Recording Hi-Res via Wasapi Loopback

Morning All, Newbie to the Forum so Hi to everyone. Been using the program for a few years and find it to be one of the best softwares available for audio processing.

I’d like to pick your brains if I can! I have a project where I will be recording 24/96 and 24/192 via Wasapi loopback. The project needs to be in 16/44.

So if I use 32 bit float and 44.1 as the recording options then choosse 16bit signed when exporting will that achieve my goal properly. Or is there any other settings that i should enable for resampling or adding dither.

Many Thanks & Best Regards

Yes, that should be OK if it records.

If it doesn’t record, you may need to match the sample rates (all the same sample rate) when recording. To export as 44.1 kHz, Audacity’s “Project Rate” (lower left corner of the main Audacity window) must be set to 44100 when you export.

Thanks Steve much appreciated. What I couldn’t get my head around was it downsampling say 192 to 44 in real time or just simply recording a higher sample rate souce at a 44.1.

Best Regards

I’m not sure exactly where the conversion occurs when recording a 192 kHz source as 44.1 kHz - it may be platform dependent.

If you record at 192 kHz and then either “Resample” to 44100 (“Tracks menu > Resample”) or simply Export from 192 kHz as 44100, then the conversion is handled by “libsox” which is a very high quality resampling library in Audacity.

I will be recording 24/96 and 24/192 via Wasapi loopback.

Wasn’t there a post that the Windows internals didn’t support fancy/advanced digital rates? You may get away with that for one-pass as in straight recording, but not multi-pass where Windows has to do re-direction.


You may be thinking about sample formats.

When using MME or Direct Sound, the data received by Audacity will always be 16-bit, regardless of what settings you use.
As far as I’m aware, that limitation does not apply to WASAPI (though I’ve not tested, and I don’t have the necessary hardware to be able to test).

data received by Audacity will always be 16-bit, regardless of what settings you use.

That’s it.


You usually don’t know what Windows is doing…

I just did an experiment and I got some screwy results - I generated a 1kHz sine wave file file and a 50kHz file at a sample rate of 192kHz. Then I ran Amplify at -50dB twice (to get below the -96dB 16-bit limit) and exported as 24-bits.

I played the files in Winamp while recording with WASAPI loopback.

The good news is when I re-amplified it by +100dB it wasn’t silent so it’s not limited to 16-bits.

The ban news is the “waves” were all fouled-up/distorted. Maybe aliasing, so there’s probably some bad/weird re-sampling going-on.

When I tried again after exporting as 16-bits (to get a silent 16-bit file) I was able to play & record digital silence and after amplifying by -100dB it was still silent. At least THAT worked as expected.

Thanks for all the responses much appreciated. I have used dbpoweramamp in the past to batch resample, set to 32 bit float, set to 16 bit and then add triangular dither with good results. I was hoping with Audacity I could cut out the middle man. I’ll still give it a bash and listen to what I can achieve through Audacity this way. One thing I have noted though is when in Wasapi (Shared) Windows still passes the audio through the mixer at the settings set in the control panel. It seems complete pass through of the singal only happens when in Wasapi (Exclusive). But then you can’t record as the player assumes full control of the audio path.

Best Regards

Yes you should be able to cut out the middle man. Audacity’s resampling is excellent quality. By default Audacity will use “shaped” dither, which is generally considered a better option (lower noise) than triangular, though Audacity also offers triangular, rectangular, or none as options (details here: Audacity’s default settings are usually the best options.

Thanks again Steve for the information, fantastic forum…