(Post)Gain and DPL2 Downmixing

[Razor04]

Normally I decode my AC3 with BeSweet and allow it to apply post-gain. I notice that in the latest profile that comes with the GUI there is a static gain of 10db applied during the decode process. Someone mentioned that you should do this in another thread about DS2 in BeSweet. I was wondering if someone could explain the purpose of applying this gain during the transcode and then post-gain afterwards, instead of just using post-gain.

Thanks!

[Trahald]

i vote for NOT using a static gain as it would cause clipping if there is a peak that with +10db added would go too high.. ive had a few vids that auto picked only 9 db and sometimes less which would mean 10 would e too much

[DSPguru]

Originally posted by w00kiee
ive had a few vids that auto picked only 9 db and sometimes less which would mean 10 would e too much really ?!
- did you use it together with dialog normalizaion ("-n1" in the profile)
- what was the downmix type ? (dolby surround in the profile)
- can we see those logfiles ? (10q :))

[Razor04]

This is what I was referring to in particular, that frank posted in the thread on BeSweet's DPL II downmix matrix:

That means the DS2 downmix has an attenuation of -9.8 dB.
The probability of overflow is very small. You should start a 1-step decoding with a static gain of +9.8 dB. The resulting dialog level will be unchanged in this case.

So should I use this gain or not? I am really confused by all of this, although every movie I have done seems to have a final gain of 12-17db done in post gain, so it should be fine (right?).

[DSPguru]

Originally posted by Razor04
So should I use this gain or not?we need to study the behavior of the dpl2 downmixing and scale the matrix coefficients to have similar gain to the dpl downmix process.
you might have noticed that OggMachine sets 10db for dpl1 and 5db gain for dpl2. i was being cautious.
in the future we'll have a scaled matrix, and both modes would have 10db of gain.
don't forget that v1.4 wasn't released yet, and we still tweaking our betas.

every movie I have done seems to have a final gain of 12-17db done in post gain, so it should be fine (right?). right :).

[frank]

...and both modes would have 10dB of gain.Yeah, I agree!

DSPguru, I had downscaled -9.8 dB in the new matrix, but forgotten in the line comments. You can full apply the 3/2 output from Azid. Then we would get the right results.

I had again damn overflows from Azid at static gain greater 7 dB.
(DRC normal). I had to split 6 dB + ota. Awesome.

Q: How does ota work? 16-bit or floating point?

Greetings
frank

[DSPguru]

Originally posted by frank
Q: How does ota work? 16-bit or floating point?floating :D.

[frank]

Then I could make the whole gain in ota without quality loss?
For 16-bit output waves (aif) too?

[DSPguru]

yes.