I’m recording many tracks for a choir song (about 32 tracks in total) and when I record each track individually and play them back, there is no buzz, it sounds perfectly fine. But when I play it back, there is significant buzz, I’m not sure why?
Running Audacity 2.0.0 and using an Art M One USB mic.
If you have all 32 track one over the other you can use MUTE and SOLO to tell Audacity which tack you want to listen to. Does the problem get worse as you add tracks? You only described the two extremes.
Overdubbing has a relationship problem. One voice is 1/32 of the show, but any buzz on the microphone is present on every voice and will be 32times louder.
You have an even worse problem than the ACX AudioBook readers. They just have to get one pass to be quiet.
USB microphones in particular can have odd noise problems. There is a “frying mosquitoes” or “Yeti Curse.” Does his sound anything like your problem? Turn the volume up after I stop talking.
Thank you very much for your replies. I will definitely try those. I also applied compression and noise removal to all the tracks if that has any bearing on the buzzing. Also, I did go through all the tracks individually and there was no audible buzz, even at +10dB amplification.
Before you get too far along (and remember to always keep archive copies of the raw, unprocessed work), USB microphones have a nasty habit of producing sound you can’t hear.
Normal analog microphone systems produce sound down in pitch to about 20Hz (truck driving by, earthquakes and thunderstorms). Some USB microphones are just getting warmed up and produce “noise” down to 10Hz, 5Hz and lower. This is nasty stuff, can be “louder” than the show and can, in some cases, prevent you from passing an AudioBook noise test “for some reason.” Remember, you can’t hear any of this stuff.
It can also (and here’s the Agatha Christie moment) get so loud it overloads the sound channel and suddenly becomes audible.
“How come my songs are clean but the composite buzzes?”
I attach the spectrum analysis from a random clip submission. The only valuable sound (purple color) is between 20Hz on the left and 20,000Hz on the right. All the purple to the left of 20Hz is garbage, not audible and note that it’s higher (louder) than the show.
Drag-select the first ten seconds of one music track and Analyze > Plot Spectrum. Use the settings in that attachment (click on it to reveal the whole picture).
Compression can raise the noise-floor, which in your case includes (mains?) buzz noise.
In this scenario: buzz when playing 32 tracks at once, but no conspicuous buzz on a single track, I’d use noise-gate rather than noise-removal, a/k/a/ noise-reduction. The noise-gate will make the “silent” sections of tracks truly-silent, buzz-free, flatline. The noise-gate, if properly applied, won’t have any effect on the singing, whereas applying noise-reduction will inevitably effect the singing to some degree.
Do you hear the buzz where it’s silent/quiet (maybe at the beginning or end) or during loud parts?
If it’s happening during loud parts, it’s distortion (not noise). Mixing is done by summation and you can get severe clipping (distortion) by mixing 32 tracks unless you reduce the levels. (You can get clipping if you mix two tracks with 0dB peaks.)