Windows 10, Audacity 3.1.3.
I have two FLAC files, the original was recorded mid-April, the second is an edited version of the original recorded at the end of July. I was asked to make four small changes. The original track is 51m08s, but the edited track is 51m37s. My proof-listener and I both are puzzled, because we do not see twenty-nine seconds of newer text in the edits. One edit was to delete text, so four edits with a net increase of 29 seconds is more mysterious.
My recording studio (“The study”) is an acoustic disaster as a natural consequence of increasing the window glass area in a nine-foot by ten-foot room from 725 square inches to 5,610 square inches and removing all but three bookcases
My theory is that recording in my bedroom instead of my old study might mean a change in posture etc resulting ion a slower mode of the spoken text.
The bulk of the 51-minute track is unchanged. Or ought to be. It is the original FLAC and so, I think, should match bit-wise at least until the first edit.
Is there a simple way to have Audacity compare two tracks bit-wise, and then halt at the first mis-match?
If so I could use this technique just four times and check that the only bit-wise mis-matches occur at the start of each scheduled edit, and more importantly, that the immediately surrounding audio is as it should be.
I did a search for topics with “compare” over the past five years, and I think that my situation differs from those examples given.
You can open both tracks in Audacity. The trick is to use File → Import → Audio for the 2nd track instead of using “open”.
When you hit the Play button they will play at the same time or there are mute buttons to left of the waveform if you want to mute one at a time.
Perhaps easier to use the “Solo” buttons to play just one of the tracks at any one time
DVDDoug>> “You can open both tracks in Audacity. The trick is to use File → Import → Audio for the 2nd track instead of using “open” When you hit the Play button they will play at the same time or there are mute buttons to left of the waveform if you want to mute one at a time.”
Hi Doug, solutions similar to this I have seen here and here, but all these solutions involve a human judgment that is non-binary. I don’t need, nor do I want, to listen to the first 30 minutes of what is identical audio. Note that I cloned the April track and made what believe to be minor patches at four specific locations.
I am looking for a mechanical, bit-for-bit comparison of that leading 30 minutes, that ought to be an exact copy of itself. FLAC files, remember, not MP3.
Had I made no changes, then the two tracks should be bit-wise absolutely equivalent. Right to the point where my first (of four) changes was made, I expect the FLAC tracks to be bit-wise identical. Why would then not be identical, bit for bit?
If Audacity can compare two tracks bit-wise and halt at the first discrepancy, then I can select the area where the discrepancy occurs and use Audacity to listen to the audio to inspect the difference.
I ought not to sit for 30 boring minutes doing what a computer can do in seconds.
I have many applications that do bit-wise analysis of files, but none of them provide an audio output to let me hear what is going on at the point of difference.
WaxCylider >> Perhaps easier to use the “Solo” buttons to play just one of the tracks at any one time
Hi WaxCylinder. As noted in my reply to Doug, I do not want to listen to the tracks. I want to listen to the tracks only after a difference has been found.
“Is there a simple way to have Audacity compare two tracks bit-wise, and then halt at the first mis-match?”
Note to all: I am not comparing two close-to-similar tracks here; nor am I comparing two tracks recorded at different times by the same person. I am comparing one track with a clone of itself, the clone having (we believe) only four areas where brief (total of 30-second) changes were made.
What happens after the first change is detected? I deal with it, then select a chunk of audio immediately after the change, and ask Audacity to locate forwards that bit-pattern in the other track. By this means the starting-point for the second comparison is established. And so on.
I am comparing one track with a clone of itself
Was the machine that made the clone the same machine that made the original FLAC? We are warned that digital audio files are created with reference to the computer’s internal time system (that little clock chip down there on the motherboard).
I don’t need, nor do I want, to listen to the first 30 minutes of what is identical audio.
Nor do you have to. If nobody comes up with an automatic way to do it, I would use the Half and Half Again method. As above mount both tracks and listen briefly to the beginning and then 30 minutes in. It should be possible to hear a distinct timing error at 30 minutes. Then go half-way into the selection and listen again.
If half-way has no errors, go half-way again.
I think you’re going to find a time error that gets slowly worse over the course of the performance due to computer hardware timebase errors. I don’t think there’s going to be one single “Aha!” moment you can fix.
Hi. I am not sure this could help you but you can try:
- Create a new project and import both tracks
- Invert only one of the tracks (Effect > Invert)
- Then mix them both into a new track (Tracks > Mix > Mix and Render to New Track)
If both tracks are exactly the same they will be mixed into silence until the first difference is found.