Convert 3/1 (4.0) to 3/2 (5.0)

[Xorp]

What's the best way to convert 3/1 (4.0) to 3/2 (5.0)? Because a lot of players still mess up 3/1 and play it as 2/2, I want all my 3/1(4.0) stuff to be 3/2(5.0)

I know the start is:
-decode DTS-MA/PCM 4.0 to wavs
-make a copy of the rear channel, label them BL and BR.

then what?

[LigH]

-make a copy of the rear channel, label them BL and BR.

then what?

Then you will have to encode some compressed audio format again from the per-channel WAVs. You may not easily purchase any dts encoder, but as a free AC3 encoder, you can use Aften and different GUIs supporting it, like EncWAVtoAC3 (https://github.com/wieslawsoltes/wavtoac3encoder/releases), it has a multiplex interface where you can select channel WAVs and build a call for Aften with MUX file (AFAIR, long ago...).

[tebasuna51]

Instead copy BC and create BL and BR with the same volume you must atenuate the volume to preserve the global balance.

eac3to BC.wav BL.wav -3dB
eac3to BC.wav BR.wav -3dB

[Thunderbolt8]

would you also have to attenuate channels if you transform a mono 1.0 track into a mono 2.0 track by simply copying the center channel and making L, R out of it?

[tebasuna51]

would you also have to attenuate channels if you transform a mono 1.0 track into a mono 2.0 track by simply copying the center channel and making L, R out of it?

That is not needed, only optative, because don't change the balance.

In 3/1 to 3/2 you duplicate the rear volume, maintaining the front volume, if not attenuate.

[pandy]

I would do:
Copy Back Surround to BL and BR, then BL-(R*F), BR-(L*F) where F is a constant (for example -12 ... -24dB).

[pandy]

would you also have to attenuate channels if you transform a mono 1.0 track into a mono 2.0 track by simply copying the center channel and making L, R out of it?

Yes, signal should be reduced by 3.103dB (1/SQRT(2)) .

[tebasuna51]

I would do:
Copy Back Surround to BL and BR, then BL-(R*F), BR-(L*F) where F is a constant (for example -12 ... -24dB).

What is this? That mixes don't have sense for me.

The question is emulate the source, not create new effects.

[AlexKane]

-make a copy of the rear channel, label them BL and BR.
then what?

After creating BL & BR from your (original) surround channel, you only need to attenuate both BL & BR by 3.103dB. If you choose to not apply attenuation, your new surround channels will be 3.103dB louder compared to the original.

[pandy]

What is this? That mixes don't have sense for me.

The question is emulate the source, not create new effects.

The question is:

What's the best way to convert 3/1 (4.0) to 3/2 (5.0)?

IMHO best way - of course this is my personal preference best for me. :D

[tebasuna51]

IMHO best way - of course this is my personal preference best for me. :D

Ok.

Well, for that exist the rule:

12) How NOT to post on this forum:
...
Do not ask "what's best" because this question cannot be answered objectively. Each and everyone has their own view about what's best in a certain area. The best is what works best for you!

[pandy]

Ok.

Well, for that exist the rule:

Yes but seem that OP ignored this rule and moderators accepted this... :D don't shoot messenger...

[tebasuna51]

But my obligation like moderator is say than BL-(R*F), BR-(L*F) is, maybe, the best for pandy, but not to preserve the original sound.

[pandy]

But my obligation like moderator is say than BL-(R*F), BR-(L*F) is, maybe, the best for pandy, but not to preserve the original sound.

And? Best =/= preservation.

but OK - feel free to remove my answers in this topic - have no will to argue with moderator.

[tebasuna51]

We can discuss about a topic, puting the arguments, and the users can choice the best for each one.
I am a moderator but I do not want impose my arguments, I want convince.
And I not need remove any answer if are technical arguments without personal offenses.

- Now my arguments about:
BL = 0.707 * BC
BR = 0.707 * BC
(0.707 = -3.01dB)

When we listen the same sound in BL and BR speakers our brain listen a phantom channel at the middle position. Just the position of the BC speaker.
The acustic power is proportional to the square value of volume, and added the two channels:
(0.707*BC)^2 + (0.707*BC)^2 = BC^2
The same than provided by the original BC channel.

Then with a 3/2.0 system we listen the same than we have a 3/1 system, like original source was.

- Now some arguments for what I don't like your method:
BL = BC - k*FR
BR = BC - k*FL
where k is a constant (for example -12 ... -24dB).

1) The acustic power is always big than original sound

2) Maybe this mix enforce the surround effect with some, not always for all, music sound.
But here we speak about sound tracks from movies and we can obtain undesired effects.
When a actor are at the left side of the screen many dialog volume go to FL speaker. For what we need listen this dialog (added or substracted) from the BR speaker?

[pandy]

We can discuss about a topic, puting the arguments, and the users can choice the best for each one.
I am a moderator but I do not want impose my arguments, I want convince.
And I not need remove any answer if are technical arguments without personal offenses.

- Now my arguments about:
BL = 0.707 * BC
BR = 0.707 * BC
(0.707 = -3.01dB)

When we listen the same sound in BL and BR speakers our brain listen a phantom channel at the middle position. Just the position of the BC speaker.
The acustic power is proportional to the square value of volume, and added the two channels:
(0.707*BC)^2 + (0.707*BC)^2 = BC^2
The same than provided by the original BC channel.

Then with a 3/2.0 system we listen the same than we have a 3/1 system, like original source was.

- Now some arguments for what I don't like your method:
BL = BC - k*FR
BR = BC - k*FL
where k is a constant (for example -12 ... -24dB).

1) The acustic power is always big than original sound

2) Maybe this mix enforce the surround effect with some, not always for all, music sound.
But here we speak about sound tracks from movies and we can obtain undesired effects.
When a actor are at the left side of the screen many dialog volume go to FL speaker. For what we need listen this dialog (added or substracted) from the BR speaker?

My point was only to express that involving rule 12 at end of discussion have no sense.

I've tried to find Dolby recommendation how to perform mixing for OP problem but can't find anything useful.
From my perspective separating Rear Surround by partially removing energy from Front speaker may improve channel separation - that's all.

[Xorp]

Instead copy BC and create BL and BR with the same volume you must atenuate the volume to preserve the global balance.

eac3to BC.wav BL.wav -3dB
eac3to BC.wav BR.wav -3dB

Thank you, I forgot I made this thread from November...

What software will make the 3/2 layout audio I'm desiring?

[tebasuna51]

When you have the 5 channels: FL.wav, FR.wav, FC.wav, BL.wav and BR.wav you can convert them to:

Lossless flac:
ffmpeg -i FL.wav -i FR.wav -i FC.wav -i BL.wav -i BR.wav -filter_complex "[0:0][1:0][2:0][3:0][4:0] amerge=inputs=5" -acodec flac -ac 5 output.flac

Lossy AC3 (the more compatible for players):
ffmpeg -i FL.wav -i FR.wav -i FC.wav -i BL.wav -i BR.wav -filter_complex "[0:0][1:0][2:0][3:0][4:0] amerge=inputs=5" -acodec ac3 -ac 5 -ab 640k -center_mixlev 0.707 output.ac3

Or DTS (Lossless or lossy) if you own the DTS-HD Master Audio Suite.

[Xorp]

When you have the 5 channels: FL.wav, FR.wav, FC.wav, BL.wav and BR.wav you can convert them to:

Lossless flac:


Lossy AC3 (the more compatible for players):


Or DTS (Lossless or lossy) if you own the DTS-HD Master Audio Suite.

Thank you! In the past it's been recommended to use SoX -M (merge) with the wavs, is ffmpeg the superior way to do this? Or equivalent?

[tebasuna51]

The merge method is equivalent. Is a lossless operation.