data loss when re-editing 16bit audio

I’m recording audio from LPs. The Audacity docs suggest archiving audio using “32-bit Float” wav. I did that. I ultimately want my recordings to be stored as flac as I want the ability to edit the audio files at anytime without audio loss while using minimum disc space. Besides the fact that the 32-bit wav files and resulting flac files are double in size, I also noticed that after converting these flac files back to wav, the new wav files load into the editor much slower than the original wav files do. After some research online, I found that the flac format does not support 32-bit Float and thus assume that those flac files must “padded” due to their large file size. I also assume that using 32-bit resolution is a waste and might even cause audio loss going back and forth between the two formats. I don’t need 32-bit audio, 16-bit is acceptable for all of my audio files.

The question: Is there any loss of audio data when importing and exporting 16-bit audio multiple times, being that the Audacity editor natively works in 32-bit Float mode?

I’m new to Audacity and currently use good ol’ Cool Edit 2000. I’ve been editing 16-bit wav in Cool Edit without fear but now I wonder.

Thanks in advance.

They suggest recording in 32-bit float, and exporting as 16-bit. If they suggest exporting as 32-bit, please give the exact URL and we’ll correct it.

If you want to load WAV files more quickly, you can always choose the “Read directly” option when importing the file: Audacity Manual.

If you insist on recording in 16-bit you can change it in Audacity’s Quality Preferences.

Be aware that Audacity always processes internally in 32-bit float. So if you choose 16-bit quality you will get downconversion loss if you run filter effects on the audio. Audacity will upconvert to 32-bit to apply the effect, then downconvert to 16-bit to return the audio to the track. Any effect which changes the sample values will have that result.

All those internal downconversions will have dither applied, which adds some intentional very low level noise to mask the rounding errors from the downconversion. You can turn the dither off in “High-quality Conversion” in the Quality Preferences if you prefer - but you may just possibly hear harmonic distortion instead of the dither noise.

There will be rounding loss. But less loss than if you record in 16-bit and apply lots of effects.

Bear in mind that Windows Vista and later performs upconversion to 32-bit float before Audacity gets the audio. That could be another reason to stick to 32-bit float in Audacity. You can’t avoid those Windows conversions unless you use the alpha builds of Audacity and record with Windows WASAPI (in Exclusive Mode) or Windows WDM-KS.


Gale