I have imported the audio section from an AC3 5.1 track. (originated from a recording TV broadcast and stored in the container MPEG-TS (.ts))
It results in 6 mono tracks (please see the attached foto). Does anyone know what to do which each mono track in a “mix-down” to a stereo track in the most optimally way ?
There is a table [u]here[/u].
Translated to dB:
Front Left 0dB → Left
Front Right 0dB → Right (corrected)
Surround Left -3dB → Left
Surround Right -3dB → Right
Center -3dB → Left & Right
LFE not used*
\
- That doesn’t mean there is no bass. The regular bass is contained in the other 5 channels. The “point one” channel is LFE for “extra” Low Frequency Effects. (On most home theater systems with “small” surround speakers, the regular bass from the 5 surround channels plus the LFE all gets sent to the subwoofer.)
Thanks for the reply
Yes I have also see something similar to this:
Front Left 0dB → Left
Front Right 0dB → Left ---- You mean right
Surround Left -3dB → Left
Surround Right -3dB → Right
Center -3dB → Left & Right
But I have 6 mono tracks - So just to be sure - Are “Sample1 track”, the Front Left? (see the attached photo)
And so on:
Sample2 track = Front Right
Sample4 track = Surround Left
Sample5 track = Surround Right
Sample3 track = Center
And the last question (for now) When i divide the Center track between channels 1 and 2 (after -3dB) will it then be 50% to both or 100% to both?
But I have 6 mono tracks - So just to be sure - Are “Sample1 track”, the Front Left? (see the attached photo)
That, I don’t know. I’d guess the “weak” channel is the LFE but that one should be easy to identify by listening.
The other two lower-level channels are probably the rear but I don’t know which one is left or right.
If it’s a move the main dialog should be in the center.
And the last question (for now) When i divide the Center track between channels 1 and 2 (after -3dB) will it then be 50% to both or 100% to both?
-3dB is half power or 0.707 times the amplitude. So if you had 10 Watts going to the center channel and you move it to left & right at -3dB, that’s 5W in each channel or 10W total again.
The rear channels are reduced in volume when downmixed.
One more thing. It looks like you have plenty of headroom but re-import and check for clipping after exporting (because you are increasing the left & right volume.) Or, export to 32-bit floating-point WAV (which can go over 0dB without clipping). Re-import and Amplify/Normalize to bring the levels down (if necessary) before exporting to your desired final format.
There’s a short-form for this.
Center is usually the only channel that talks. Reduce by half (-6dB) and add it to the two main channels. It should never overload because there is no or very little voice in Main Left and Right. Check them for overload anyway. A little graceful Limiter may be called for. You should not be able to hear it working.
Surround L and R are usually echoes, reverb and environment trash. Leave them out.
Unless you can deal with earthquakes and thunder, you can leave out the Low Frequency Effects LFE.
So that’s three channels which should be “reasonably” easy to identify.
If you can’t split up the channels easily like that, then somebody may have done a bad job of making the 5.1. Then it’s entirely up to you. Good luck.
Koz
In your illustration, I would ignore the bottom three tracks. I would guess 3 is Center. Reduce by 6dB and add it to both tracks 1 and 2.
Check for overload.
That thing with the voices isn’t as strange as you think. The last thing you want the audience to do is flap their heads back and forth trying to follow the dialog. The camera will almost always put the lead actor(s) in the center or prime area of the screen.
Koz
Hmmm it might be appropriate to define further what type of sound the audio file contains.
As I said, it’s a TV brodcast. It features a live recording of a semi-acoustic recording with song and instrumental accompaniment. So there is no talk or dialogue. I just wanted to save the soundtrack in the best stereo quality as possible
AND again thank you very much for your nice response
Still, the rear two are usually taken up with Room Ambience, Echo and Reverb, the exact thing that most recording engineers spend their lives trying to avoid.
So the real question is what did they do to the Center. If it’s a video music show, my bet would be to follow the video shots, which might give you oddball effects as the camera switches between the instruments and singers.
I’d still cut track 3 in half and add it to the two front channels. Check for overload. Ignore everything else. You can set View > Show Clipping ON and it will give you red marks where the channels have gotten too loud.
I would reduce the volume of everything until they go away. If you try to fix it locally one at a time, you will be stuck remixing the performance and that way lies madness.
Post back what you decided to do and how it went.
Koz
It’s amazing that you’re so wonderfully curious;) And I would like… - and I’m really thankfully to post the results (see/hear the attached)
The TV broadcast I recorded contains 2 audio files: a 192 kbps mp2 stereo and then the aforementioned AC3 6 channel audio at 448 kbps.
192 kbps MP2 is not the best quality in the world - so I would like to capture the audio from the AC3 track and save it as a Stereo file (I’ve used OGG)
I have now created three files:
The first one I Reduced the center channel with the recommended 6dB and mixed all 5 tracks down to a stereo track (called “Mix down 5 to 2.ogg”)
The second I just used the two front channels (“Mix down 2 front to 2.ogg”)
and finally in the third I just extracted the sound from the mp2 track and saved the file (“MP2 track extract”) (you could call that for a reference for the way the mix down should sound - luckily, they have “sent” the “reference” - though it was in a slightly poor quality)
Please try listening. I personally find it difficult to judge
PS Let’s not spend time writing about distortion / clipping, etc. I think I have a good idea of that stuff 
.
Mix down Samples.zip (1.96 MB)
I personally find it difficult to judge
I can’t hear any difference.
Koz
I did a frequency analysis. Analyze > Plot Spectrum.
There are very minor tonal distribution differences between the MP2 version and the ordinary stereo mixdown. The most important difference is the MP2 only has tones up to 16,631Hz and the mix from the front two has tones up to 18,300Hz. So your dog may know which one you picked.
Koz
Ha ha - or maybe my son at 3