Audacity Forum

waxcylinder wrote: ↑

Tue Oct 29, 2019 8:25 am

For MP3 export you need LAME and not FFmpeg.

oops my bad

For MP3 export you need LAME and not FFmpeg.

LAME is now already bundled in the latest version of Audacity.

WC

... audio is imported using that version. Why might the different version of Audacity lead to a timing discrepancy? Might it be something about the FFmpeg build the two versions are using? Thanks in advance for any guidance! Best, Julia

whykai wrote: ↑

Mon Oct 28, 2019 10:42 am

Is it possible to use the ffmpeg mp3-decoder in Audacity instead of libmad ?

Yes. Use the File menu "Import" command, and in the file type filter, select "FFmpeg compatible files". Audacity will then try using FFmpeg first when importing the selected file(s).

... in the mp3-definitions, it would be a good practice in modern decoders to provide provisions to fix such cases in the most convenient way. ffmpeg obviously already does this. It would be much appreciated, if the other decoders would also employ it. Depending on the length of the bitreservoir ...

I am currently using macOS Catalina Version 10.15, which I updated to today. Prior to that I was using Audacity version 2.1.1.0. I had the FFmpeg library before the update and was able to export audio to AMR (it was also included in the File Type dropdown whenever I'd export audio). Now it ...

liafail wrote: ↑

Wed Oct 23, 2019 9:42 pm

I opened an AMR file, took some time to open, but then played the file with no problem.

Does that file still work in Audacity? Can you still import it into Audacity?

I have FFmpeg properly loded in the Library; all three editions show up. I opened an AMR file, took some time to open, but then played the file with no problem. Onto file 2, and every file thereafter: upon opening or importing, ...

Well, uhh thanks! I did do "uncompressed files" and then that RF64 or whatever. I actually tried encoding it in Matroska (.mka) in signed 16-bit PCM with FFmpeg, and it completely worked.

... peak levels beyond 0 dB (~ +2.5 dB). Just a single frame of 1152 samples seems to be affected. For comparison I have decoded the mp3 also with ffmpeg 4.0.2 to wav. That version shows just a clean 'well behaved' (low level) transient at the same file position. I assume that this "glitch" ...