What audio level to record at?

I just installed this 30 minutes ago. I am recording a radio station streaming their music. I will leave it recording all night. I will make an MP3 of it to play at work. What should I set the volume to? Right now it’s peaking at about 12. I can’t listen to it to do test because everyone is asleep. Think 12 is ok for playback on the radio in my truck?

Radio stations are usually quite consistent with the level of their broadcasts.
For commercial stations with adverts, the adverts tend to be the loudest part.

I would recommend a maximum peak level of around -6 dB, which will give a waveform about half the track height.

When the recording is complete, the track can be normalized to -1 dB before export (http://manual.audacityteam.org/man/normalize.html), which should be a good level for playback on the radio in my truck.

Check the computer settings and ensure that the computer will not go to sleep or hibernate.
Ensure that you have plenty of free disk space.

Note that the maximum size for a WAV file is 4GB. Attempting to save a WAV file bigger than that will result in a corrupt file. That won’t be a problem unless you are recording at a high sample rate or making a very long recording. 8 hours of 16-bit 44100 Hz stereo should be fine.

What should I set the volume to?

It depends on where they got the stream from.
If they got the stream after the broadcast compressors, you can move it up to -6dB or so because the volume won’t drift. If they got it from the automation or other production supply -12dB may be about right.

The Car Talk radio show in the US is available as both off-air and streaming. The two shows are miles apart in production volume control. The air show—after the compressors—is nice to listen in the car, but the stream is almost unlistenable in the car or headphones. Ray mumbles in his beer and Tom has a thermonuclear laugh.

If you find the stream wanders in volume, for example switching in and out of network news or drift between shows, you might try Chris’s Compressor.


I move the first number Compress Ratio from default 0.5 to a stiffer 0.77 and the processed stream is almost an exact match for the broadcast.

Chris needs some sound at the beginning of the show to compress right. I leave a ratty edit sticking out at the beginning and then go back later after compression and cut it off.