Mp3 file duration and conversion to wav

Hello,
I have a short mp3 file:
http://52.28.125.94/audio/input-mp3/part0.mp3
When loaded in Audacity on Windows it shows duration around 0.078 seconds, but if looked at in foobar2000 or MediaInfo the duration is about 0.104 seconds (4608 samples at 44100Hz).
Who is right and why please ?
Also, when converting to wav using something like Ffmpeg we get 0.104 seconds with the first 0.078 seconds identical to the Audacity’s output and the rest being some garbage.
Thank you.

The file that you attached is about 1 minute 39 seconds long, but it has damage of some kind that is causing only the first 0.078 seconds to open in Audacity.

I loaded the file into [u]MP3 Diags[/u] and I got several “strange” errors…

“No MPEG audio stream found.”
“Invalid MPEG Stream. Stream has fewer than 10 Frames.”
“Broken stream found.”
“Unknown stream found.”

Based on the reported bit rate of 96 kbps that various tools agree on (or any sane bit rate) it could not be longer than a second or less given the file size?

Gale

Guys, thanks for your feedback, the sample is one of the series:
http://52.28.125.94/audio/input-mp3/part1.mp3
http://52.28.125.94/audio/input-mp3/part2.mp3
http://52.28.125.94/audio/input-mp3/part3.mp3
etc and these are produced by Ffmpeg like so:
require(‘child_process’).spawn(“/usr/bin/ffmpeg”, [
“-i”, URL,
“-strict”, ‘experimental’,
“-vn”,
“-b:a”, “96k”,
“-acodec”, “libmp3lame”,
“-ar”, “44100”,
“-reservoir”, “0”,
“-f”, “mp3”,
“-ac”, “2”,
“http://”+httpServerAddr+“:”+httpServerPort,
], {detached: false});
Do you suggest that each separate chunk of mp3 stream cannot be decoded ?
Could you point me at what is missing in the chunks ?

Also, if chunk size is increased MP3 Diags does not complain anymore, but audacity still seems to cut end of the file - 0.810 vs 0.836 seconds.

I’m not sure exactly how you are making these files, but certainly that first file was malformed. There are several forums specifically for ffmpeg, and they are probably better able to help than we are.

Can you explain what exactly you are trying to do and what the source audio is that you are encoding to MP3? For example, is this an audio stream that you are trying to encode as 1 second fragments?

Why do you need such short fragments?

What is the version number of FFmpeg that you are using and what configuration was it built with? You can just type “ffmpeg” at the command-line and post the output to let us see that. What operating environment is FFmpeg running in - cygwin?

I believe the problem Audacity has with these files is that there is no LAME header. Foobar 2000 does not truncate the files.

But FFmpeg 2.2.2 does not write files without LAME header, in a normal Windows environment.


Gale