For this release the corresponding GstVideoCodecFrame before
pushing the buffer. The buffer will now be writable unless
the subclass still holds another reference to the buffer or
the frame.
When we get a new buffer, always call the parse function, even if it is a 0
sized buffer. For theora we need to also decode 0 sized buffers.
Ideally we would like to make theoradec be packetized but that fails currently
because of oggdemux and because of the assumptions that the base class makes.
This allows elements to specify a function to upload
a buffer content to a specific OpenGL texture ID. It
could be used by the vaapi elements to provide a way
for eglglessink or WebKit to upload a VA surface to
an GL texture without the respective sinks knowing
anything about VA.
Helps when using dvbsuboverlay in connection with vaapisink
or some other video sink that wants ARGB pixels (dvbsuboverlay
attaches pixels in AYUV format, and we then convert as needed).
Alignment should not be a problem here.
DTS and PTS usually have a non-zero offset between them in MPEG-TS,
so assigning DTS to PTS is almost always wrong. The other, newer
timestamp recovery code does it correctly if we leave it as invalid.
For interlaced vertically subsampled images we need to combine alternating
chroma lines with alternating luma lines. That is line 0 and 2 are combined
with the first line of chroma samples and line 1 and 3 with the second line
of chroma samples.
See also: https://bugzilla.gnome.org/show_bug.cgi?id=588535
The _1_0 suffixed environment variables override the
non-suffixed ones, so if we're in an environment that
sets the _1_0 suffixed ones, such as jhbuild, we need
to set those to make sure ours actually always get
used.
And only return the proportion. The earliest time already can be
retrieved from get_max_decode_time() and by renaming we allow this
to be more extensible in the future.
Add a getter for the QoS proportion and earliest_time to help
subclasses do better estimations based on the proportion.
API: gst_video_decoder_get_qos_info()
https://bugzilla.gnome.org/show_bug.cgi?id=687991
We only allocate 8 bits per component for our temp buffers, which
causes invalid memory accesses if we try to unpack formats that
unpack into a format with 16 bits per component such as e.g. v210.
We don't support blending onto those yet, so just bail out.
This reverts commit e39fbe6b7e.
Looks like we need to pass the full .la file after all in a setup
with libtool, or it might not find the library, e.g. like
ERROR: can't resolve libraries to shared libraries: gstfft-1.0
Conflicts:
gst-libs/gst/audio/Makefile.am
gst-libs/gst/pbutils/Makefile.am
Also see https://bugzilla.gnome.org/show_bug.cgi?id=603710
Monitor for reordered output timestamps, and then avoid oldest DTS
as PTS approach, and try for an oldest PTS as out PTS approach,
if at least all valid PTS available.
Avoids bogus estimating upon sparse available input PTS, and tries
to handle all-keyframe input, or input PTS which are actually DTS.
Hold both the stream and the object lock to modify the output_state,
this way it can be safely modified while hold either one or the other.
Also, only hold the object lock in the query
https://bugzilla.gnome.org/show_bug.cgi?id=684832
... by having some more timestamp tracking in a private frame field.
Not doing so would lead to (a.o.) losing the needed minimum timestamp in
an earlier sent frame.
... rather than to output segment, which will only be set
to current input segment if some output is produced
(coming from non-clipped input).
Also fixup debug message.
Don't try to take STREAM_LOCK on upstream events such as QOS.
Protect qos-related variables with object lock instead. Fixes
possible deadlock when shutting down in certain situations.
https://bugzilla.gnome.org/show_bug.cgi?id=684658
Drain out the decoder when encountering a gap. Needed for DVD 'still'
sequences which consist of a single video frame, and a large gap
while audio plays.
Only hold back events until the first buffer is generated, then just
send them directly. Otherwise, important events like 'still-frame' are
held forever, waiting for a frame that'll never arrive.
Expose the gst_video_info_align function that adds padding and does stride
alignment on a GstVideoInfo. Move this previously private function to
video-info.co
Move the definition of the alignment structure to video.h where it can be picked
up by both the bufferpool and the video-info.
Sometimes the encoder would need to use the allocator for something else than
just allocating output buffers, for example, querying different parameters.
This patch expose a getter accessor for the negotiated memory allocator.
Sometimes the decoder would need to use the pool or the allocator for
something else than just allocating output buffers. For example, the querying
for different parameters, such as asking for a bigger number of buffers to
allocate in the pool.
This patch expose a two getters accessors: one for the buffer pool and the
other for the memory allocator.
Unifies the code and ensures that:
* subclasses needing to use the frame_number on a void* field will
always work
* wraparounds will be automatically taken care of if we have to deal
with more than 2**32 frames
Fix copy'n'paste bug which made us allocate a slice of the
size of a rectangle for the overlay composition, but then
free it passing the size of an overlay composition, which
is not something GSlice takes to kindly, resulting in scary
aborts like:
***MEMORY-ERROR***: GSlice: assertion failed: sinfo->n_allocated > 0
Also, g_slice_new already includes a cast, so remove our
own casts, without which the compiler would probably have
told us about this ages ago.
https://bugzilla.gnome.org/show_bug.cgi?id=680091
Add a method to get the offset and scale values to transform the color values of
a format to their normalized [0.0 .. 1.0] range. This is usually required as
the first step of a colorspace conversion.
Add an unpack option to specify what to do with the least significant bits of
the destination when the source format has less bits than the destination. By
default we will now copy the most significant bits of the source into the least
significant bits of the destination so that the full color range is represented.
Add an option to leave the extra destination bits 0, which may be faster and
could be compensated for in the element algorithm.
The x/y values are meant to be signed.
This bug was introduced by 76c0881549
Conflicts:
gst-libs/gst/video/video-blend.c
gst-libs/gst/video/video-blend.h
Check that we have a valid output_state before attempting to use it to calculate
the duration of a buffer. It is possible that we don't have a state yet, for
example when we are dropping the first buffers.
Make sure the frame deadline was set before calculating the
max_decode_time. Fixes problems with ffmpeg skipping frames when
it doesn't need to, when the input doesn't have full timestamping
(divx in avi)
Interpolating the timestamps from the picture numbers
does more harm than good, getting it wrong in a lot of
cases (especially reverse playback). Removing it in favour
of simply incrementing the timestamps until there's
something better
Use g_list_free_full instead of walking lists twice when freeing
them.
Remove pointless clause in gst_video_decoder_chain that doesn't
actually have any effect.
Other changes to make the code slightly more like the 0.11
version.
Move processing of the gather list into the flush_parse function.
Add a last ditch attempt to apply timestamps to outgoing buffers
when walking backwards through decoded frames. Requires that each
gathered region has at least one timestamp.
Make sure to remove decoded packets from the decode list when
they are sent - otherwise the list just grows on each cycle, with
more and more frames being decoded and then clipped away.
Break out of the processing loop early on a bad flow return to make
seeking more responsive.
Use the gst_video_decoder_clip_and_push_buf function in reverse
mode, instead of pushing all buffers arbitrarily.
A couple of small efficiency gains in the list handling, by moving
list elements directly and not reallocating, and by reversing
and concatenating the gather list instead of moving it one node
at a time.
Rename the gst_video_decoder_do_finish_frame function to
gst_video_decoder_release_frame.
Rename gst_video_decoder_have_frame_2 to
gst_video_decoder_decode_frame and pass the frame to process
directly, rather than using the current_frame pointer as a holding
pen.
Move the negative rate handling out of the function to where it
is needed, and remove the process flag.
The frames are the owners of the buffers. In cases where a decoder
would keep around reference frames, we need to ensure they don't
disappear early.
To handle this, we pass downstream a complete sub-buffer of the output
buffer, ensuring that the buffer will only be released when downstream
is done with it *AND* the frame is no longer used.
Conflicts:
gst-libs/gst/video/gstvideodecoder.c
Don't replace the initial frame's timestamp with a bogus
one calculated from the (incorrect for Ogg) frame number just
because the 'sync time' hasn't changed.
Also, don't output a bogus warning about the output_frame being
NULL when it's being dropped/skipped due to QoS.
Use a separate variable to describe the amount of lines that will be used in
packing instead of abusing the h_sub variable. Some formats might have no
subsampling but need to operate on multipe lines.
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
Remove Y800 and Y16 wich are the same as GRAY8 and GRAY16_LE
Add const to the GstVideoFormatInfo when used in argument
Add GRAY8 and GRAY16 pack/unpack functions
Add support for supporting chroma subsampling correctly in the pack
function.
Fill in the pack and unpack functions for most formats.
Add some missing pack/unpack functions to the orc file.
Add a flag argument to the pack and unpack function so that we can expand it
later when needed. We could for example prefer a High Quality pack/unpack
operation later.
Add 10 bits I420 format definitions
Move encoded format as second entry in the array so that it doesn't end up in a
weird place when we add formats.
See https://bugzilla.gnome.org/show_bug.cgi?id=665034
When need to push out all the previously received events, concatenate all the
events from the previous frames (instead of leaking the old ones)
Improve debugging a little
Conflicts:
gst-libs/gst/video/gstvideodecoder.c
Frames receive a refcount when added to the frames list so release that refcount
in gst_video_decoder_do_finish_frame(). Also release the ref on the frame
because gst_video_decoder_do_finish_frame() takes ownership of the passed frame.
This allows subclasses to override it, as is necessary for e.g. the
video-crop meta. It is now necessary that after decide_allocation()
there is always a allocator and a configured buffer pool inside the
query.
Video base classes and theora plugin still needs to be ported again
Conflicts:
docs/libs/gst-plugins-base-libs-docs.sgml
docs/libs/gst-plugins-base-libs-sections.txt
docs/libs/gst-plugins-base-libs.types
ext/theora/gsttheoradec.c
ext/theora/gsttheoradec.h
ext/theora/gsttheoraenc.c
ext/theora/gsttheoraenc.h
gst-libs/gst/video/Makefile.am
gst-libs/gst/video/video.c
gst-libs/gst/video/video.h
gst/playback/gsturidecodebin.c
tests/check/libs/video.c
tests/check/pipelines/theoraenc.c
win32/common/libgstvideo.def
Some container formats (like AVI) set DTS on the buffers instead of
PTS.
We detect this by:
* detecting if input timestamps are non-increasing
* detecting if the order the frames come out is the same as the order
they were inputted (meaning the implementation is reordering frames).
If the decoder reorders frames, but input buffer timestamps were not
reordered, that means the buffers has DTS and not PTS as their timestamp.
If this is the case, we use set the PTS of the outgoing frames in the
same order as they were given to the decoder.
This fixes the issue for any decoder using this base class (yay).
Rename the frame_flags to flags. Because they are flags on the frame object it
does not need the redundant frame_ prefix.
Change the order of the metadata constructor so that the flags come before the
format and dimension arguments.
There's a new GstVideoFrameFlags enum now that contains the frame
specific flags only. GstVideoFlags does not contain the TFF/TFF/ONEFIELD
flags anymore because these are strictly frame specific.
Also add fallback to parse these fields from the GstBufferFlags in
gst_video_frame_map() if there's no GstVideoMeta attached to the buffer.
When setting its config, the pool increase the ref count of the allocator, but
at finalize the ref count is also increased rather than decreased.
This one-liner patch changes the gst_allocator_ref() for gst_allocator_unref()
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=674011
gst_buffer_set_qdata() will leak the structure passed to it
when called incorrectly (e.g. on a non-metadata-writable buffer).
This is expected, but we must avoid doing that in valgrind.
Try to avoid floating point maths for each pixel to be blended in
inner loop, and try to avoid the multiplication entirely for the
most common case of the global alpha being 1. Could probably be
refactored a bit more.
extract_alpha and apply_global alpha always return TRUE really,
so just do away with the return value. Convert a g_return_if_fail()
into a g_assert(), since this is only to check internal consistency
and not a guard for public API. Add some locking.
https://bugzilla.gnome.org/show_bug.cgi?id=668483
If we are asked to (un)premultiply,we need to create the new rectangle
with the right flags, so we can find it properly on subsequent cache
lookups (also because it's wrong otherwise).
https://bugzilla.gnome.org/show_bug.cgi?id=668483
We need to copy the pixels before messing with them, not least
because the buffer creation code below assumes it's ok to take
ownership.
Fixes crash caused by double-free.
https://bugzilla.gnome.org/show_bug.cgi?id=668483
We require the videometa activated before we can implement the alignment of
buffers. Users of the bufferpool should do this manually based on the results of
the allocation query.
