitrate automatically, apparently on the same value as media file. but "pause" does not work.
'Default' sounds slightly less loud, and some detail is missing.


For MPC HC
- WASAPI Exclusive is possible through internal filter in MPC HC. It uses limiter and volume filters. At least it does not lie about 32bit FLOAT output and reports correctly its being reduced to PCM 24bit padded before sent to DAC.
- ASIO Renderer allows better transform filter than windows default, again allows output samplerate and bitrate to be selected automatically and in bit-matched manner.
- There is little to no difference when using 'Default', again slightly less loud...

Both ASIO audio renderers are exclusive by default, in case of Foobar introducing few small advantages.

 

[Offler]

Ok, lets take the scientific approach.

1. I found 1KHz sinewave 24bit 192KHz samplerate 0dB Wav file.
2. I played it through various audio renderers in Foobar2000 and Media Player Classic - Home Cinema.
3. In every case signal went through optical SPDIF and landed in external DAC of NAD d3020v2.

Lets start with MPC-HC:

Default audio renderer.

There is a significant "Windows" distortion around 1KHz.

Then i applied -0.14dB correction in Equalizer APO:

and the distortion is gone. This is a way how EQ APO may fix music and sound in games. (afterwards i turned EQ APO off again)


Then there is MPC-HC internal renderer:

Results were basically identical for both exclusive and shared output. Sound is slightly less loud, by about 5dBs.


And then there is MPC-HC with MBSE Multichannel ASIO Renderer:

I expected it to be perfectly identical to Internal renderer, but there are two spikes around the 1KHz and harmonics seem to be slightly higher, but nothing audible, but the main signal is slightly closer to 0dB.


Then lets go for Foobar2000, Default audio renderer:

Interesting. Foobar applied its own volume limitation by about 8-10dB, so the sound is not distorted.

Then lets go for WASAPI Exclusive, with 24bit rendering:

Now the signal is much closer to the 0dB and harmonics are louder as well.

Lets check ASIO:

Almost identical to WASAPI Exclusive.

 

[SeriousApple]

Does ASIO or WASAPI really optimize sound quality? I'm not an audio expert but I've noticed that these ''bit-perfect'' drivers improve sound in one way and worse in another, definition and accuracy improve (low latency) but sound vibe (so to speak) is a bit tiresome and ''forced''', Direct Sound being more limited sounds slightly congested but at the same time provides a less artificial sound, oddly enough. Searching forums I didn't see anyone commenting on it, would it be my ears, personal taste or is there something real about it? many DAC's like mine come automatically with ASIO support, I've used some others before and I realize the same thing.

Use WASAPI only asio is meant for use with fl studio and stuff for production. WASAPI is definatley the best and the latest. In my opinion it sounds best.

 

[Offler]

Use WASAPI only asio is meant for use with fl studio and stuff for production. WASAPI is definatley the best and the latest. In my opinion it sounds best.

If you are not a musician who uses Windows PC for recording, the only other extra benefit of ASIO compared to Wasapi Exclusive is for Zoom calls, to reduce latency.

 

[czt]

After shared, exclusive WASAPI is coming too natively (without component) in the next stable foobar2000 (as in current beta).

 

[Trell]

If you are not a musician who uses Windows PC for recording, the only other extra benefit of ASIO compared to Wasapi Exclusive is for Zoom calls, to reduce latency.

Depends on the features the ASIO driver provides. For RMI ADI-2 DAC with ASIO driver one can use the DigiCheck application. With an upcoming firmware one can even use DigiCheck on the DAC output after DSP.