AC3 to stereo aiff

In the past I have used mac3dec to convert ac3 file to stereo aiff file.
Am trying to find a substitute.

Does not work on upgraded setup Early 2009 mac mini intel
rosetta translator does not work for all PPC apps.

Actually the mac3dec part seems to work
just did not realize it was really a link to sound converter.
it is the sound converter that is “missing”
soundconverter apparently lives happily on as linux but that’s not me.

Audacity with attached library can open ac3 files
I got as far as saving as a mono file.

wikipedia has a complicated filter to translate ac3 to stereo
which may be in the attached library but …

am I missing the easy answer or is the answer no?
Audacity does not do ac3 to stereo aiff.

Me no speak Mac , but If AIFF is not shown on the drop down menu shown below ( when exporting audio from Audacity) …
save in AIFF format option in Audacity in Windows.gif
Then I think you need to install something called FFmpeg into Audacity to be able to export any audio from Audacity in the AIFF format …[u]How_do_I_download_and_install_the_FFmpeg_Import.2FExport_Library.3F[/u]

[ IIRC AIFF is native to Audacity : you shouldn’t have to install FFmpeg to export in that format ]

Thanks for you time.

Already installed library and split AC3 file into 5 tracks.

Did not see any easy way to combine into 2 stereo tracks and export as air.

Will look again but am not holding my breath.

Maybe FFmpeg does this but looks pretty scary to me.
The box does not come over so well, but it is how to demux? the 5 into Left Total and Right Total

The Dolby Stereo Matrix is straightforward: the four original channels: Left (L), Center (C), Right (R), and Surround (S), are combined into two,
known as Left-total (LT) and Right-total (RT) by this formula:

Dolby Stereo Mix Left Right Center Surround
Left Total 1 0 frac {1}{sqrt 2} +j frac {1}{sqrt 2}
Right Total 0 1 frac {1}{sqrt 2} -j frac {1}{sqrt 2}

where j = 90° phase-shift

I presume that you mean “as AIFF”?

You don’t need FFmpeg to export AIFF (mono or stereo).

When you Export from Audacity, all tracks that are not muted are mixed down to a single file.
If all of the tracks are “mono” tracks, then the exported file will be mono.
If any of the tracks are stereo, or set to “Left Channel” or “Right Channel” or panned left or right, the the Exported file will be stereo.

To set a mono track to “Left Channel” or “Right Channel”, use the track dropdown menu:
To combine two mono tracks into a stereo track, use “Make Stereo Track” (also in the track dropdown menu).
To “pan” a track left or right, use the track “Pan Slider”:

When you have the tracks and channels as you want them, export from Audacity and select AIFF as the export format.

I think the magic word is “downmix” …

[ I’ve not downmixed 5.1 to stereo , DVDdoug may be your man for that job ]

I’m sure I’ve seen plugins to do that , e.g.
2.102 Surround matrix encoder (surroundEncoder, 1401)
but I’ve never used that myself.

Yeah, this is the wiki info I was looking for audio mixing

Stereo downmixes / fold-downs[edit]
Left total /Right total (Lt/Rt)[edit]
Lt/Rt is a downmix suitable for decoding with a Dolby Pro Logic decoder to obtain 5.1 channels again. Lt/Rt is also suitable for stereophonic sound playback on a hi-fi or on headphones as it is.

Lt = L + -3dBC + -3dB(-Ls -Rs)

Rt = R + -3dBC + -3dB(Ls + Rs)

(where Ls and Rs are phase shifted 90°)[7]

Have no idea what mac3dec did, total black box
open ac3 file save as stereo aiff.

For all I know it may have just saved “L” and “R” whatever the heck that means.

Now that it does not work on intel, do know it used sound converter which seems to still be available in the linux world.
Rosetta can not / will not run sound converter.

I have my ppc setup , I can go back
but was hoping I could do something with Audacity

here is as far as I get when exporting as apple aiff

“Your tracks will be mixed down to a single mono channel in the exported file”

On the other hand,
the choices on FFmpeg export are beyond bewildering.

There are formats: “aiff” is one
12 codecs and
10 general options

nothing that says stereo aiff you dummy ; )

That’s because all of your tracks are mono.
In my previous post I described how to create a stereo mix down.

As I wrote previously, you don’t need to use FFmpeg to export AIFF files.

Ok can see that can make tracks left channel or right channel
I am looking at 6 tracks
5 make sound, 1 does not seem to be able to be played by my computer.
(its solo is very quiet)

Any idea which track goes where?
Is it
left “highlight”
right “highlight”

You do need FFmpeg to open the AC3 file.

I am not trying to do anything fancy.
I used mac3dec to make a copy of the Buffy singing episode for my iPod. : )

I am not trying to do anything fancy

Yes, you are. You happen to have a program line-up that makes it look easy. In addition to the obvious L and R, you have dialog almost always in the Center (separate track), low frequency effects (rumble and boom) and Dial Norm which is the automatic volume control that allows the system to simulate the real sound of thunder and jets taking off. All that is or can be hiding in a surround AC3 performance and must be untangled for a stereo mixdown.

My fuzzy memory is that nobody tries to re-use the surround tracks. I could be corrected.


We have different ideas about easy.
Easy to me is here are 6 tracks in this AC3 file,
do you want to output mono, stereo or same number of tracks as source?

I guess I worry mac3dec > soundconverter may have done a really bad conversion
but it sounded ok to me and something is better than nothing.

Surely there is some rhyme or reason about this format
so that some kind of default 2 channel split could be done.

Easy to me is here are 6 tracks in this AC3 file,

Seven counting DialNorm.


I can give you the poor version.

Left goes to Left, Right goes to Right. Center is added to both Left and Right. Throw away Low Frequency Effects and both of the Surround Channels.

There’s a trick with the Center Channel Dialog. I think you make that into an Audacity mono track if it’s not there already and slide the panning adjustment to Anything Else and then back to the middle. That will force the track into center panned stereo (secret aural teachings).

Pray to goodness they didn’t have explosions, thunder or jets.

That should give you a not awful stereo mixdown.


Yeah. Here it is.

Tracks top to bottom. L, R, C. Throw away everything else.

From the drop-down menus on the left.

Top track: Make Left
Next Track: Make Right

Bottom Track: Mono (if it isn’t already).

Top track again: Make Stereo Track.

Bottom track: Pan off center and then back.

File Export. It will warn you it is going to mix everything down to a stereo show. [OK].
Name the file.

Pick a test with dialog so you know if it worked.


What’s that about koz?
If a track in audacity is mono, it’s mono - no need to mess about panning it off centre and back again.

and from pete’s other posts, the 5 tracks are clearly all mono because that was his original problem - Audacity was mixing down to mono.

The usual difficulty mixing down multi-channel audio is working out which channel is which (which goes where).
Assuming that your AAC decoder split AAC in the “correct” order the first three channels should be Left, Right and Centre, but they may not be in that order. Different decoders will often swap the channels around.

If there is speech in the recording, that is usually mostly in the centre channel. If you get left and right the wrong way round it is not likely to matter much unless there are sounds that you know should be on the left or right.

These three channels are the most important for a downmix, and you should be able to get good results using only these three tracks.

The back left and back right speakers are likely to be much quieter than the front and centre channels.
The LF channel will sound bassy and muffled.

To listen to just one track at a time, use the track “solo” button (in the panel on the left end of the track).

When you have decided which channel is which:
Pan the left channel all the way to the left using the track pan slider.
Pan the right channel all the way to the right using the track pan slider.
Leave the pan slider central for the centre channel (no need to do koz’s dance with the slider).

The other three channels may be closed (click on the [X] in the top left corner of the track)

Optionally, you may also want to adjust the amount of each channel. To do that, use the “track gain” slider (immediately above the “track pan” slider.
When you are happy with how it sounds, export it. The result will be a stereo track.

Using the Pan controls, doesn’t each side in the final show effectively double in volume? Add the dialog and the possibility of overload goes way up. That’s why I picked "Make Left, etc. That’ stays the same volume.

One of the two techniques is going to have a mono/dialog imbalance compared with the rest of the show. It makes my head hurt to figure out which one.

Yes the sliding off center dance is redundant.


Thanks very much for your help.

Am using the FFmpeg library to split the AC3 file.
only see six tracks not seven.

looks like first(top) is L/R
second is L/R
third is C enter
fourth has a silent solo +/- 0.1
fifth has music
sixth has music
the movie is Blues Brothers, so car chases and maybe an explosion
but only interested in the music.

Is there no way FFmpeg can pull out track “annotation”,
some kind of name like “left track”. etc.

L R and C is a start
but I am afraid some background vocals are trapped on the fifth and sixth tracks.
The bottom 2 tracks are +/- 0.25.

Should I treat them like the third Center track
or will real explosions ensue?

I will poke at FFmpeg to see if any annotation can be coaxed from the file.

ok looked at FFmpeg and 180 page ATSC documentation

so is simpler and not so simple.

There is
Left Surround
Right Surround
Low Frequency Effects.

I plan to drop LFE and smoosh the others into 2 tracks
following suggestions given earlier.

The best way would seem to be to master the FFmpeg commands and calculate a 2 channel file
following standard practices
or have software that does that for you.

My guess:
1: L
2: R
3: C
4: LF
5: LS (left surround)
6: RS (right surround)

I’d suggest working with just 1, 2 and 3. “LF” is not always used - video players can often “extract” low frequency sound from the other channels when LF is not present. LS and RS will not add much to the stereo mix, and may make the sound a bit echoey.

No. Audacity is a little unusual here, but in this case it is to our advantage. When you pan a track to the left, it makes the right side quieter - and no change to the left side. When panned all the way to the left, the right side is silent, and the left side is unchanged.

Mine too, so I looked it up :wink:
A good starting point is (with the track gain sliders), drop the centre channel to about -6 dB and drop the left and right channels to about -12 dB (panned hard left and hard right). Tweak it from there to taste.

Tip 1. Make the playback meter wide The meter toolbar can be dragged out of its docking position, resized and re-docked. When tweaking the levels, keep the peaks below 0 dB so as to avoid distortion (“clipping”)

Tip 2. The track gain sliders can be a bit fiddly, but “View > Mixer Board” provides a BIG version of the track gain sliders

Yes, much of this can be done directly in FFmpeg. The big boys use advanced options in FFmpeg (or libavformat or Mencoder) to tweak the results, but that is scary stuff. If you will be doing a lot f this, then it’s probably worth mastering the intricacies of FFmpeg, but otherwise it’s much easier to do in Audacity once you have the basics under your belt. (libavformat is a fork from FFmpeg that is gaining popularity on Linux).