Can I add a compressed audio track to an Mjpeg?

Question

Is there a case where a video file could contain both mjpeg frames and a sound layer? I know originally, people used to place a 8khz PCM uncompressed track along with their mjpeg movie since it is streamed/decoded/played frame by frame with no motion prediction needed. Can some decoder accept an Mjpeg with a more recent audio format?

[EDIT 1]

What I'll first try is to check if ffmpeg handles the conversion of Audio/Video movies to MJpeg with audio, and I'll explore the header and the layers with an hex editor.

[EDIT 2]

OK. I've studied a Mjpeg with audio: ffmpeg -i some_movie_with_music.mp4 -f avi -acodec mp3 -vcodec mjpeg mjpegWithSound.avi 

And there's an MP3 file splitted into the total number of frames under each jpeg plus some changes in the header. So it's easy to implement in a context where a mobile application would offer to the user the opportunity to add an MP3 files to a serie of jpeg or to a movie. So, one more reason to use Mjpeg when a platform has no encoder yet.

Answer 1

It's fun to watch your application take shape. :-) I'm going to assume this is a follow-on to your last question and that you want to write C# code to accomplish this task. Are you still writing this into an AVI container? AVI stands for "audio/video interleaved" and is designed to transport both audio and video.

So, yes, you should be able to write both MJPEG and audio into an AVI file.

Guess what! You have lots of options for audio codecs too. We haven't cataloged quite as many audio codecs as video codecs (but close). Good news, though: Implementing a basic audio encoder in pure C# should be much simpler than trying to port even an MPEG-1 video encoder. Alternatively, check around to see if you can find an MP3 encoder written in pure C#. AVI accommodates MP3. If not, try IMA ADPCM. It's easy to implement and gives you 4:1 compression. Thus, if you have a monophonic, 44100 Hz, 16-bit stream, that requires 88200 bytes/sec. IMA ADPCM will give you roughly 22050 bytes/sec (plus small overhead).