Audio DTS 5.1 to AAC Stereo.

[cmartinez]

Hello All,

i have volume related issue when trying to go through audio track .DTS 5.1 to AAC stereo.
Part of the process is done with xVid4psp (5.xxx), and as far as i know, in the process, the software does the AAC encoding first, and then the downmixing. (i deducted that after reading xVid4psp logs, and checking the file with mediainfo)
Will the result be better if i downmix to stereo first, and encode AAC after?
About the normalization, how to obtain the best result ? Each channel serparately or the global audio file.
Everything is done with avisynth scripts or / and function. I know that one way to do it, is to use soxfilter.dll, but i can't understand the way it works, and it seems very hard to use.

Thanks to all for reading me, and for those who have advices to give.

thanks again.

Cyril MARTINEZ.

[Selur]

doesn't make any sense to:
1st decode the dts
2nd encode to aac
3rd decode aac
4th apply downmix
5th reencode again

------------------

There a bunch of posts in this subforum which describe different ways how to do dts->aac downmixing, also there are a bunch of gui for this purpose, which mainly use the soxfilter.dll or the sox cli to do the downmixing.
Personally I would:
Decode the audio with ffmpeg, pipe to sox, do the downmixing (+ audio adjustments) in sox, pipe to the aac encoder of my choice (qaac, fhgaac, neroaacenc, fdkaacenc,...) and use it to encode to the output I want.

Cu Selur

[tebasuna51]

the software does the AAC encoding first, and then the downmixing...
Will the result be better if i downmix to stereo first, and encode AAC after?

I don't know xVid4psp but I don't think so.
To downmix you need a uncompressed audio, then first decode the original audio, then downmix and after encode to AAC.

About the normalization, how to obtain the best result ? Each channel serparately or the global audio file.

If you normalize each channel separately you losse the correct balance between channels. You must normalize the global audio.

Everything is done with avisynth scripts or / and function. I know that one way to do it, is to use soxfilter.dll, but i can't understand the way it works, and it seems very hard to use.


Seems you want apply some Dynamic Range Compression to avoid low volume dialogs.
You can use the Compand Sox Filter, but yes, is not easy to use.

You can try for instance:
SoxFilter("compand 0.1,0.3 -90,-90,-70,-64,-43,-37,-31,-31,-21,-21,0,-20 0 0 0.1") # Film Standard

[cmartinez]

doesn't make any sense to:
1st decode the dts
2nd encode to aac
3rd decode aac
4th apply downmix
5th reencode again

I am agree with you, and it's not my choice. My purpose is to finally have a stereo track. But i've read in the log that the DTS track has been converted in AAC 6 channels. When i check my Movie file with mediainfo, it is said that audio track has only 2 channels. That's why my conclusion. Perhaps i've done somthing wrong. As i am a noob in avisynth, i could try to call my downmixing script before the audiodub argument.

Thank you very much for your help and advices.

Cyril MARTINEZ.

[cmartinez]

You can see in my reply to Selur why i thought it was firstly encoded to AAC (same channels number) then downmixed to 2 channels.

About the normalization separately for each channels, i didn't know that.

About the Compand Sox Filter, xVid4psp use it if i ask <ConvertToStereo()> in the script. But i wasn't able to find wich parameters were applied (firstly), and the result on the audio track wasn't very convincing (secondly).

I will run some tests with your parameters, and post the result.

Thank you very much for your time and your advice.

Cyril MARTINEZ.

PS : Do you know a tutoriel for Compand Sox Filter ?

[tebasuna51]

When you download the SoxFilter package (http://sourceforge.net/projects/sox/files/sox/) there are pdf's with technical info.

[cmartinez]

thank you very much for your help Tebasuna51.

Cyril MARTINEZ.