[audacity4blind] Re: Normalizing Versus Amplification Versus Gain

  • From: Robert Hänggi <aarjay.robert@xxxxxxxxx>
  • To: audacity4blind@xxxxxxxxxxxxx
  • Date: Sat, 6 Dec 2014 19:53:29 +0100

2014-12-06 18:44 GMT+01:00, Rich De Steno <ironrock@xxxxxxxxxxx>:
> Robert, thanks for this detailed response.  I pretty much understand it,
> except for this last part:
> My question:
> After you mix and render all tracks to a single track, you say to also
> select the mixed track and then "Normalize or amplify all selected
> tracks, including the mixed track,  to e.g. -3 dB, and then to delete
> the mixed track.  Why include the mixed track in the normalization or
> amplification?  It seems redundant to do so , since it duplicates the
> material in the separate tracks and then you are deleting it anyway.

It is important to include the mixed track and to use amplify (not normalize).

The mixed track is the sum of all other tracks.
The amplify effect proposes a value that reduces the gain to reach 0 dB.
It takes the loudest track to calculate this value.
That's of course the mixed track.
It then subtracts this dB value from all tracks equally which
automatically results in a perfect mix.
Let's say that our tracks have the following values:
Track 1 -6 dB (0.5)
Track 2 -6 dB (0.5)
Track 3 0 dB (1.0)
mixed 6 dB (0.5 + 0.5 + 1.0 = 2.0)
Amplify will now show "-6 dB".
All tracks are therefore amplified by -6 dB, that is their values are halved.
And this gives 0.25+0.25+0.5 = 1.0 (without the mixed track, as you can see).
In order to reach an overall level of -3 dB, you can either directly
enter -9 dB instead of -6 dB or amplify  a second time with -3 dB (the
proposal will be 0 dB).

If you would omit the mixed track, the loudest track would be taken (0
dB in our example) and thus 0 dB amplification will be proposed.

If you want to recreate this example, make sure that you have three
identical tracks. The proposed amplification value will be smaller
when the tracks are different (try e.g. white noise).
The reason is that all individual samples are added together, for
white noise, this could result in anything from -2 to 2 but in average
+ 3 dB instead of +6 dB for two full scale wave forms (i.e. 0 dB or -1
to 1).

However, the procedure works for all cases, just the amplification
factor will change.

I know, that's even more confusing than the last reply...


The audacity4blind web site is at

Subscribe and unsubscribe information, message archives,
Audacity keyboard commands, and more...

To unsubscribe from audacity4blind, send an email to
with subject line

Other related posts: