Default sample format ist set to 16-bit
Imported mp3 is still being imported as 32-bit, which then can be converted to 16-bit via context menu, but shouldn’t it be imported as 16-bit right when importing when default sample rate is set to 16-bit?
I’d like to save space and time when editing these non-audiophile recordings and work with 16-bit files without having to convert them every time one file is being imported.
I believe that Audacity does all its editing functions in 32-bit float arithmetic.
Audacity works internally at 32-bit floating point. You can export as 16 or 24 bit, or as MP3.
Digital signal processing is easier in floating-point. Plus, if you boost bass (or something) you can go over 0dB without clipping and everything will be OK as long as you lower the volume before exporting.
If you open a 16 or 24-bit file the conversion to floating-point and back is lossless.
Note that if you export as MP3 you are going through another generation of lossy compression and there is some accumulated “damage”. You may not notice any quality loss but it’s something to be aware of.
Yeah, but I set the default sample format to 16-bit in settings and still wonder why the imported files are not imported as 16-bit files…
I’m not sure what the defaults do… When I once did an experiment with 16-bit default it was still “acting like” floating point and able to go over 0dB. I’m pretty sure all of the effects are done in floating point no matter what. That’s standard with audio editors/DAWs.
Except the sample rate will take effect while recording and if your soundcard is feeding it a different sample rate, it will be re-sampled to the default.
If you open a file, the default sample rate will be ignored and you’ll get the sample rate of the file. If you have a file already open and you import a 2nd track with a different sample rate, the 2nd one will be re-sampled. (Maybe that happens every time you import instead of open.)
It’s not a problem except that it takes more temporary space, and more space if you create an Audacity project, and you might have to remember the original format if you want to keep the same format.
Hey guys, i’ve read the manual here: Audio Settings Preferences - Audacity Manual
And it says “This sets the Sample Format (bit depth) which will be used each time Audacity is launched”
Add i set it to 16-bit. Because the files are still imported as 32-bit, I considered this a “bug”, because the behavior should be as stated above (in the manual) to import the files as 16-bit, isn’t it?
I’d say the manual isn’t clear. Or maybe the manual is just “wrong”.
It shouldn’t be a problem (except for maybe the larger temporary and project files) because you have to select the format when you export anyway.
I’m not a developer but it s not practical to write all of the effects/processing code to work at different bit-depths, and 3rd-party plug-ins wouldn’t work.
And there’s a good chance of losing quality if you stick to integers. If you simply lower the volume you lose resolution which isn’t a big deal if you leave it at lower volume but if you re-amplify the resolution doesn’t come back.
What makes you think the file has been imported as 32-bit?
Which exact version of Audacity are you using?
If you right click in the track settings area you can call a context menu and in format the check mark is next to 32-bit floating.
When I change this to 16-bit, Audacity starts to calculate and change the bit depth.
When saved, the project file is much smaller.
That’s what I would like to have when importing, and even the manual says so. 
The version of Audacity is 3.7.5
If you create a new track in an empty project, do you get the expected 16-bit format?
Yes! The checkbox next to 16-bit is checked in context menu > Format.
With that 16-bit track in that project open:
That MIGHT be because the MP3 decoder decodes to floating point and for whatever reason Audacity doesn’t re-convert it.
(MP3 doesn’t have a regular bit depth because it doesn’t store individual samples.)
I raised this same question many years ago.
The answer is that Audacity is correctly setting the default project format, and the default project format is used when adding new tracks, generating audio, mixing down, and exporting (which involves mixing down).
The default sample format setting does not apply to importing - When an audio file is imported, Audacity just imports it. Audacity does not automatically import and convert the format - it just imports. If you want to import and convert the sample format, you have to explicitly convert the sample format, for example by rendering the track (Mix and Render).
Note that the importers are not all guaranteed to behave in exactly the same way - it depends on whatever the codec does.
Maybe if you told us a bit more about the jobs. Plain, ordinary audio editing can be dangerous without the 32-bit talents.
For example, Audiobook Mastering software automatically reduces rumble, sets the overall loudness of the voice whether the tips and peaks overload or not. Then it gently reduces tips and peaks to conform to audiobook standards with no damage. This process wouldn’t work in plain 16-bit.
Koz
I’d simply want to process a recording (file), reduce it to mono, normalize it and save the project with the name of the recorded file and save the project in 16-bit to save space on disk and then to be able to quickly open a project if needed to show it to someone and be able to quickly change the project to show different recordings, it is necessary to see the waveform everytime, and therefore the best is to save the project.
Because there are many recordings (files) to process that way I’m looking to reduce the steps involved to a minimum that I have to do manually one by one at the moment.
I think you don’t need a AUP project at all. (I rarely make one). You can Export emphasized text (rather than Save) as WAV, or whatever format you want.
Thanks for your thoughts on my task. Exporting emphazied text, would mean exporting by generating labels, isn’t it? I know of this feature and it is a great feature that i used already to export clips, but with this task, i don’t want to change the recording other than changing to mono, reducing bit-depth and normalize it, but keep the full length and being able to see the full or zooming in/out of the waveform. It helps to analyze occurencies in the recording to be able to recognize the time the peaks occur.
This topic was automatically closed after 30 days. New replies are no longer allowed.