Pushing the buffer via gst_pad_push () in transform_ip () function
causes downstream elements to process the buffer with a reference
count > 1. This leads to performance issue if there are downstream
elements which modify the buffer memory.
However, in drop-only mode this reference is not required.
So, let GstBaseTransform push the buffer in drop-only mode.
Fixes#4258
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/9532>
When using the custom WebKitMediaSrc element (used by WebKit and able to
perform an initial seek in playbin), a stall caused by streamsynchronizer
was detected during an initial seek. The flow of events revealed that the
intertwining of the initial configuration of the streams with the reset
caused by the flush events from the seek left streamsynchronizer in an
inconsistent state:
streamsynchronizer0:sink_0 (video) events, starting before the seek:
stream-start --> Sets the stream to wait
flush-stop --> Clears the stream wait flag
caps
tag
segment
stream-collection
(buffers start to come and flow properly)
streamsynchronizer0:sink_1 (audio) events, happening after seek:
(no flush events, because the stream hadn't been initialized when the seek happened)
stream-start --> Sets the stream to wait
caps
segment
(stalled because the stream is in wait mode!)
The code in streamsynchronizer expects that all the streams are in wait
state before releasing all of them at once. The flush on the video stream
broke that assumption and that's why the audio stream is never released in
that scenario.
Avoiding the clearing of the wait flag on flush-stop isn't an actual solution
to the problem, as it creates other side effects and at least makes the
gst-editing-services/seek_with_stop test to timeout. The alternate solution
implemented in this patch consists on analyzing if the other streams different
from the one newly added (after the flush) aren't waiting (which would mean
that they've all been unlocked after all of them were waiting before), and,
in that case, mark the new stream as also not waiting.
A new test_stream_start_wait test case has been added to demonstrate this
problem. The test case creates a video stream, pushes a buffer, then
simulates a seek by pushing flush-start, flush-stop, stream-start and segment
events. Note that the flush-stop clears the video stream waiting flag.
After that, a new audio stream is created and stream-start and new segment
events are sent. Note that stream-start will set the audio stream to wait.
Then a buffer is pushed on each stream. In the failing case, the test hangs.
In the working case (after this fix), the test runs properly because the
fact of having seen a stream-start also helps to clear the wait flag.
A second new test_stream_start_wait_sparse test has also been added to prove
that this mechanism can also work with sparse streams (a special case of the
current stream-start handling code). This test behaves like the previous one,
but there's no video buffer after the seek (it'll come in the future, as the
stream is sparse, but actually never comes). The buffer after the seek in the
audio stream starts at its due time. Streamsynchronizer is able to ignore
the wait for the video stream and produce audio buffers on time.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4544>
The 'new-pref' property sets the preference to use the new (next)
instead of the old (previous) buffer. The default is set to 0.5 to get
a similar behaviour as before the change.
Value 0.0 makes sure that only frames are shown where it's known that
the frame content is visible at that time, always show the old frame
until the new frame timestamp is reached.
Then, if the next buffer replaces the previous buffer the new buffer
is pushed as often as possible until PTS is reached. Before the new
buffer was only pushed once the new next buffer arrived.
Use GstClockTimeDiff because it's known that the current buffer time
is inside the time interval of previous buffer and next buffer the
calculation can be done with building absolute values. Special macros
are not needed here.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/8579>
Variable f1 is never used, so just skip that loop for now.
The test has never actually tested actual resampling because of
that bug it seems, and the test fails if fixed to actually resample.
For now we just avoid the pointless 126*12 pipelines that were just
testing the same thing (nothing) over and over again.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/7827>
It was adding and subtracting the segment base here and there, but it
was also doing so incorrectly, leading to various calculation errors.
Fixed a few bugs uncovered, related to getting a new segment:
* If we reset base_ts/next_ts/out_frame_count, also reset prevbuf
* Only do so if the new segment is different than the previous one
Also replaced a few occurrences of GST_BUFFER_TIMESTAMP with
GST_BUFFER_PTS for consistency.
Integrated the tests of
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2186
, now passing. The test_segment_update_same test had to be fixed,
because it was wrongly assuming that we would not fill the gap inside
the new-but-same segment.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/6031>
Windows does not support fork() so all tests will run in a single
process, and global variables will be reused across multiple tests.
Thus, each test should reset global variables.
Also, setup pad chain/event functions before playing state
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5752>
Fixes a potential GPU stall if an immediately freed texture/buffer is
attempted to be reused immediately by the CPU, e.g. when uploading.
Problematic scenario is this:
1. element does GPU processing reading from texture
2. frees the buffer back to the pool
3. pool acquire returns the just released buffer
4. GPU processing then has to wait for the previous GPU operation to
complete causing a stall
If there was a reliable way to know whether a buffer had been finished
with across all GPU drivers, we would use it. However as that does not
exist, this workaround is to keep the released buffer unusable until the
next released buffer.
This is the same approach as is used in the qml (Qt5) elements.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5144>
These 10bit formats are identical to NV12_16L32S, but 64bytes of data is being
prefixed with 16bytes data with four pixels of lower 2bits per byte. For
MT2110T, the lower two bits set so each bytes contains a column of 4 pixels,
also describe as tiled lower 2 bits. MT2110T has been chosen as a name to match
the vendor chosen name. This format is unlikely to exist for other vendors.
For MT2110R, the 2 low bits are in raster order.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3444>
The current limit is `x10`, which allows just `+20 dB` of gain.
While it may seem sufficient, this came up as a problem
in a real-world, non-specially-engineered situation,
in strawberry's EBU R 128 loudness normalization.
(https://github.com/strawberrymusicplayer/strawberry/pull/1216)
There is an audio track (that was not intentionally engineered that way),
that has integrated loudness of `-38 LUFS`,
and if we want to normalize it's loudness to e.g. `-16 LUFS`,
which is a very reasonable thing to do,
we need to apply gain of `+22 dB`,
which is larger than `+20 dB`, and we fail...
I think it should allow at least `+96 dB` of gain,
and therefore should be at `10^(96/20) ~= 63096`.
But, i don't see why we need to put any specific restriction
on that parameter in the first place, other than the fact
that the fixed-point multiplication scheme does not support volume
larger than 15x-ish.
So let's just implement a floating-point fall-back path
that does not involve fixed-point multiplication
and lift the restriction altogether?
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5063>
Don't assume that compositor will output only single buffer
for single input buffer. If buffer's running time is not completly
aligned to output buffer running time or duration, compositor
can generate multiple buffers. If that happens, two threads,
one is aggregator output thread and main thread were trying
to modify buffer in this test. Clear the buffer after
shutting down pipeline to avoid the race.
Fixes: https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/2836
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5081>
If gst_buffer_pool_set_config() fails then the pool will use its old
config. This may include different width or height when
pic_width/pic_height != frame_width/frame_height.
As a result, the assertions in theora_handle_image() will fail.
So check the result of gst_buffer_pool_set_config() and only use the pool
if it succeeds. Otherwise let the parrent decide_allocation() create a new
pool.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4600>
The first serialized events that can be send on a src pad are a CAPS and then a
SEGMENT event.
When handling events from user in appsrc, we used to send a segment
automatically if the SEGMENT has not been sent yet.
This breaks if the CAPS event was not send either as we were now sending
a SEGMENT before the CAPS.
Fix this by delaying such events until the CAPS has been configured.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4297>
Adding propose_allocation is to meet the requirement of Application to
request buffers. Application sometimes need to create buffer pool
and request buffers to maintain buffer management itself, and Gstreamer plugin
import Application's buffers to use. So, add propose_allocation in
appsink like waylandsink and kmssink etc.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4185>
