How to know if audio track is mono or stereo?

[Name]

Hi all,

I got a video file where apparently the following 3 audio tracks are present:
German PCM 2.0 Stereo
German PCM 2.0 Mono
English PCM 2.0 Mono

The audio tracks have the following information:

Audio #1
Format                                   : PCM
Format settings                          : Little / Signed
Codec ID                                 : A_PCM/INT/LIT
Duration                                 : 1 h 40 min
Bit rate mode                            : Constant
Bit rate                                 : 1 536 kb/s
Channel(s)                               : 2 channels
Sampling rate                            : 48.0 kHz
Frame rate                               : 30.000 FPS (1600 SPF)
Bit depth                                : 16 bits
Stream size                              : 1.08 GiB (3%)
Title                                    : Stereo
Language                                 : German

Audio #2
Format                                   : PCM
Format settings                          : Little / Signed
Codec ID                                 : A_PCM/INT/LIT
Duration                                 : 1 h 40 min
Bit rate mode                            : Constant
Bit rate                                 : 1 536 kb/s
Channel(s)                               : 2 channels
Sampling rate                            : 48.0 kHz
Frame rate                               : 30.000 FPS (1600 SPF)
Bit depth                                : 16 bits
Stream size                              : 1.08 GiB (3%)
Title                                    : Stereo
Language                                 : German

Audio #3
Format                                   : PCM
Format settings                          : Little / Signed
Codec ID                                 : A_PCM/INT/LIT
Duration                                 : 1 h 40 min
Bit rate mode                            : Constant
Bit rate                                 : 1 536 kb/s
Channel(s)                               : 2 channels
Sampling rate                            : 48.0 kHz
Frame rate                               : 30.000 FPS (1600 SPF)
Bit depth                                : 16 bits
Stream size                              : 1.08 GiB (3%)
Title                                    : Stereo
Language                                 : English

Since the two German tracks show the same details, I assume that one is the "real" stereo track and the other one should be mono converted to stereo.


I thought I use FFMPEG with the following command to find out which one is the real stereo track:

ffmpeg -i input.wav -filter_complex "stereotools=phasel=1" -ac 1 output.wav

I would expect that the converted stereo shows a flat line, while the real stereo shows some differences.

But when I check now the output, I see one German track flat out, and the other two tracks with some wave forms.

For my understanding, the English track should also be flat out, or not?

Am I interpreting the FFMPEG command wrong, or what am I seeing here actually?


[Attachment 91474 - Click to enlarge]

[hello_hello]

In theory the English track should also be flat if it's mono in 2 channels (no waveform), but if the audio you're working with was encoded with a lossy encoder at some stage that'd probably be enough to cause it not be totally silent. Or even if there's a slight difference in phase or volume between the channels for some reason.

You can do what you're doing in Audacity itself.
Import an audio track.
Right click somewhere on the area to the left of the waveform and select "split stereo to mono" so you have two mono tracks.
Left click on the area to the left of one of the mono tracks to select it, then under the Effect menu select Special/Invert (at least that's where it is for my version of Audacity but I think the menus change now and them).
The waveform probably won't look any different but if it's truly mono audio in 2 channels, when you play the tracks together the output should be silent.

I just tried the above using a couple of old Billie Holiday tracks that should be mono, but after inverting a channel the volume didn't drop to zero, however it did drop very substantially and what remained didn't sound like the difference between two stereo channels. Maybe try it that way to see what your ears think.

Once you've inverted a track you can also select them both and then use the "Tracks/Mix/Mix And Render To A New Track" menu. This was the result after I did that for a Billy Holiday track.




And the same again for a stereo track.





They were music tracks though. It might be harder to tell the difference for video soundtracks if they're predominantly dialogue and therefore mostly mono anyway.

[Name]

I get the same result with your direct approach in Audacity.
In this case I believe the statement that the English track is mono, is simply wrong.
There must be 2 stereo and 1 mono.

Beside anyone else has an explanation for this behavior.

[davexnet]

Explanation? It looks like your FFmpeg method is sound. What let you to believe whether they were mono or stereo in the first place?

[cholla]

You might check this video with ffprobe.

They all say "stereo"
.
Why do you believe one is "mono" converted to "Stereo" ?

[timtape]

A "Stereo" file just has the capability/potential to carry a stereo signal. Formats like VHS/Beta Hifi Stereo/DVD etc were by design stereo capable. But many original audio recordings including TV and film sound tracks were mono. So when converted to a stereo capable format they will normally be "dual mono", the mono signal identical on each channel, left and right. Generally a merely mono capable format will have left and right channels of an original stereo recording mixed into a single mono file. So it can go either way: stereo can sometimes be mixed into mono, or mono often split into two identical left and right tracks in a stereo capable format, but still mono because that's what it natively is.

If it's a true stereo recording in a stereo file it will sound like stereo. In the end it's about the sound. The proof is in the listening.

[Name]

Why do you believe one is "mono" converted to "Stereo" ?

On the Blu-ray cover it is written. Since there was no indication which track was the stereo one (and I couldn't hear it ), I tried to find a way to find it out, and ended up with FFMPEG.

Now I just believe the English track might be stereo too.

[abolibibelot]

In theory the English track should also be flat if it's mono in 2 channels (no waveform), but if the audio you're working with was encoded losslessly at some stage that'd probably be enough to cause it not be totally silent.

Here you meant “if the audio [...] was not encoded losslessly at some stage”, right?
(Or: “if the audio [...] was encoded lossy-ly” if that's a word...)

[hello_hello]

Here you meant “if the audio [...] was not encoded losslessly at some stage”, right?
(Or: “if the audio [...] was encoded lossy-ly” if that's a word...)

Yeah I meant if it was encoded lossily, or with a lossy encoder.
I don't know where my head was at, but I'll edit the post to correct it.

[hello_hello]

Actually it looks like I was talking crap anyway.
I just made a file that was definitely mono in 2 channels, exported it as a wave file, and then re-encoded the exported audio as an MP3 and again as AAC.
I imported both versions and inverted one of the channels, and in both cases the output was still totally silent, so lossy encoding of a track that's mono in 2 channels doesn't seem to make a difference.

[davexnet]

Actually it looks like I was talking crap anyway.
I just made a file that was definitely mono in 2 channels, exported it as a wave file, and then re-encoded the exported audio as an MP3 and again as AAC.
I imported both versions and inverted one of the channels, and in both cases the output was still totally silent, so lossy encoding of a track that's mono in 2 channels doesn't seem to make a difference.

Yes. I did the same, taking a stereo file amd converting to 2-channel mono (takes about three steps in Audacity)
Invert one channel and convert to mono - it's zeroed out.

Audacity works with uncompressed digital samples, so the compressed/uncompressed properties of the source
seem to be irrelevant (as you say)