We’re growing the Active Theory team and hiring four roles that will help shape the next chapter of our work:
• Executive Creative Director:
https://t.co/WMSmuo0WXE
• Associate Creative Director:
https://t.co/7vCuUhV87M
• Senior Designer:
https://t.co/YLuQJZ8xbw
• Associate Technical Director:
https://t.co/FE3F28Gfq9
All roles are full-time and fully remote. We collaborate primarily across EST to PST time zones, with a preference for candidates based in the Americas.
If you care deeply about craft, systems, and building work that pushes the web forward, then please apply.
@joeydi@Anemolito@active_theory By using keyframe=1 you are not leveraging intra frame compression. Which means the file weight is similar to a series of a single jpg per frame
It’s our 20th birthday, but the gifts are for you 💚🥳
Join Party of the Year(s) on Spotify to rediscover your all-time listening history.
Reply with your results below ⬇️
Another little something we squeezed out of @claudeai & @figmaweave. This time playing around with the activeframe library from @active_theory. Demo link: https://t.co/pz7kYpsOVJ
@goellner@active_theory It could, but to get better browser support it's just easier to have two videos and composite them in build. With ActiveFrame you can keep in sync multiple videos, like in the demo
For an upcoming project at @active_theory I've been experimenting with a file format i'm calling ActiveFrame.
Like a video format, leveraging H.264 compression, allows random-access to any frame and it's fully hardware accelerated. No runtime depedencies, works eveywhere.
@vanilagy@active_theory Thanks for creating Mediabunny. I think that online demo is a bit misleading – I mostly made it for other kind of purposes: Biggest need was to sync multiple videos and use them in a webgl context
And leveraging intraframe compression to get smaller files
https://t.co/NjV3LMDg7T
For an upcoming project at @active_theory I've been experimenting with a file format i'm calling ActiveFrame.
Like a video format, leveraging H.264 compression, allows random-access to any frame and it's fully hardware accelerated. No runtime depedencies, works eveywhere.
Many people asked, so – I just published ActiveFrame open source on GitHub.
It's a stripped version of what we've been using so far at @active_theory – but has all the main ingredients. Still lot of work to do
https://t.co/rPF8u5ezX7
https://t.co/EOjVpSCXUy
@jo_chemla@active_theory@vanilagy Not sure about max video duration, do not depend on VRAM, just regular cpu bounded memory (and the limitation set by the browser)
My use case was
- small files, <2mb
- fast as possible, no runtime demuxing or other dependencies
- mostly for webgl effects, syncing multiple videos
@Anemolito@active_theory I think that might open the way for some memory leaks, better to just have your texture upload sync (at least that’s how I like it). I don’t know much threejs, but you can do that easily with webgl directly 🙌
@PMazhuga@active_theory Firefox on Windows? I haven’t tested it, but I think either a smaller GOP (more keyframes) on the video, or set the decoder to ‘prefer-software’ would help. Will test if I have some time in the next few days
Many people asked, so – I just published ActiveFrame open source on GitHub.
It's a stripped version of what we've been using so far at @active_theory – but has all the main ingredients. Still lot of work to do
https://t.co/rPF8u5ezX7
https://t.co/EOjVpSCXUy
@bandinopla@active_theory Gotcha, in the demo I’m using h265 on desktop. Likely not supported on Linux by default. The demo needs a better fallback strategy, thanks for the heads up
@Anemolito@active_theory You have precise control on what frame to show. Regular video tag is unreliable for that. You’ll also know when the frame is rendered - you can sync multiple videos together, something that you can’t do with regular <video>