Directly forwarding the input buffer to the next filter means that
drawtext no longer owns any references to it and thus shouldn't refer to
it in any way.

The buffer is passed on to the next filter, so we shouldn't keep any
pointers to it.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

OpenJPEG can decode in lower resolution or decode only a number
of enhancement layers.

It already skips any extra bytes at the end, and apparently there are some
samples in the wild with larger 'kuki' chunks.

Needs to be done separately for each element.

Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>

This will make multi-channel implementation simpler.
Based partially on a patch by Andrew D'Addesio <modchipv12@gmail.com>.

Use get_bits_long() in decode_scalar().
Use unsigned int for decoded value.

Avoids unneeded interleaving and allows for reusing the AVFrame output buffer
as the internal buffer for 24-bit and 32-bit sample size.

Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>

This will simplify the multi-channel implementation.

It does not need to be unsigned.

It is already unconditionally set to 0 prior to this, so we can modify it
only when needed.

x_modified is just unnecessary, and final_val can be removed by simplifying
the unsigned-to-signed conversion.

avoids some unnecessary arithmetic in certain situations

reduces the number of parameters to decode_scalar() and slightly simplifies
the code