* Move NotifyOnProgress processing to ErrorData
* added HighPass filter, LowPass filter, Audiogate and Silencedetection
* re-added corresponding AudioFilterOptions
* Update LICENSE
* Update ci.yml
* Fix argument
* Bump action versions
* Ignore Uri_Duration until root cause found
* Use action that can install specific ffmpeg version
* Remove ignore
* Revert "Remove ignore"
This reverts commit d85a4b81ab.
* Bump dependencies
* Use setup-dotnet@v2 since v3 seems to install .NET 7
* Init
* WIP
* Add Directory.Build.props
* Add SupportedOSPlatform attribute on tests using SDC
* Fix using temporarily
* Add IgnoreIf attribute to only run SDC tests on Windows
* Cleanup pipelines
* Cleanup
* Cleanup using directives
* More cleanup
* Simplify attribute
* Fix attribute
* Add missing test file
* Added blackdetect and blackframe arguments
* Added log levels
* Add missing using directive after rebase
* fix extension is not png lost path
* Apply fix to methods in new location
* Add Uri support for Frame Analysis,
skipped Stream support as this cannot support MP4's with moov atom in the end of the file, and input pipes do not support seek.
* Add select multiple streams
* Add other stream types to Channel (V,s,d,t)
* Add negative mapping to select stream (deselect)
* Update test
* Add pad video filter
* Update PipeHelpers.cs
* Fix GetPipePath() for MacOS
* Add SampleAspectRatio property to VideoStream
* Update year
* Always use Path.GetTempPath() on linux and macos
* Use FedericoCarboni/setup-ffmpeg@v2
* Include macos in ci matrix
* AddDeviceInput similar to AddFileInput and FromDeviceInput
* fixed hwaccel parameter not working in 5.0
* a hack to unconditionally kill ffmpeg when parent .NET process exits
* Remove PInvoke.Kernel32
* Remove unneeded cast
* Update test
* Added ability to retrieve bit depth from media streams for lossless encodings (#359)
* Added ability to retrieve bit depth from media streams for lossless encodings
* Shortened sample AIFF file used in tests
* Cleanup after splitting into two packages (#386)
* Move PosterWithAudio to FFMpegCore
* Reduce windows only tests
* Update Directory.Build.props
* Create .editorconfig
* More cleanup
* Enable implicit usings
* Remove unused method
* Apply dotnet format
* Fix unused variable in AudioGateArgument
* Fix boolean conditions in AudioGateArgument
* Merge boolean conditions into pattern
* Use target-typed new
* Add linting to CI
* Add CUDA to HardwareAccelerationDevice enum
* Increase timeout for Video_Join_Image_Sequence
* Adjust Video_Join_Image_Sequence timeout
* Fix expected seconds in Video_Join_Image_Sequence
* Increase timeout for Video_TranscodeToMemory due to macos agents
* fix: Switch source for rotation property from 'tags/rotate' to 'side_data_list/rotation' (incl. test case) (#388)
* Init (#389)
* build master branch after merge or push (#390)
* Update ci.yml
* Add codecov
* Remove codecov for now
* Add coverlet.collector and codecov
---------
Co-authored-by: keg247 <44041557+keg247@users.noreply.github.com>
Co-authored-by: Wilbert Bongers <msdnwilbert@muziekweb.nl>
Co-authored-by: Artemii Gazizianov <107502822+ArtemiiimetrA@users.noreply.github.com>
Co-authored-by: Thodoris Koskinopoulos <me@koskit.me>
Co-authored-by: 赵宁 <602726286@qq.com>
Co-authored-by: jeroenvanderschoot <jeroen.vanderschoot@kinetiq.tv>
Co-authored-by: Sky Z <qe201020335@sina.com>
Co-authored-by: Gleb Moskalenko <gleb.moskalenko.general@gmail.com>
Co-authored-by: ep1kt3t0s <86835785+ep1kt3t0s@users.noreply.github.com>
Co-authored-by: Victor Nova <lostfreeman@gmail.com>
Co-authored-by: Tom Bogle <tom_bogle@sil.org>
Co-authored-by: pklaes <10601494+pklaes@users.noreply.github.com>
- on large recordings (e.g. radio transmissions), ffprobe might return number of hours which is too large for TimeSpan.Parse (exception: The TimeSpan string '149:07:50.911750' could not be parsed because at least one of the numeric components is out of range or contains too many digits.)
- use regex groups to extract components (hours/minutes/seconds/millis) then parse/create new timespan from that
- NOTICE: this will discard microseconds provided by ffprobe, not sure if this is significant
- ffprobe has inconsitencies with how it represents millisecond component. Sometimes it may return just `82` for 820 milliseconds, so padding with 0s is required on the left. Likewise, sometimes it might return microseconds past milliseconds (first 3 significant figures); this is currently discarded
- Added InternalsVisibleTo to help with unit testing *just* the duration parsing function
