Riemann Computing's Founder - Introduction to the .qdxv File Format for Videos

追加: 2026-05-14

[00:00:01]Hello guys. So, what is up? I am trying to do the video compression in the QDXV format that I created.

[00:00:09]Instead of this running sole Python like the last demo, because video is a bit ex- exhaustive, this does use Cygwin and it runs C++ and C code. Again, this is the compression without the full machine learning and without the machine learning and recursion in in place as well as any grid compute or anything like that. Bare-bones, this is extracting the frame and the mathematics being really the core. This provides a basic bench line example really without the grid compute and the recursive learning.

[00:00:46]And this is trying to do it on test synth, which is a single gradient file.

[00:00:52]So, because it's single gradient, it's doing it on top of.mp4, which is already compressed codecs. So, it doesn't do this from raw 12 to or raw to the actual QDXV format, but it's doing on this on top of mp4 with mp4, as you know, is already a compressed codecs format. So, the hardware specs you could see right here and we're trying to enable CPU and CPU acceleration.

[00:01:26]Let's get started. So, drag and drop test synth and um you could see I set the quality at 30%, which is not the best. Again, trying to get as close to a maxed out ratio as I can. We did try running this before. This is version 20.04. We did try running this on a 150 megabyte file in version 4 and it got down to 352 kilobytes in the same quality format, but unfortunately that file was corrupted. Again, this could be because we're making new codecs. There was something going on with the header and there's a lot when you're trying to create new codecs. So, getting rid of those issues and then trying to enable some sort of playback, this is the max compression ratios that we could really come up with bare-bones without again any of the exhaustive machine learning processes and grid compute that are in place.

[00:02:19]So, let's get started.

[00:02:22]And you could actually see the speed for 96.9 kilobytes compressed down to 1.2 kilobytes, 77.9x smaller.

[00:02:32]Again, test since that QDXC. So, Oh, wait. I forgot to press decompress.

[00:02:38]So, as you could see, QDXV, we decompress and you can open it in the player and you can actually see the gradient change. Now, it's not as great as the original, but again, this is single gradient and this is at 77x.

[00:02:55]This is on top of a compressed MP4 and this is extremely good results for what it's worth, especially bare-bones.

[00:03:04]I think in terms of pushing the limits of information theory, we actually could get pretty close. So, anyways, if you guys like this video, don't forget to like, comment, and subscribe and I'll see you guys next time.