Does it harm mp3 dynamics Normalizing one channel?

[datauser]

I downmixed a 5.1 track to a PrologicII 160kb/ps mp3 with Belight and after checking the balance of channels:

L =94.4 R=95.1


I see the right has a 0.7 difference using Sound Normalizer 3.9.3 so moved the normalizing level dial on the right side to -1.7% and the channels came out more evenly balanced after a right channel normalize and save.

L=94.2 R=94.6

My question to the audio gurus out there then is does this harm the dynamic range or cause any loss of information to the original track which originally was declipped -10.5db to 84.4db (from my overamplified resultant mp3 160kbps track processed in Besweet) using mp3Gain and now after my fine tuning: equalizing with Sound Normalizer it now shows as 84db?

Some articles I read say no harm whatsover but others says loss of information might occur when normalizing channels to lose rather than increasing.:thanks:

[tebasuna51]

If you change the balace between L-R you lose surround info.

With BeLight use the Azid settings -> Normalize to your desired value (limit 100%) and don't touch the output mp3.

[SquallMX]

Try doing the downmix using eac3to with the "-phaseShift" option. This fixed a similar issue that I had. :helpful:

[tebasuna51]

Try doing the downmix using eac3to with the "-phaseShift" option. This fixed a similar issue that I had. :helpful:

Do you think the light difference (L =94.4 R=95.1) is a symptom of the problem with the unshifted SL-SR channels?

@datauser
You can try the SquallMX suggestion and see if you obtain a more balanced output:

eac3to input.ac3 -down2 -phaseShift -normalize | Lame -b 160 - output.mp3

[datauser]

Hi
Forgot to ask! Would it destroy the surround phase too if you increase the lower channel L to balance L +R channels instead of decreasing one of them(R) ? :thanks:All

[tebasuna51]

"If you change the balace between L-R you lose surround info."

You can increase/decrease both channels at time but not change the balance (only increase/decrease one channel).