Commit graph

52 commits

Author SHA1 Message Date
Sebastian Dröge
4ad101b53b Use once_cell crate directly again
The glib crate does not depend on it anymore and also does not re-export
it anymore.

Also switch some usages of OnceCell to OnceLock from std.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1441>
2024-01-31 18:07:57 +02:00
Sebastian Dröge
1ef47cb48e Fix a few new clippy 1.75 warnings
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1415>
2023-12-22 16:40:18 +02:00
Sebastian Dröge
763739e3ae Update for Buffer / Memory API changes
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1409>
2023-12-17 14:07:12 +02:00
Sebastian Dröge
16b917abb1 Update for gst::Rank API changes 2023-11-02 14:10:59 +02:00
Stéphane Cerveau
68c2d27e8d fmp4mux: specify the fragment duration unit
The fragment duration is expressed in nanoseconds.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1348>
2023-10-04 12:47:15 +02:00
Sebastian Dröge
3661b4f95b fmp4mux: Fix draining in chunk mode if keyframes are too late
We would create another chunk that ends after the fragment end, and
would from then on consider the stream always filled for the chunk
because it starts after the current fragment end (i.e. nothing would go
into this fragment).

This is obviously wrong because the actual fragment end moved further
ahead because of the additional chunk.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1277>
2023-07-18 07:57:20 +00:00
Bilal Elmoussaoui
dd2d7d9215 Use re-exported once_cell
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1268>
2023-07-06 17:50:49 +03:00
François Laignel
f85106b86a mp4, fmp4: fix byte order for opus extension
The "Encapsulation of Opus in ISO Base Media File Format" [1] specifications,
§ 4.3.2 Opus Specific Box, indicates that data must be stored as big-endian.

In `write_dops`, `to_le_bytes` variants were used.

Related to [2].

[1] https://opus-codec.org/docs/opus_in_isobmff.html#4.3.2
[2] https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4875

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1247>
2023-06-19 16:32:07 +02:00
Sebastian Dröge
2e83107c18 fmp4mux: Don't wait for more data if a stream has no GOP starting before fragment end
Simply don't output anything for this stream and only include it in the
future.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1229>
2023-06-01 19:46:06 +03:00
Sebastian Dröge
a5fcd66c95 fmp4mux: Consider a stream filled if the earliest GOP starts after the current chunk
There's not going to be any buffer to output for this stream in the
current chunk.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1229>
2023-06-01 19:25:44 +03:00
Arun Raghavan
b05c21680d Revert "fmp4: Return a running time in get_next_time()"
This reverts commit 04bb7b4db0.

As Sebastian points out, the chunk PTS is already in running time, so
this was wrong from the start.

Fixes: https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/issues/363
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1219>
2023-05-23 09:27:00 -04:00
François Laignel
7ba0073052 use Pad builders for optional name definition
Also, apply auto-naming in the following cases

* When building from a non wildcard-named template, the name of the template is
  automatically assigned to the Pad. User can override with a specific name by
  calling `name()` on the `PadBuilder`.
* When building with a target and no name was provided via the above, the
  GhostPad is named after the target.

See https://gitlab.freedesktop.org/gstreamer/gstreamer-rs/-/issues/448
Auto-naming discussion: https://gitlab.freedesktop.org/gstreamer/gstreamer-rs/-/merge_requests/1255#note_1891181

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1197>
2023-05-12 12:55:31 +02:00
Antonio Kevo
cf21bfabf2 fmp4: Use updated start_pts when checking stream filled
After calculating the earliest pts, the fragment_start_pts and
chunk_start_pts in State are updated. However, when checking if the
stream is filled, the previous start_pts (set to None) is used instead.
This means that chunk_filled and fragment_filled will be false the first
time aggregate() is called, assuming timeout is false, all_eos is false,
and the sinkpad is not EOS. This requires aggregate() having to be
called a second time before the first fragment is sent.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1207>
2023-05-10 16:09:25 +02:00
Arun Raghavan
04bb7b4db0 fmp4: Return a running time in get_next_time()
We were currently returning a value based on the next chunk PTS, but the
expectation in GstAggregator is that we return a running time. This
resulted in spurious wakeups and warnings like:

0:00:01.501685123 1552995 0x55899715c1e0 WARN                 fmp4mux mux/fmp4/src/fmp4mux/imp.rs:1818:gstfmp4::fmp4mux:👿:FMP4Mux::drain_buffers:<fmp4mux0:sink_1> Don't have a complete GOP for the first stream on timeout in a live pipeline

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1127>
2023-03-10 10:58:38 -05:00
Seungha Yang
6b15e772ac fmp4mux: Ignore framerate update
like mp4mux in -good does already

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1094>
2023-02-16 02:23:56 +09:00
Sebastian Dröge
5965ff4364 fmp4mux: Accept more data on already filled streams if the remaining streams need more data for finishing a GOP
In other words, continue queueing buffers in sync from all streams until
all of them are ready for draining instead of stopping to queue buffers
on every stream that is already filled individually.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/issues/310

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1081>
2023-02-09 20:36:42 +02:00
Sebastian Dröge
3a408c0146 fmp4mux: Handle GOPs ending after the desired fragment end correctly
Either create further chunks if enough data is queued or simply start
the new fragment at a later time if the keyframe is later.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1077>
2023-02-06 19:20:34 +02:00
Sebastian Dröge
4c3ae6f8ce fmp4: Add support for AV1
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1075>
2023-02-04 13:48:37 +00:00
Sebastian Dröge
cef6fef079 fmp4: Add support for VP8
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1075>
2023-02-04 13:48:37 +00:00
Sebastian Dröge
ecb26a0b16 fmp4mux: Fix a couple of assertions by handling these cases cleaner
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1071>
2023-02-01 14:57:00 +02:00
Sebastian Dröge
1998ecab45 fmp4mux: Refactor and clean up code
Split many longer functions into multiple functions and simplify various
parts. Overall this is functionally still the same as before.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1068>
2023-01-30 14:41:33 +00:00
Sebastian Dröge
2a3d962dc5 fmp4mux: Add support for sub-fragments / chunking
Allow outputting sub-fragments (chunks in CMAF terms) that are shorter
than the fragment duration and don't usually start on a keyframe. By
this the buffering requirements of the element is reduced to one chunk
duration, as is the latency.

This is used for formats like low-latency / LL-HLS and DASH.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1059>
2023-01-27 19:28:27 +00:00
Guillaume Desmottes
abe4efc4a2 fmp4mux: add 'offset-to-zero' property
Add it only to 'isofmp4mux', the onvif variant already does this and
CMAF and DASH are always single-stream so you rely on inter-container
synchronization via the running-time.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1063>
2023-01-25 12:29:30 +00:00
Sebastian Dröge
ee8249eec7 fmp4mux: Don't write the first sample flags into any trun but the first
The flags are based on the first sample of this fragment so writing it
into any trun but the first is not very useful.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1060>
2023-01-24 19:54:12 +02:00
Sebastian Dröge
1ceaea844a fmp4mux: Fix decision whether per-sample flags are needed in the trun
Previously it would never use per-sample flags if any later sample
needed different flags than the first two.

Also comment the code a bit better.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1060>
2023-01-24 19:54:11 +02:00
Sebastian Dröge
458b2386ed Update for glib API changes 2023-01-21 18:13:48 +02:00
Sebastian Dröge
6132788b02 Update for caps/structure-related string API changes
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1048>
2023-01-15 22:58:44 +02:00
Sebastian Dröge
e9bbf804ba fmp4mux: Remove obsolete comment
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1037>
2023-01-06 14:52:56 +02:00
Sebastian Dröge
af9d9c0a5c fmp4: Fix compilation after glib::List API changes 2023-01-02 19:22:51 +02:00
Johan Bjäreholt
71c268da14 fmp4mux: Only push fragment_offset if write_mfra is true
This is done so that the fragment_offset vector does not infinitely
build up when write_mfra is disabled.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1023>
2022-12-22 10:31:27 +01:00
Sebastian Dröge
13b6f8fad4 fmp4mux: Skip gap buffers earlier to consider them for the sample durations and fragment start durations
Otherwise dropping the gap buffers would offset the timestamps of
following samples.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1016>
2022-12-16 17:43:44 +02:00
Sebastian Dröge
3f904553ea Fix various new clippy warnings
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/1011>
2022-12-13 11:43:16 +02:00
Jordan Petridis
a84eeeb240 mux/{mp4, fmp4}: Hard depend on feature v1_18
Else --no-default-features was failing to compile.

v1_18 is needed to for the aggregator code.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/588>
2022-11-29 21:06:12 +02:00
Sebastian Dröge
6b3bb2c747 fmp4mux: Handle EOS correctly if it happens before a fragment start time was determined
Whatever earliest time we have at that point is going to be the start
time.

Also handle the case correctly where all inputs are EOS before any
buffers were received at all.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/issues/270

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/982>
2022-11-19 19:24:43 +02:00
Sebastian Dröge
e3f645af19 fmp4mux: Don't overflow negative composition offset calculation
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/982>
2022-11-19 19:24:04 +02:00
Sebastian Dröge
fd910cb828 fmp4mux: For video with N/1001 framerates use N as timescale
See https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3049

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/981>
2022-11-18 16:07:11 +02:00
Sebastian Dröge
c553ac7402 fmp4mux: Re-work buffer dequeueing and calculations of timestamps
Especially simplify calculation of ONVIF UTC times. As a side-effect
this reduces the number of times the running times of a buffer are
calculated, and also causes streams to be interleaved correctly in ONVIF
mode if there is a non-constant UTC-to-running-time difference.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/979>
2022-11-18 15:00:00 +02:00
Sebastian Dröge
f062b7cf0d fmp4mux: Make media/trak timescales configurable
And refactor a bit of code for easier extensibility.
2022-11-07 18:06:29 +00:00
Sebastian Dröge
7ac29827d2 fmp4mux: Don't allow VP9 for CMAF
This would require setting the correct compatible band for VP9 in CMAF,
which is not implemented yet.
2022-11-03 16:53:01 +02:00
Sebastian Dröge
6706f3a4b4 fmp4mux: Add initial Opus support
Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/issues/239
2022-11-03 16:53:01 +02:00
Sebastian Dröge
a8250abbf1 Fix various new clippy warnings 2022-11-01 10:27:48 +02:00
Sebastian Dröge
31d4ba2eae fmp4mux: For VP9, write resolution into the tkhd and include a stss box to signal that not all frames are sync samples
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/957>
2022-10-31 15:20:19 +02:00
Sebastian Dröge
938f4e4f1c fmp4mux: Clip negative PTS to zero/last PTS instead of erroring out
This can happen at the beginning of a stream if upstream is
rtpjitterbuffer and it has problems figuring out timestamps in the
beginning due to resetting / skew.
2022-10-27 14:47:16 +03:00
Sebastian Dröge
627bf756b1 fmp4mux: Send force-keyunit events for now if the ideal position has already passed 2022-10-27 12:53:09 +03:00
Sebastian Dröge
a56435801d fmp4mux: Add debug log when writing the mfra box 2022-10-26 17:22:20 +00:00
Sebastian Dröge
27fac33c44 fmp4mux: Reset timing infos to None if a stream only contained gap events for a whole fragment 2022-10-26 17:22:20 +00:00
Sebastian Dröge
cab4cd3b8c fmp4mux: If a stream is longer than the main stream at EOS, simply include all of its buffers in the last fragment nonetheless 2022-10-26 17:22:20 +00:00
Matthew Waters
d067fb2ec8 fmp4mux: don't require dts for predictive-only formats like vp9 2022-10-26 11:42:13 +00:00
Matthew Waters
8c8384c711 fmp4: add support for muxing VP9 streams in cmaf, dash and iso fmp4
As specified in https://www.webmproject.org/vp9/mp4/
2022-10-25 18:33:42 +11:00
Sebastian Dröge
9a68f6e221 Move from imp.instance() to imp.obj()
It's doing the same thing and is shorter.
2022-10-23 23:08:46 +03:00