I’m narrating an audiobook using Audacity 3.7.1 on a MacBook Pro running macOS Sequoia.
My finished Audacity project looks and sounds normal. Before exporting, ACX Check gives these passing results:
Peak level: –3.10 dB
RMS level: –22.68 dB
Noise floor: –77.16 dB
However, after I export the chapter as a WAV file and import that WAV into a new Audacity project, the entire waveform appears as a solid red block and sounds distorted in Audacity.
The strange part is that the same exported WAV sounds normal when played in QuickTime Player. I also completely re-recorded the chapter, but the exact same problem occurred after exporting and reimporting the new WAV.
This has happened with multiple chapters. The original Audacity projects look normal; the problem appears only after exporting the WAV and importing it back into Audacity.
I’ve attached screenshots showing:
The original project with its normal waveform and passing ACX results.
The exported WAV after it was imported back into Audacity, showing the solid-red waveform.
Is this an export problem, an import/display problem, or a known Audacity issue on macOS? What information or settings should I check?
That’s a little hard to see. I like dark gray desktops. The top is the actual player and the bottom part is the analysis. Note that it’s a mono file at 44100 Hz sample rate, 16 bit.
You may be accidentally exporting in the wrong format, (wrong bit-depth ?). MediaInfo (runs in browser) will show the the specifications of an audio file.
You can also check the format of the original, unexported audio data by opening either the context menu on the track’s LHS controls area (the box containing the volume and pan sliders), or opening the three-dots menu at the upper right corner of that box.
The Format and Rate submenus will show you the encoding format (16-bit PCM, 24-bit PCM, or 32-bit float) and samplerate (44.1kHz, 48kHz, or any of more than a dozen other possibilities) for the recorded project data.
It may be that your recording device (since you said this is recorded narration) is feeding in audio in an esoteric format that doesn’t match how you’re exporting it. Normally I would expect Audacity to take care of that on export, just like I’d expect it to use the metadata of the exported .WAV file to import it correctly, but it seems like something is going wrong there.
But I can get a .WAV file to import as a solid block of noise by, say, exporting as a stereo, 48kHz, 32-bit float WAV, then using File > Import > Raw Data… to load that WAV back in as if it contains 1-channel, signed 8-bit, 16kHz data. It’ll load both much longer duration than the original, and containing nothing but a solid-rectangle waveform filled with noise. (Or solid red, with Show Clipping enabled.)
Like I said, normally I’d expect the encoded WAV file to contain a header that tells Audacity how to interpret it, so it would prevent me from messing up the import like that unless I forced the issue. Especially since it plays correctly in other programs, so they can clearly interpret the data correctly.
@Trebor’s suggestion of using Mediainfo to examine the .WAV file headers is a good one. You can also use soxi filename (or sox --i filename) to get the same information using Sox, if you have that installed.
