Sox Audio Effect Filters v1.1 beta

[manolito]

Thanks for the links...
I did not know that there even exists a standalone CLI version, on Moitah's main site you cannot find it.

It should be easy to integrate this into my AVSToDVD plugin. Just one question about the -b option:
The command line help says:
Option to put bass in side channels, useful if your goal is vocal removal.

Vocal removal is not exactly what I have in mind, do you believe this option should be used anyways?


Cheers
manolito

[manolito]

OK, I integrated the CenterCut method into my AVStoDVD plugin, no major problems. The only thing which is a little bit annoying is that only 16bit int WAV files are accepted as input. Feeding 24bit int or 32bit float files to CenterCutCL.exe causes an error.

At this point I am at the end of my wits because I have no way to test the different upmix methods on a real home theater setup. I uploaded some converted AC3 files here:
http://www23.zippyshare.com/v/B0MHag96/file.html

Maybe someone finds some time to play the files through a good quality multichannel system.

The source is a typical low quality clip soundtrack, mainly synthesized stuff. When I play the files on my stereo setup (downmix done by the Windows Sound Mapper), only the "UpAction" file sticks out because the strong phasing is unbearable. The CenterCut method with the -b option sounds almost too good, the bass frequencies are overemphasized quite a bit.


Cheers
manolito

[tebasuna51]

The CenterCut method with the -b option sounds almost too good, the bass frequencies are overemphasized quite a bit.
Of course, there are 3 channels (FL,FR,LFE) with bass frequencies, without the -b there are only 2 (FC,LFE).

Like I say I recommend the -b option with LFE empty, the extraction of bass frecuencies with CenterCut to FC can be different to sox extraction to LFE.
When the audio system extract the bass from FC to add to LFE to send the output to subwoofer maybe there are problems.
If you let bass in FL-FR untouched the extraction is sure.

In your samples always the SL channel have high volume than SR channel
Max peaks in -> SL SR
---- ----
CenterCut -b.ac3 0.8 0.3
CenterCut.ac3 0.9 0.26
SoundOnSound.ac3 0.36 0.13
UpAction.ac3 0.5 0.3

I see your bat file and I can't understand because always use:

rear.wav -c 1 rear_left remix -m 1v0.473,2v-0.473 delay 0.02
rear.wav -c 1 rear_right remix -m 1v0.473,2v-0.473 delay 0.02

or

rear.wav -c 1 rear_left remix -m 1v0.668,2v-0.668 delay 0.02
rear.wav -c 1 rear_right remix -m 1v0.668,2v-0.668 delay 0.02

and always seems SL = SR

[manolito]

In your samples always the SL channel have high volume than SR channel
Max peaks in -> SL SR
---- ----
CenterCut -b.ac3 0.8 0.3
CenterCut.ac3 0.9 0.26
SoundOnSound.ac3 0.36 0.13
UpAction.ac3 0.5 0.3

I see your bat file and I can't understand because always use:


Could it be that Aften has its own mind in determining which (inaudible) peaks to keep or to discard?

I tested this using the temp WAV files, loading them into WaveLab and performing a Global Analysis on them. SL and SR always came out with exactly the same peak values.


Otherwise I tried your suggestion for CenterCut to use the -b parameter together with an empty LFE channel. The stereo downmix still had more bass than the source, but a little less than with the LFE channel. Probably irrelevant because the downmix parameters have an influence on the channel balance.


Cheers
manolito