Import RAW Audio Showing Combined File, Trying To Get Isolated Track

I recorded a session with myself and someone else on a Zoom H4N with our Shure SM7B microphones each plugged into it and recording two separate MTR files in Mono each (WAV, 16 Bit / 44.1 Khz) into a SanDisk 32GB memory card. The H4N said it was recording fine the whole time (we were continually checking) but then when I put the memory card in to check on my computer to get the files, it wasn’t playing each file. The files on the memory card were saying one was 1GB and the other was 0KB. But in Disk Utility, it says there’s 2GB taken up on the card.

For the 1GB file, I was able to get it to work. However, the only time I can hear or get close to the 0KB file is when I import the RAW card image copy into Audacity, it brings it in as one combined file of the two recordings. So it’ll cut after a few seconds between each file. I need to get a separate proper isolated audio file of the other track.

The settings below are what get that combined file:
Encoding: Signed 16-bit PCM
Byte Order: Little-Endian
Channels: 1 Channel (Mono)
Start offset: 0 Bytes
Sample rate: 44100 Hz

It is pulling from the DMG of the memory card, so that may be why it’s drawing both (and since that 0B file seems to have something otherwise for regular importing). I tried making another image copy of the memory card but with deleting the other file (the 1GB) but it still brings in both files still.

However, based on being able to clearly hear the audio in the combined version, I know it’s in there somewhere. I just don’t know how to locate it or import it as an isolated single track. Any help on a way to import it but getting separate audio tracks?

I am using Mac 10.15.3 OS Catalina and Audacity 2.3.3.

“MTR” ?

“combined” in what way? Like two different recordings playing at the same time, or like two recordings playing one after the other, or combined in some other way?

MTR as in Multitrack

Combined, as I mentioned, as in it’s one long file that plays for a few seconds then switches to the other and back and repeats this process

Does it switch after exactly the same amount of time throughout the recording?
(for example, 0 to 2.5 second = mic 1, then 2.5 to 5.0 seconds = mic 2, then 5 to 7.5 seconds = mic 1, etc…)

Yes, that’s correct.

What’s the length of each section, measured in samples?

If we were able to extract all of the “mic 1” parts from the “mic 2” parts, then put all of the “mic 1” parts together and all of the “mic 2” parts together, would you have a complete show, or would there be parts missing (or damaged)?

If you can play back OK on the Zoom you can re-record by connecting the headphone output to line-in on an audio interface (or if you have a soundcard with line-in).

…Ironically, stand alone recorders are usually more reliable than computer recording just because there is a LOT more that can go-wrong in a computer (more settings/options and multitasking, etc.) and eventually it does.

The H4N said it was recording fine the whole time (we were continually checking)

:stuck_out_tongue: In the days of analog tape, if you had a 3-head recorder you could monitor what you just recorded a fraction of a second earlier. But you’re still in trouble if something goes wrong while recording a live event with no chance for “take-2”.

Unfortunately even if I were to extract all the pieces (which would be insanely time consuming), it’s not the smoothest transition when putting those pieces together and is a little jagged. So the isolated separate regular track is needed.

And it doesn’t play back on the Zoom itself, sadly.

I was thinking that we may be able to automate much of that so that it isn’t insanely time consuming.

If the pieces won’t fit together smoothly, then I think that’s the end of it. We can only recover from what is there, not from what isn’t.

I’m wondering why the Zoom might have failed in the first place. How long was the recording? Was it a 2 channel recording (one mic in one channel and one in the other)?

How would it be able to be automated?

It would be worth a shot to see and maybe trying millisecond calculation adjustments of where it cuts to see at least how smooth it could be.

IF the audio switches cleanly from one mic to the other mic at precisely constant intervals, then we could:

  1. Make a duplicate copy of the track
  2. Modulate the first track with a square wave that has precisely the same interval as the “switching” to mute alternate sections.
  3. Modulate the second track with a square wave that is 180 degrees out of phase with the first, to give us the other mic.
  4. Use “Truncate Silence” effect to remove the silences.

Steps 2 and 3 would be done with a Nyquist script (which we would need to write).

Should I send you the audio track to try it?

Three solutions to alternately chop at equal-intervals are covered in this thread … automated silence interval

Tried a few times coming back to this to make sense of it but I think I’m a bit out of my depth in trying those solutions myself…