Stereo Downmix with Besweet

[Richard1485]

I am trying to downmix 5.1 to stereo (not DPLI or DPLII) with Besweet v1.5b31. I have six mono wavs and have made a .mux file.

The channel order in the .mux file is as follows.

"C:\L.wav"
"C:\R.wav"
"C:\C.wav"
"C:\LFE.wav"
"C:\SL.wav"
"C:\SR.wav"

I am using the following command line.

"BeSweet.exe" -core( -input "file.mux" -output "file.wav" -2ch -logfile "file.log" ) -azid( -s stereo --maximize ) -ota( -g max )

When I input my command line, the process that finds the maximum gain runs until completion, but when "transcoding..." appears nothing happens, no matter how long I leave it. An empty file appears on my hard drive that is labelled SR_1.WAV and the logfile suggests that nothing has been transcoded.

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : C:\.mux
[00:00:00:000] | Output: C:\SR_1.wav
[00:00:00:000] | Floating-Point Process: No
[00:00:00:000] | Overall Track Gain: 0.714dB
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +---------------------

It does not appear that anything other than the last wav in the mux file (i.e. SR.wav) is being read. I also tried the Belight GUI, but the same happens. When I load the .mux file into Belight, the filename given in the GUI is SR.Wav and the description next to it is WAV - Mono - 768kbps - 48000Hz, which deepens my suspicion that only that last wav is being read.

I am aware of other methods of downmixing to stereo but am unhappy with the output and want to try Besweet's to compare. Any help would be appreciated.

Operating system: Windows 7 Home Premium (64bit)

[tebasuna51]

Only AC3 sources can be downmixed, because is the Azid decoder the only soft included with BeLight/BeSweet than can do the job.

Please explain your problem with other downmixing methods.
You can obtain the 3 Azid methods (stereo, dpl I, dpl II), or make your personal downmix, with BeHappy.

[Richard1485]

Only AC3 sources can be downmixed, because is the Azid decoder the only soft included with BeLight/BeSweet than can do the job.

Ah! That would explain the problem. My original file is TrueHD and Besweet didn't like that either. I didn't realize that it could downmix only AC-3.

Please explain your problem with other downmixing methods. You can obtain the 3 Azid methods (stereo, dpl I, dpl II), or make your personal downmix, with BeHappy.

When I downmix to stereo with Behappy, the result does not sound as balanced as a DPLII downmix made with Eac3to. Some sound effects are too prominent, noises that sound as if they should be in the background seem too loud, in some places voices sound too quiet, and the volume level seems to change from scene to scene, as if the dynamic range is too high – even if I encode the resulting wav to AC-3 using a DRC profile.

This is the procedure I followed with Behappy.

Downmix
Normalize to 100%
Convert sample to 24 bit int

[tebasuna51]

With BeHappy you can select between modes: DPL II, DPL II (LFE), DPL I and Stereo

If you select DPL II and Normalize 100% is the same than use eac3to -down2 -normalize
(I passed the same parameters from BeHappy to Madshi to make the downmix)

[Richard1485]

With BeHappy you can select between modes: DPL II, DPL II (LFE), DPL I and Stereo

Yes, I know. That's the point. I'm choosing stereo (not DPLII) in Behappy and comparing the result to the DPLII downmix in Eac3to, and the pure stereo downmix doesn't sound as good as the DPLII downmix for the reasons I mentioned above.

By the way, I thought that Eac3to normalizes automatically when -down2 is selected and that -Normalize is no longer needed. Is this not true?

[tebasuna51]

Yes, I know. That's the point. I'm choosing stereo (not DPLII) in Behappy and comparing the result to the DPLII downmix in Eac3to, and the pure stereo downmix doesn't sound as good as the DPLII downmix for the reasons I mentioned above.

Of course the mix is different, but if is better or not is a question of your taste.

The mix for DPL II (BeHappy or eac3to) is:

fl' = 0.3254xFL + 0.2301xFC + 0.2818xSL + 0.1627xSR
fr' = 0.3254xFR + 0.2301xFC - 0.1627xSL - 0.2818xSR

The mix for Stereo in BeHappy is:

fl' = 0.2929xFL + 0.2071xFC + 0.2071xLFE + 0.2929xSL
fr' = 0.2929xFR + 0.2071xFC + 0.2071xLFE + 0.2929xSR

I can add to BeHappy other personalized mix if you want.

By the way, I thought that Eac3to normalizes automatically when -down2 is selected and that -Normalize is no longer needed. Is this not true?

Nope, you need -normalize.

[Richard1485]

I can add to BeHappy other personalized mix if you want.

Thank you for the generous offer. :)

One difference seems to be that stereo downmixes the LFE, whereas the DPLII does not. In terms of what you posted above, I do not understand how the center and surround channel are matrixed in with DPLII and how to avoid this matrixing in when making a pure stereo mix.

Thanks for the information about the -Normalize switch in Eac3to.

[robertcollier4]

eac3to.exe E:\Sourcefile.ac3 E:\Outfile.wav -down2 -mixlfe