this is a very plain question, and i swear i havent found any real satisfactory answer in the faq…
→ what’s the best way to set recording volume levels when e.g. recording streaming radio??? ok, i think i’ve seen somewhere that peaks should just reach 0.5, but recording level can be set both at the device’s output (realtek mixer) and audacity’s input, so is there a trick to get the best results, i.e. highest volume with smallest distortion? does the input/output ratio matter, or do i just want to be at 0.5?
i remember old (analog of course) recorders where recording volume could be set automatically for best results… i suppose there is no such thing in audacity?
thanks in advance - jm

<<<i remember old (analog of course) recorders where recording volume could be set automatically… >>>

And almost universally produced terrible results. Pumping, breathing, and gating effects. Nobody doing any serious work used AGC. The Nagra tape machines had that Modulometer and peak limiting, but it was far from Set and Forget. You still had to get it close.

It doesn’t matter because Audacity doesn’t do anything in real time except capture and play back. Everything else is performed in post production – later that day when you’re editing the show.

Remembering that everybody capturing internet audio using the Mix-Out business is really taking the cheap way out. Real capture software (WireTap, Total Recorder) doesn’t have volume controls because it captures the stream and saves it as a file with all the original characteristics. At worst, it converts to other sound formats and only very rarely has volume controls as part of the process.

Internet Audio shows have an interesting problem. They’re frequently taken from the studio master sound feeds, but don’t go through the radio transmitter compressors and limiters. So the nice gentle, always the same volume show you listen to in the car suddenly becomes an exercise regimen of reaching over to turn the volume up and down every time the actor leans into the microphone.

The only solution I ever found is to capture the whole show making sure the waves and sound meter never hit maximum. Then apply DC Removal or Normalize > Remove DC, followed by Chris’s Compressor. That is as close to the radio broadcast you can get by capturing a streaming version of a broadcast radio show.

That super-loud, peaks at 1 (or 0 depending), dense sounding show you’re used to hearing is produced in post production, not during capture.

…unless you literally capture the radio broadcast. Those are stablized at the radio station and you don’t need post production for those at all.


But even with that you can oversaturate the signal and push the recording into clipping unless you preset the recording level carefully.

I regularly capture FM off-air from the BBC, but setting the recording level correctly in advance of the show I want to capture gives good results.


There is an experimental new feature that is currently being developed to do this.

It works slightly differently to the “Automatic Gain Control” found on some hardware sound recorders and avoids the “pumping, breathing, and gating effects” that Koz spoke of. I don’t know whether it will make it into the Audacity 2.0 release or will come out later.

Basically it samples the peak recording level for a period of time set in Preferences and checks this level periodically during that time (the frequency of checking can also be set in Preferences). If the recording level hits 0dB, the recording level is turned down a bit. This could, if enabled, mess up the start of your recordings, but if recording from an external source, you would either start the recording a bit early and snip off the beginning, or start the recording, wait for the level to be set, then start the recording from the beginning again.

… posting this for those who like me cant afford that iphone and onthego 3g, and owe much to audacity (thanks really) for being able to mp3-can their music…

does the input/output ratio matter?

just tried recording with either max input volume + output vol set for peaks to barely pass 0.5, or max out / 0.5 in => no difference to my imperfect ears (got great bass with the former but the tunes were different so i cant really tell)…
steve: looking forward to that agc notwithstanding koz’es pessimism… :wink:

Koz isn’t the only one to not like AGC. My personal view of AGC is that it is an abomination, it is the curse of the dark side of technology, a malign invention to taunt sound technicians and torture music lovers. (I could go on but I think you get the idea). Fortunately the Automatic recording level that is being developed in Audacity is a little different. Rather than it riding the dynamics constantly up and down and irreversibly mangling the recording, it “listens” to the first few seconds (or longer) of the sound being recorded and adjusts the level, much in the same way that one would do it manually. Once the correct level has been determined (with an allowance for a bit of head room), the recording level stays at that level for the rest of the recording, thus avoiding the evil that is usually associated with AGC (so yes, I’m looking forward to trying it too) :smiley: