Hi Folks,
[Windows 10, Audacity 2.1.2)
I’ve been trying to do more with bird sound recording for eBird.org, using my Samsung Galaxy S4 (Android phone) for recording (with RecForge II Pro app) and simple trimming of ends and normalizing with Audacity. One of the things I’m not fully grasping yet has to do with how important it is to match the sample rate and format (depth) from the point I do the recording, through import to Audacity, editing, then exporting a file that will be uploaded to eBird.
I have RecForge set up to record at 48 kHz and apparently it is a fixed 16-bit depth.
When importing/opening in Audacity, I discovered that it opens the file in whatever format the preferences are set at. They have been set at defaults of 44100 Hz and 32-bit float. This of course does not match what I have recorded. So, is it “resampling” my recording, converting it to a format different than the original? Should I set my Audacity preferences to 48 kHz and 16-bit depth?
Then, when I go to export, it exports it at the 44100 Hz and 32-bit float settings. Seems like it would be best to have Audacity set to the same specifications that RecForge is recording with, but I don’t know how much difference it makes. I would appreciate any input on whether or not this makes much difference in the quality, technically. Thank-you.
Matt Hunter
There’s nothing wrong with 48000 sampling. That’s what video uses.
As far as bit depth, Audacity does everything internally at 32-floating to make effects and filters easier (or even possible), If all you’re doing is trimming the two ends and out the door and you start with 16-bit (highly likely), you can leave everything where it is and turn off dither in preferences. Dither controls the conversion back down to 16-bit at the end. Assuming you started with 16-bit, the transition should be perfect.
Edit > Preferences > Quality > Conversion
The instant you change the sound in any way, volume, filters etc, you’re dead. You need to put the dither back and what few tiny errors there are come back.
Koz
You said you were normalizing. That’s changing the sample amplitudes and so you should leave Default Sample Format at 32-bit float, so that there is no downconversion of bit depth as a result of Normalize. If the track was 16-bit there would be downconversion after Normalize processed at 32-bit then returned 16-bit audio to the track.
Audacity exports WAV at the bit depth shown in the export dialogue. To ensure compatibility for all player applications, it’s best left at the default 16-bit PCM. If you make that decision then in my opinion you should not turn dither off because you want it to manage the downconversion from 32-bit project to 16-bit WAV.
Gale