I’m trying to batch process some files with a simple macro which normalizes each file and then exports as WAV. I have verified that all my default audio settings are 44.1k and 32 bit float and no matter what I do, the files in macro-export always end up as 16 bit. There seems to be no way of getting anything other than that.
Try setting encoding to 32-bit in the Export dialogue window …
I understand the desire to use the tools you are already familiar with, but if you are willing to explore something made for batch processing, investigate looping ffmpeg with the -af loudnorm parameter.
Yeah I had tried that, it doesn’t make a difference. Single manual exports do obey that setting, but then using the macro reverts back to 16 bit.
Export scripting is under-developed. Audacity 4 will have a new system for these things, where you simply can use the actions from the UI. In the meantime, I’d also suggest using ffmpeg directly to batch convert stuff.
Thanks, basically what I’m doing for now is exporting to FLAC, which works fine, and then just using a separate app to convert those to whatever WAV format I want. It’s annoying that there’s an extra step but not terrible.
I learned recently that saving to a FLAC file will not save 32-bit float data. It will be 24-bit signed data. To save audio in 32-bit float you will need to save it as a RF64 or W64 file.
Oh, interesting, I did not know that! Thanks for the tip.
