Saving an audacity file larger than 4gb?

I recorded a cassette (about 90 minutes long) at 96000Hz. Attempting to save it gives me this error:

“You have attempted to Export a WAV of AIFF file which would be greater than 4GB. Audacity cannot do this, the Export was abandoned.”

Is there a workaround? I suppose I can split the track into two.

I read that you can export in FLAC. If I do this, are there any disadvantages when opening it up for editing again as opposed to saving it as an AUP file?

The 4GB limit is the result of the 32-bit file-size field in the WAV file header. FLAC is lossless. The files are smaller and there is no inherent size limit. AUP projects are also lossless.

Ok, so say I recorded something, saved it as FLAC or AUP, and am now going to reopen it for editing.

There’s no difference between opening it as a FLAC vs an AUP?

There’s no difference between opening it as a FLAC vs an AUP?

There’s a big difference. There is no AUP sound file. You are saving an Audacity Project which is a collection of files and folders.

A Project will save your individual tracks and settings from an edit and allow you to “freeze” a process in the middle of a show. It will not, however, save UNDO and it is not a good format for archive storage. Projects save work as collections of six-second sound files in that folder and there can be thousands of them. If the computer drops the ball anywhere, you lose the AUP manager file or damage any of the tiny sound files, the show will be trash.

As near as I can tell, the only limitation to FLAC is support. It’s probably not a good idea to send a FLAC to somebody else without first checking if they can open one. WAV files are as universal as you can get.

I recorded a cassette (about 90 minutes long) at 96000Hz.

What was the goal for doing that? You’re also using 24-bit, right? “24/96” is a studio format. You might do that if you’re planning on complex editing and re-mastering for distribution in the open market. Was the original cassette Dolby C and do you have a Dolby C player? I don’t remember, does Dolby C have that calibration tone at the beginning like B does, and did you do the calibration?

44100, 16-bit is the format of Audio CD. That’s usually a good goal for home work including mastering for audiobook reading. If you don’t do that you can have what you got. Awkward, odd sound files that everybody may not be able to play.

And yes, splitting it into A side and B side may reduce the sizes enough to get you out of trouble.


There is one other layer to this. You didn’t post how you would be listening to the work. If you have a Personal Music Player or music system in the car, it’s perfectly acceptable to copy the WAV files to MP3 at high bit rate (quality) settings such as 192 or higher Constant Bit Rate. The MP3 files will be indistinguishable from the originals and much more convenient to portable storage.

The trick is you can’t edit them. The instant you try to edit or change an MP3, the quality goes down. MP3 is an End Format. Make an MP3 and Full Stop. Enjoy listening to it. If you have to make changes, go back to your WAVes, make the change and make a new high quality MP3.


You will, of course, be unable to resist the urge to break the show up into individual songs (or acts in the case of an opera). Do that at the WAV level and do it to a copy of the WAV transfer. We publish tutorials for that.


Is there any harm for me to save it at 96000Hz? I don’t care about distributing it to people at this moment, it’s just for archiving purposes.

These cassettes are recordings of me as a child so I want the recordings to be as pristine as possible since the cassettes will likely degrade very soon.

Is there any harm for me to save it at 96000Hz?

The only downside is larger files. (Except there have been “reports” of worse performance from “pushing” certain soundcards to their limits.) As Koz says, 24/96 is the pro studio standard, so if you were running a pro studio you’d use that by default.

But, but there’s no practical reason to do that.

The guys who do blind ABX testing have pretty-well demonstrated that nobody can reliably hear the difference between “high resolution” audio and a copy downsampled to “CD quality”. And CD quality digital is obviously far-far beyond cassette, especially a homemade voice recording on cassette.

Even good-quality MP3 can sometimes sound identical to a high-resolution original in (and again it beats cassette) but MP3 is lossy compression so if you go back and edit it, it get’s decompressed and re-compressed again and the “damage” does accumulate with each generation of compression (similar to making multiple generations of analog copies).

One up side of high bitrate masters is the ability to downsample them to other formats with little or no damage. Certainly no audible damage. So you can always do that even you have a system that can’t actually play them.