On 01/18/2011 08:58 PM, John Stebbins wrote: > I am attempting to update some code to work with the new float implementation > of the ac3 encoder. I made what I believe > are the necessary changes, but the audio produced is just a buzz. mplayer > generates these errors: > > [ac3 @ 0xd86060]exponent out-of-range > [ac3 @ 0xd86060]error decoding the audio block > [ac3 @ 0xd86060]delta bit allocation strategy reserved > [ac3 @ 0xd86060]error decoding the audio block > [ac3 @ 0xd86060]delta bit allocation strategy reserved > [ac3 @ 0xd86060]error decoding the audio block > [ac3 @ 0xd86060]delta bit allocation strategy reserved > ... > > I have generated a minimal test program to demonstrate the problem I am > having. This program works with the fixed point > implementation from ffmpeg r25689, but does not work with current git head. > There is a define to change from fixed to > float mode at the top of the code so I can switch easily between. The > program takes as input a raw stereo pcm 48khz > file and outputs a raw ac3 file. > > Could someone knowledgeable take a look and see if there is anything obvious > I am doing incorrectly?
The floating-point sample format is in the range -1.0 to 1.0. You need to scale the samples when you're converting from int16. -Justin _______________________________________________ libav-user mailing list [email protected] https://lists.mplayerhq.hu/mailman/listinfo/libav-user
