Bitrate for recording with Realtek and low volume

Hello. First post from a casual user of Audacity, using it for the past couple of years for vinyl archiving and restoration of (mostly my friends!) record collection. Needless to say I’m thoroughly amazed what your team has developed.

The reason for my post is first just clarification of what sampling rate I should set my on-board Realtek ALC889A soundcard. I understand that the Realtek is certainly not the best available, but it seems to work fine for my hobby use. The ALC889A defaults to 16/48, but there are many other bitrates available (see attachment). I also understand the minimal audible returns above 16/44 and that Audacity recommends capturing at 16/44, so should I set my card to 16/44 to match Audacity or set Audacity to the Realtek’s default 16/48 (or some higher rate for both) for creating my original project? Or does it even matter if they match? My end result generally gets exported as 16-bit PCM WAV for CD burning, but sometimes FLAC.

A second question regards the low levels I experience using what was “Stereo Mix”, but now is just the output loopback (just upgraded to 2.1.1 from 2.0.6 last night and seemed to have lost Stereo Mix as a working source, but have workaround). I have just started using Audacity to capture web content, so I’m very new to this. When recording I have an average level around -20 with all levels at 100%. I know I can use amplify/normalize to fix this, and to this point it doesn’t seem to have raised the noise floor too much. Maybe it doesn’t matter, but I wanted to make sure I wasn’t making a big mistake.

Also, of possible general interest, I found I was able to split the digital outs on the motherboard, sending the coax to my receiver and loop the motherboard’s optical output to an optical input on the same board, making the optical-in a recordable source. This board doesn’t have an optical in stock, but I bought the manufacturer’s (Gigabyte) add-on bracket for the headers present. Still results in low volume for web content, but responds normally to other sources. Thought that was interesting!

Audacity 2.1.1 Win Server 2008 R2

Thanks again for all your work and dedication and any responses.


It depends what host you choose in Device Toolbar (and probably does not matter too much if you are not making overdubs). See Missing features - Audacity Support.

With stereo mix recording you have to play the audio loud enough to record it. And stereo mix does not record the optical output or output of USB headsets or USB speakers.

If you get stereo mix enabled again, check if its input levels are maximum in Audacity and Windows Sound.

If you mean Windows WASAPI loopback, its achieved recording level usually maxes out about -6 dB. Usually (but not always), turning the audio source up or down does not affect that achieved recording level. See Tutorial - Recording Computer Playback on Windows.


Thanks for the quick reply.

I guess “Choose a project rate that the device supports, and use that same rate everywhere.” pretty much says it all. So I’ll probably just use 48/24 on the card and 48/32 float in the app since it will just be exported to a lesser rate and minimally processed.

Yes, I meant WSAPI, and yes, all the volumes are 100% (mixer, master, the individual I/O’s, the apps and even the control on the web page). I can live with it.

Thanks again.

My end result generally gets exported as 16-bit PCM WAV for CD burning, but sometimes FLAC.

If you’re making a CD, 16/44.1 is fine and it saves any conversion. I usually use 16/44.1 for CD and 16/48 for video.

There’s no harm in using higher bit depth or higher sample rate other than larger files. WAV and FLAC files will be proportionally larger. MP3 and AAC file size depends on the bitrate (kbps) or quality setting. Windows drivers will up-sample, so you never really know if your soundcard is recording in high-resolution or if it’s recording at 16/44.1 (or 16/48) and upsampling.

Some people claim analog vinyl has “infinite resolution”, but that’s false because the noise limits the true usable resolution. The bottom line is, CD quality is better than human hearing, and vinyl records are worse than human hearing. And, once you’re better than human hearing there’s no audible benefit higher resolution… It’s like looking at a regular-size TV across a football field… You can’t see the difference in standard resolution and high-resolution Blu-Ray.

Thank you, that’s interesting. I wonder if this info from the datasheet on the ALC889 would tell me what it’s really doing. This is from the “Hardware Features” section of the datasheet. Thanks again for your reply.

That tells you what formats are supported, but does not tell you how it is doing it. For example it doesn’t tell you whether 88.2 kHz is a clock doubled version of 44.1 kHz and is just interpolating data captured at 44.1 kHz. Recording at the “destination” sample rate will mean one less processing step in Audacity.

Setting your hardware to capture in 24-bit could be a significant advantage for recording low level signals IF the noise floor is below -90 dB. However, if the noise floor is significantly higher than -90 dB (very likely), then the increase from 16-bit to 24-bit has little benefit (better quality surface noise :confused: ), but still has the disadvantage of larger files and greater load on computer resources.

Interesting again. So would there actually be any way to tell the “native” resolution of the chipset? Maybe the default 16bit/48K set in Windows (with the latest Realtek driver)? Or just don’t bother chasing shadows and enjoy the music? :smiley:

As I understand it, the Windows Sound “Default Format” for sample rate and bit depth is sent to Audacity when the device is in “Shared Mode”, which happens if Windows DirectSound or Windows WASAPI host is chosen in Audacity and both Exclusive Mode boxes for the device are disabled in Windows Sound.

With those Exclusive Mode boxes checked and running under DirectSound or WASAPI, Default Format should in theory only determine the number of channels sent to Audacity, while Audacity requests the sample rate and bit depth directly from the device.

It is true that (perhaps this is what Doug refers to) that in the Vista and later audio stack when Shared Mode is enabled, Windows upconverts all audio to 32-bit float before Default Format is applied. That upconversion does not happen if Exclusive Mode is enabled.

However, Windows DirectSound is 16-bit only as far as I know, so there will still be downconversion to 16-bit if necessary even with Exclusive Mode enabled. And PortAudio (the audio API that Audacity uses) always captures in 32-bit float, so there is always upconversion there.


Thanks for the explanation. Back to reading! Boy, where’s Betty Crocker when ya need her?!