Missing/Deleted AUP file

I have inherited a bunch of Audacity recordings from an organisation and they’ve asked me to convert them all to MP3 files - no problem. The majority have been easy, just load the project and export to MP3.

However there are a significant number of projects for which I have the data folders but no .AUP file corresponding to it. Is there any easy way/tool to rebuild the project from simply the data files?

Things to consider:

  • There are literally dozens of projects in this condition and most of the recordings are 30+ minutes long with 600+ .au files in multiple sub folders so before you suggest it, stitching them together manually is not an option.
  • I’m pretty sure I could put together the XML myself but again, way too time consuming and not a viable option.
  • All projects are raw recordings - none have been edited.
  • All projects are recorded in stereo.
  • I have looked at the old Audacity Recovery Utility but it didn’t work, I’m not surprised as it is quite old I understand.
  • I’m using V.3.0.0 of Audacity.
  • I’m on Windows 10

I’m an IT professional with a good grasp of most technical aspects of computing but I’m fairly new to Audacity.

I’d be really grateful any help you guys might be able to provide.

That’s a good sign.
Do the .AU files still have their original time stamps?

There is no way to determine from the file names what order the AU files need to go in.

All except two files will be about 1.1 MB, which corresponds to one audio channel with about 6 seconds of audio data, assuming the default sample rate of 44100 Hz. For a mono track, the final AU (“blockfile”) may be smaller. For a stereo track there may be two smaller AU files at the end (one per channel).

To reassemble the track, you need to sort the AU files into chronological order from their time stamps.
To do that you will need to write a script to batch process the files. In Python I think you can use os.stat().
From the analysis, you could either generate the XML for an AUP file, or handle the file concatenation in your script. The former is probably easier as Audacity will handle file headers correctly for you.

I don’t know if the time stamp resolution on Windows is sufficiently precise to determine which is the left channel and which is the right. If not, there may be blocks in the left channel that belong in the right, and vice versa, but you could mix down the track to mono, which will still give a correct average for each pair of 6 second blocks.

Thanks very much for the reply. I feared that might be the case. I guess I was hoping for an easy solution to the problem, like a new feature from Audacity that I’d missed.

I see that all the filenames are hexadecimal, these are not incremental I’m guessing?

I’ll investigate the possibility of checking the date/time stamp in a windows environment but they still exist on the original iMac that recorded them so worst case I’ll go back to that.

From researching this it seems like it is a pretty common occurrence - do you think that audacity might develop a solution in the future? Also do you think it might be worth creating a commercial solution?

Thanks again

No they’re not. They are pseudo-random, which is a fast way for Audacity to generate unique names for each new .AU file without complex book keeping.

No and no, because Audacity 3.x introduces a new project format in which the entire project is one single file. The bucket full of separate .AU files is no longer used.