DCA encoder fixes
Here's proof of the concept patch to the dcaenc.c which fixes DCA encoding. It is not very clean and should be splitted into parts. Spotted and fixed/workarounded bugs: * avctx->frame_size was unset with filling temp buffer instead * small errors in bistream writing * no check was done on written data - outputting 0xFFFF in bitstream may be treated as syncword so I clear last bit of quantized value (so it can't be made from to consequent words as well). And I'd like to do AAC encoder anyway.
Kostya wrote:
Here's proof of the concept patch to the dcaenc.c which fixes DCA encoding. It is not very clean and should be splitted into parts.
Spotted and fixed/workarounded bugs: * avctx->frame_size was unset with filling temp buffer instead * small errors in bistream writing * no check was done on written data - outputting 0xFFFF in bitstream may be treated as syncword so I clear last bit of quantized value (so it can't be made from to consequent words as well).
And I'd like to do AAC encoder anyway.
Does it encode recognizable sounds now ? MvH Benjamin Larsson
On Sun, Mar 23, 2008 at 07:57:10PM +0100, Benjamin Larsson wrote:
Kostya wrote:
Here's proof of the concept patch to the dcaenc.c which fixes DCA encoding. It is not very clean and should be splitted into parts.
Spotted and fixed/workarounded bugs: * avctx->frame_size was unset with filling temp buffer instead * small errors in bistream writing * no check was done on written data - outputting 0xFFFF in bitstream may be treated as syncword so I clear last bit of quantized value (so it can't be made from to consequent words as well).
And I'd like to do AAC encoder anyway.
Does it encode recognizable sounds now ?
Not that I've tested it extensively but it does produce very recognizable sounds without artifacts to me.
MvH Benjamin Larsson
participants (2)
-
Benjamin Larsson -
Kostya