1

Is there a way I could check if an mp3 file of people talking, with some noise, was recorded fully as stereo, or maybe there were parts recorded as mono? In Audacity I could see content in both the left and the right channels, but can I somehow compare if they are exactly identical (mono) or they have slight differences (stereo)? I have a hypothesis that part of the file is recorded as mono, and part of it - as stereo.

How could I check this?

I tried bitwise comparison, but it feels it is too granular. Also, I am not sure if splitting the right track and the left track in Audacity does not tweak the original left and right tracks.

I would appreciate any practical advice on the comparison between mono and stereo pieces of a record (or more like "between mono and stereo in general").

Improve this question
5
  • 2
    Your definitions are too simplistic. Much modern music is recorded in mono, which is then panned within the stereo field. Added effects from that are also likely to be stereo. Identifying a true stereo recording within that is unlikely to prove fruitful. Things like kick drum, bass & lead vocal are likely to be panned centre, so may fit almost your definition of mono.
    – Tetsujin
    Commented Jun 4 at 9:12
  • 2
    Maybe if we knew why you wanted to know, or what you were trying to accomplish, we could advise you better. You can always use the phase meter in Audacity to see how much stereo information there is at any given time, but in a busy arrangement, you probably won't be able tell one musical part from another. Also, as Tetsujin mentioned, you won't be able to tell the difference between a true stereo recording of a particular acoustic instrument and stereo processing of a mono source.
    – Data Processing
    Commented Jun 4 at 18:41
  • @DataProcessing I edited the description - it is an audio of people talking, with some noisy pauses, that I am trying to illustrate that is manipulated. I was hoping to prove it by visualizing that there are pieces that mono and pieces that are stereo. (There are pauses with added noise, I believe, so there I hoped that it would be mono)
    – Mateva
    Commented Jun 6 at 15:20
  • 1
    @Mateva, it appears that I was wrong about there being a phase meter built into Audacity. You could look for a free third party plug-in which offers this capability or switch to a different audio editor which has one. I'm sorry I can't recommend one for you, but you can do a web search or post a separate inquiry if you need help with that part. In any case, for the purpose you're describing, if there were mono sections of audio (not just silence) inserted into a stereo recording, those would stick out like a sore thumb in a phase meter, so I still think that's your best bet.
    – Data Processing
    Commented Jun 6 at 17:41
  • Huge thanks about the phase meter plug-in suggestion, @DataProcessing! I am trying to install lsp-plugins on top of Audacity on Linux, I will post the result if I succeed.
    – Mateva
    Commented Jun 7 at 8:52

2 Answers 2

Reset to default
2

I would suggest using a stereo field vector analyzer plugin. There are several available. Some are analysis tools only and others allow you to manipulate width, pan, etc.

This image (from MStereoScope) shows how the analysis of nearly-mono audio would look:

enter image description here

Panning the mono signal would just change the angle of the nearly linear graph towards L or R.

This image (from FLUX StereoTool) shows how stereo audio might look:

enter image description here

Audio that's mostly stereo with some mono components mixed would have sharp peaks in the pan direction of those components.

Here's a discussion on topic from the Audacity help forum: stereo analysis | Audacity Help Forum | Windows

Improve this answer
1

You can do this by converting the L/R stereo coding into M/S coding. The first channel will be the "mid" which is mono (the sum of the left and right channel) and the second channel will be the "side" (the difference between the left and right channel). A stereo track that was made entirely from mono or downmixed from mono will have a completely silent side channel. If some parts are recorded in stereo and some are not then the parts with a non-silent side channel will be the stereo parts.

Please do note that a low bitrate MP3 (or any lossy format really if the bitrate is too low) might downmix stereo into mono for some or all parts of the track if the codec thinks the side channel is not worth using bits on. If the bitrate is too low then you can't distinguish a silent side channel being from a mono recording or being from the codec downmixing to save bits, at least not with the decoded PCM (you might be able to get some information from syntax elements in the encoded format's frames but that is more complex and dependent on the behavior of the encoder).

How to do it with ffmpeg (tested on ffmpeg 7):

ffmpeg -i input.mp3 -af stereotools=mode='lr>ms' output.wav

How to do it with Audacity (tested on Audacity 3.2.4) after you've opened input.mp3:

  1. Select first track, "Edit>Duplicate"
  2. Select first track, "Tracks>Mix>Mix Stereo Down to Mono"
  3. Press second track name, "Split Stereo Track"
  4. Select third track, "Effect>Special>Invert"
  5. Press second track name, "Make Stereo Track"
  6. Select second track, "Tracks>Mix>Mix Stereo Down to Mono"
  7. Select first track name, "Make Stereo Track"
  8. Select first track, "Tracks>Mix>Mix And Render"
Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.