I am producing a song. I have a piano track recorded and set at -10 dB. Now i want to mix it with my recorded vocal tracks, which are set at -1 dB. This ratio i find comfortable, as when combined, the vocal dominates over piano, but the piano still keeps it’s presence powerful enough in the song.
Problem: after mixing and rendering piano with vocals, the piano track drops decibels significantly in the part, that is combined with vocals and that totally ruins the song. I did not have this problem in the older Audacity versions - i can remember CLEARLY that when two tracks would be mixed together, no way would one of them get weaker in loudness…
that is combined with vocals and that totally ruins the song. I did not have this problem in the older Audacity versions - i can remember CLEARLY that when two tracks would be mixed together, no way would one of them get weaker in loudness…
Yeah… I’m baffled by this and I suspect it’s just a perception.
…i.e. Mixing is simple summation and if you don’t touch any of the levels you can invert one of the tracks and mix the inverted track with the previous mix to completely subtract it out.
which are set at -1 dB. This ratio i find comfortable,
The ratio is entirely up to you and your ears. The [u]Mixer Board[/u] might be handy.
But… The summing can push your levels into clipping (distortion from “trying” to go over 0dB) depending on how the peaks line-up.
From my handy-dandy spreadsheet -
-1dB = 0.89
-10dB = 0.32
Sum those and you get 1.21 (121%) = +1.6dB.
Audacity itself won’t clip internally so rather than fooling-around with calculations an easy solution is to export as floating-point WAV (which also won’t clip), re-open the mix and run the Amplify or Normalize effect to bring the level down (or up) before exporting to your desired format. (Although you could keep the floating-point file, it’s “bad practice” because playing it can clip your digital-analog-converter.)
Having read and re-read Artas1984’s post, the piano is either digital or an upright.
If its digital, surely the output would be stereo, when recorded into Audacity. If so, Artas1984’s piano track would be in stereo and could be panned for example 60% left.
Artas1984’s goes on to say the vocal is -1.0 db, that tells me it’s been normalized to that level. I would assume the vocal track is in mono. If so, one could duplicate it and pan it far right, and the other, lower, roughly at the same level as piano in the mid stereo picture.
Add a little bit of garnish such as echo and reverb panned far left and far right.
If Artas1984 has recorded the song on an upright piano (or grand) and singing whilst playing, that would explain the 10db lower volume on the piano because its just picking up the spillage, and would come up to -10db when the vocal is normalized.
It would be welcomed, for analytical purposes, if Artas1984 could post a clip highlighting the problem.
I did not have this problem in the older Audacity versions
What was the earlier version and which are you using now? There is no “latest,” we need a real number like 3.0.4.
Audacity doesn’t apply corrections, filters, or effects when recording. I can’t come up with a good scenario, but it sounds like the computer could be getting into the act. Do you have any of the Windows playback effects or “enhancements” running? Check the Windows Control Panels.
A much more sinister and evil possibility is another application. Do you like Zoom, Skype, Meetings or other chat program? They take over your sound system and you can’t stop them. I can see one of them changing your overall sound volume without telling you—and they can change over time.
Here’s one.
If you play your voice clean by muting the piano track, is it clean? Is there any piano in there at all? A very common recording problem is double recording the background track by accident. Overdubbing and Recording Internet Sound settings hate each other.
Mute only the voice track. Does the piano track volume go back to normal?
