Make function pointers NULL when nothing needs to be done.
Pass target pixels to dither and matrix functions so that we can later make
them operate on the target buffer memory directly.
videoconvert.c: In function 'videoconvert_convert_new':
videoconvert.c:287:11: error: 'Kr' may be used uninitialized in this function
videoconvert.c:287:15: error: 'Kb' may be used uninitialized in this function
Fix the calculation of the offset and scale values for GRAY formats. We also
need to set the offset and base of the chroma values to match what the unpack
function creates.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=679612
RGB8_PALETTED -> RGB8P
Fix the definition of paletted formats, store the palette in the second
plane.
Make sure we copy the palette correctly in gst_video_frame_copy()
Don't do alignment on the palette in videopool
Flesh out the video filter base class. Make it parse the input and output caps
and turn them into GstVideoInfo. Map buffers as video frames and pass them to
the transform functions.
This allows us to also implement the propose and decide_allocation vmethods.
Implement the transform size method as well.
Update subclasses with the new improvements.
With the new video bufferpool we can now implement the propose_allocation
vmethod on some video filter elements so that we can also use video metadata and
bufferpools when not operating in passthrough mode.
This reverts commit bd539753eb.
Adding the supported metadata to the query does nothing at this stage. Proposing
allocation parameters and supported metadata for upstream should use the
propose_allocation vmethod.
The output size of a buffer does not depend on the input size but simply on the
caps of the output buffers. Don't let the base implementation deal with
unit_sizes, because input buffers might not be a multiple of that when they have
padding or non-default strides. instead, implement a transform size function
that simply calculate the natural size of an output buffer based on the caps.
