To ensure better stability of future releases, we need to adopt multiple stages in the release cycle. As we already label Alpha, Beta, Candidate and Stable differently, simply adopting this classification system already does everything for us. This also allows us to maintain compatibility with the existing system, while offering something new entirely.
- Moved all auto-dependencies to a uniform subdirectory for easier caching and cleanup.
- Add an option to download or specify a path for libOBS+obs-frontend-api.
- Remove the dependency on the non-standard obs-frontend-apiConfig.cmake file.
- Add an option to download or specify a path for OBS Dependencies.
- Add an option to download or specify a path for Qt.
- Add an option to download or specify a path for AOM.
- Fix and improve architecture and platform detection.
- Fix some messages having two :, or no prefix at all.
- Fix detection of obs-frontend-api.
- Fix applying custom compiler and linker flags for MSVC and GNU-style builds.
- Use target_compile_options over add_compile_options for compatibility.
It was sometimes impossible to sign the binaries on Windows, depending on how CMake was being run. By explicitly converting to the intended native format, we avoid this problem for most situations. Additionally we now support per-project certificates, instead of global certificates only. This also allows multiple uses of cmake-codesign in the same overall project.
Originally intended to be an experiment with no future, it turned out to be very popular with streamers that move a lot. In the end it was popular enough that NVIDIA added their own variant to their Broadcast software, which works decently enough. Unfortunately my wrapper code around the library was written very poorly, so it didn't take long for it to break out of nowhere.
- Updated the AOM library to libAOM v3.1.2-882-03b6f69.
- Fixed the Denoising filter not working correctly with unexpected sizes.
- Fixed the Denoising filter not rendering at the correct size.
- Fixed the Denoising and Upscaling filter destroying the Alpha channel.
- Fixed the Denoising filter running twice on the same image, resulting in a far worse image.
- Fixed Upscaling not automatically deciding on a proper scale factor for you when the selected one is not supported.
There are a number of duplicate shader routines we should combine into a single shader to save disk space, and remove unexpected errors in one copy but not the other.
* The "Super Resolution" filter is now called "Upscaling" and in the future will support additional providers, like FidelityFX "Super-Resolution" and NVIDIA "Upscaling".
* The entire plugin now has identical logging behavior, so it should be easy to identify exact parts of the plugin in log files.
* A crash caused by log functions on Linux/Mac has been fixed, which was caused by reusing variable length arguments. (#632)
* Radeghast submitted a new 'Swirl' example shader. (#618)
* Added an AV1 encoder based on AOM-AV1, although its stability and quality may not be great so far.
* Denoising of Video content is now possible with the new "Denoising" filter on NVIDIA RTX hardware.
* Fixed a few bugs affecting the "Upscaling" (previously "Super Resolution") filter.
* Fixed references to NVIDIA missing the ®.
The previous name was too strict on what could be put into the effect, and would result in additional clutter in the Filter menu when we eventually decide to support other Upscaling methods than Super-Resolution networks.
* Fixed a bug preventing versions from being parsed correctly in CMake.
* Fixed a bug introduced between 0.9.3 and 0.10.0b1 which broke transparency support for Color Grading.
* Added support for direct rendering to some sources/filters/transitions.
* Added .pkg based Installer for MacOS, thanks to @cpyarger.
* Updated translations from Crowdin.
Enables users of the Mac Operating System to use an automated installer instead of the manual installation method which is prone to user errors. This is done via the use of "Packages", which generates a .pkg file according to the requirements we present it.
When the CMake script was rewritten, not much attention was put into the versioning code, resulting in odd behavior which was never found or fixed. For example, the automatic splitting of the suffix from the number never worked, and the build number was being stored in the wrong variable.
* 3D Transform should now appear again.
* Automatic disabling of unsupported features should now work in CMake.
* Fixed a compilation error on some versions of GCC.
- #437 Use 'git describe --tags' for versioning.
- #439 Add support for overriding the automatically detected version.
- #454 Fixed new versioning code breaking when the tag is on the current commit.
- #427, #428 Fix Qt interaction before OBS Studio is actually ready.
- #452 Fixed building without updater but with UI.
- #447 Redesigned Color Grading to support two rendering modes:
- Direct Rendering applies the entire color grading function to every single pixel, and is thus more accurate, but much slower.
- #-Bit LUT Rendering first applies the color grading function to a LUT that varies in accuracy, and then uses that LUT to render the actual output instead, which is much faster but less accurate.
- #447 Optimized Direct Rendering in Color Grading to be up to 25% faster, while using 50% less VRAM.
- #440 Removed support for Ubuntu 18.04 as it ships seriously outdated versions of libraries we use.
There is hardly any reason for us to recalculate everything all the time. LUTs can cache the work once, and then re-use it every time necessary, drastically reducing the impact of Color Grading by almost 60% (on some GPUs even more). Additionally this fixes the negative gamma issue, which plagued the filter for a while.
In the future, once PR 4199 (https://github.com/obsproject/obs-studio/pull/4199) has been merged, we can cut away one intermediate rendering step currently required to make the effect work. Hopefully this will be with the 27.x release of OBS Studio.
For simple image and video editing, LUTs (Look-Up Tables) are vastly superior to running the entire editing operation on each pixel - especially if all the processing can be done inside a single shader.
Due to the post-processing requirements for our LUTs, we are limited to 8 bits per channel - though clever use of the unused Alpha channel may result in additional space. For our purposes however, this is definitely enough.
A complete redesign of the component and dependency system is necessary in order to support additional platforms, such as MacOS and other Linux platforms. Additionally it results in a much cleaner code base, which is less confusing overall.
Eventually it might be necessary to push components of StreamFX into their own CMake projects, as it is getting kind of complex now. Especially with the push for a proper plugin manager, things get dicey for big plugins like StreamFX.
Adds support for the AMD Advanced Media Framework H.264 and H.265 encoders via FFmpeg. The majority of settings are supported, and the UI/UX experience mimics that of the NVENC implementation. Various settings are left out due to their complexity and should be controlled via the custom parameters field.
The MODULE_EXPORT and EXPORT macros in libOBS do not correctly mark a function or type as visible on GCC, which results in the newly added flag hiding everything from view, instead of just what should be hidden.
Michael Fabian 'Xaymar' Dirks
carsten.braun
Christopher P Yarger