Originally Posted by twobob
This is excellent advice for someone looking to get this going now.
I personally am interested in having a little more control of the code functionality than 3rd party app calls as this will allow us to potentially expand on the building blocks when they are completed.
But excellent solution for the actual final goal to be realised at this time, well commented and the cross platform research was a nice touch.
I'll have to try it out. Many thanks
I had success playing audio and video with two instances of ffmpeg on my Kindles: one outputs audio to STDOUT piped to aplay, and the other outputs video to STDOUT piped to gmplay. It can even play streaming 1080p x.264 video on a K5 (rescaled to 800x600 in realtime), but on a K3 there is only enough CPU power to decode lower-resolution XviD (or use an extra non-realtime step with a temp file instead of a pipe)
What I would really like is to play audio and video with a single instance of a single program. Ehhh's video player is almost there, but his video framerate is much slower than gmplay (until I get around to making a "gmplay" gst plugin for him), and his audio has too many dropouts.
I was hoping to get alsa sound working in my ffmpeg before publishing it. For now it is just an ARM build with all options turned on (except alsa). One nice thing about ffmpeg is that it can RECORD too, including recording the framebuffer. I was thinking of using it to make youtube videos of my demos, but perhaps they would be more effective filmed from a real camcorder that shows that it is really a kindle playing the video effects...