I got 4K video playback to work using Haiku with mpv. Just scale it down from “full screen” mode.
This is efficient on a 3Ghz quad-core 64-bit setup. Not efficient on my dual-core 32-bit laptop yet.
The OpenGL part is in not doing too much (aka full screen effects) or certain ‘live rendering’ effects to overload things.
One thing about certain ‘killer apps’ is exclusivity - i.e. they are usually not marketed as cross-platform (i.e. multi-OS) applications upon initial release (basically, marketed as exclusively developed on/for Haiku). Like Amiga promoted EA’s DeluxePaint which wasn’t available for other computers.until a few years - or like Newtek’s Video Toaster for Amiga or Apple’s FInal Cut Pro. An app that specifically utilizes Haiku’s strengths (and/or weaknesses) to help market the benefits of Haiku to other people.
Let me just say that I applaud any and all efforts to get a video editor running on Haiku. Editing video is a difficult endeavor with multiple unique challenges. Let me give a brief description on UltraDV, a native BeOS editor effort ported to Haiku. UltraDV was originally intended as a commercial product to be release under BeOS V4 or so. The code was open sourced and archived and is freely available. It features an interface with 8 tracks for audio, video, and stills. The interface code is mostly complete and is simple and reasonably intuitive. There is an interface for transition effects, but transitions themselves still need to be coded. The code itself is close to rendering output but not quite completed. The challenge is that the rendering was written for V4.0 Media Server functions which no longer exist. UltraDV can be easily extended to support any resolution of video.
Please, stop making me useless ports 
I believe we need to involve @Barrett, who was officially “working” on video editing under Haiku:
https://www.haiku-os.org/tags/video-editing
Latest screenshot from my Haiku native video editor. I now have A/B markers. Audio scrubbing plays live preview. Still have heaps to do, but slowly getting there. Expect to release by the time R1 is final.
Barrett been working on the media kit even recently… that work should complement Zenja’s contributions as well.
Not really, he’s working on a bunch of refactoring that isn’t affecting functionality. All the long-standing bugs are still there, untouched…
Well, it affects it in that it becomes actually functional… hopefully.
NLEs don’t use GPU acceleration at all outside of virtual reality. even high-end encoders don’t touch the GPU.
the media kit is already the backbone of a much more robust video suite than any other open source editor project out (and most commercial ones)
This is … blatantly false. Where do you get your information?
It’s the exact reason the Radeon SSG even exists… https://www.amd.com/en/products/professional-graphics/radeon-pro-ssg
Also also why RED has their Rocket accelerator cards (though the SSG is a good bit faster than a rocket accelerator and more versatile).
(Message deleted by moderation team, personal insults.)
I’m not an engineer, but Quick Sync Video was specifically introduced by Intel (from their Sandy Bridge CPUs) for video encoding and decoding and is exploited by many modern NLEs:
Indeed it can be used for draft quality output, decoding previews as well as for streaming things live, in the end you typically aways use a CPU encoder for the final one though.
the NLEs i use in order to feed myself, running on systems with no acceleration.
“GPU acceleration”
those exist, those are not part of most edit suites.
