I recently noticed that using the internal soundcard of my Dell computer I have a lot better audio results using ASIO4ALL drivers in any sound editor/player (Audition as editor, AIMP as player) than with the built in MME/directplay Windows drivers (by Realtek the audio chip maker).

Playing thru the windows brand's drivers really reduce the overall quality, loudness and bass response of my audio file (tested both AIFF and WAV, 48khz/24 bits).

It's very audible that ASIO4ALL offer the full audio spectrum, even in Ableton Live.

Strange that the same audio chip, using just different drivers, produce so different results ....

Any explanation on this ??

 

The only thing I know it's better for latency, if you work with the same sample rate to me normally the sound should not be better or worse, but maybe the explanation is on the science side. I'm curious to understand to

 

So same converters, same signal path, same everything except for the drivers?

I'd call that bias. But if you can prove it somehow, that would be pretty interesting.

 

Could be the sample rate conversion. Check which sample rate the hardware is really running at.

 

Maybe there are some "enhancements" active which are not being used while using the lower latency ASIO4All driver.

https://learn.microsoft.com/en-us/w...rs/audio/audio-processing-object-architecture

That's the point where system wide processing like EQ-APO or Sonarworks takes place.

 

Two options:
1) It doesn't happen and your brain tricks you;
2) Something's not good with your Windows/Realtek drivers.
Would be nice if you could somehow record the same file both on ASIO drivers and on the other ones. Either through loopback or, if you can't, Stereo Mix.
To do this, you'll need to enable the Stereo Mix in your sound options, launch the built-in sound recording tool from Windows, select Stereo Mix as the source, play the file in question and record it. Then upload the result and the source file itself. You may cut them so only, say, the same 30 secs fragment is playing.
Generally tho, ASIO drivers are indeed more stable for realtime audio processing, so you should always select at least ASIO4ALL, and your audio interface's drivers if you have one.

 

This could be the issue that is going around with Real Tek chipsets on Windows 10.

Windows 10 sometimes puts the Real Tek cards as a Mono, even though you have the settings in windows 10 set to stereo. Using the Asio4all drivers bypasses this and this happens quite often when you have the Combo mic/line out headphone jack on your laptop.

You can try this :
In your Daw set it to be "Direct Out"

Import a known stereo track ( MUST BE STEREO)

Hook up a pair of working headphones to the Mic/line out jack of your laptop

In your Daw on the master Bus play the track and Pan the track Left to Right

If you hear the Audio moving from the full left to full right and full right to full left ( as you pan back and forth)
then this means your Direct Sound/ ASIO4All is identical

If you hear the sound being panned but you still hear it on the other channel then you have the Real Tek Windows 10 driver problem ( Stuck in Pseudo Mono)

This happened on a laptop I was working on for someone who thought the Mic/Headphone jack was not working as he only got Stereo sound if he used Asio4all, the built in Speakers and Mic/Headphone jack all output Mono when using the Real Tek Windows 10 drivers.

Make sure you get the Real Tek Drivers and not use the Windows 10 built in ones, this was the problem the Windows 10 High Definition Drivers.

With Asio4all this bypasses the drivers being stuck in Mono mode, I believe this is why you are having a massive difference in Sound as Stereo is not only louder but the Sound Spread is much more apparent.

 

The difference is pretty easy to explain when you know how the audio system in Windows works. ASIO provides a direct, bit perfect audio stream directly to your hardware. Everything else (except WASAPI afaik, same as ASIO) goes out through Windows mixer and is subject to poor resampling quality. The difference should be subtle, though. Like between using a good pro-grade DA converter and a line out on your laptop. Hmmm, the difference is not so subtle, actually.

That's also why ASIO and WASAPI can achieve such low latency, because there's no OS processing, mainly resampling, in between.

 

Last edited: Feb 6, 2023

Long time Mr Wave!! "hope your well bro"

Ive personally found that using either Fl Asio or Asio4all reduces the cpu load by a fair margin compared to cubases generic driver etc etc when using live 9/10/11/FL20/bitwig blah blah

 

If we only talk about quality, or "sounds different", that's very likely the answer.
ASIO drivers skip the Windows settings including stock or external enhancers and go directly to low-level sound card output/input.

In the case of ASIO4ALL might also be important to match your sample rate/bit depth to your windows settings. Being a generic driver can't do what native drivers can. I'm not 100% positive but almost.

 

Thanks for this explanation bro'

 

There is no necessarily a mental bias or a matter of settings, etc. Quality of sound reproduction is dependent, among other things, on how many stages of processing an output driver entails in a given operating system. The more processing given, the heavier the CPU load, the more exceptions, the more digital noise, and more sound blurring. Software exceptions inevitably bring delays in time domain causing jitter and parasitic frequencies. For the same reason sometimes you even get drop-outs.
(In the same manner, CPU load that is imposed by memory resident system drivers and services is widely recognized as a factor that greatly impairs the sound quality of a system - just try 'Fidelizer' or a similar program).
In my experience smaller buffers also bring more sound quality, no matter what many people might think.
Also exclusive mode renders better sound quality than shared mode with the same output driver. The same is with push/pull modes. Pull mode brings even more sound quality in exclusive mode.
Not all ASIO drivers sound the same. For instance Steinberg Generic Lower Latency ASIO Driver sounds better than ASIO4ALL, probably because it can work in exclusive mode, and/or because of a better design.
So the idea of 'bit perfect reproduction' is not realistic in non audio-dedicated operating systems like Windows. Not counting the drivers, the minimal number of services in which Windows 10 can work is 7 (by some claims on internet), but PC isn't a CD player still and it never can be.

 

Ok I recorded the same WAVE file, first part with ASIO4ALL driver, second part with standard Realtek driver.
Same output, same quality, no enhancing option.

 

Part 1 has more bass, second part more middle range less bass ...

But it's the exact same audio chip.

The only difference is the driver used to play it back ...

I assume my brain didnt ticked me right ? There IS an audible difference ...

 

aaand that's usually because Realtek audio driver by default applies various sound enhancements, including EQ, limiter, and even some slight compression depending on model,
it's also a reason why LOUD AS FUCK playback doesn't distort, but it can easily distort through ASIO driver which bypasses those Windows/Realtek sound "enhacements"

 

In Windows and Realtek driver all enhancement are OFF, and third party enhancers (some WAVES stuff) uninstalled ...
Are there any sound enhancements that are enabled by default and cant be disabled ?
But as far as ASIO4ALL also use the same Realtek chip, it should also be enabled by default if it's a "feature" of the chip itself ?

 

There's definitely some dynamics processing going on. Are you sure there are no options enabled that could be causing this? Like for "loud listening environments", "same playback level" or something like this.

 

Nope!! it dint trick ya , i find the very same thing myself
Its Akin to when you resample in live 10/11, you always loose some of the bottom end n overall punch and have to duplicate the channels and pan them against eachother and add eq to push the bottom end up again to get close to the original punch prior to resampling
to be clear im reffering to the resample function in live not freezing or consolidating

 

Last edited: Feb 7, 2023

yes, in Windows audio device settings, I think it's even tickbox called "sound enhancements",
also make sure samplerate and bit depth is correct/matching in Windows audio device settings too

 

As I stated before all windows related enhancement are turned off and the file is played in audition with two different drivers using the exact same setting 48khz 24 bits.