there isn't necessarily a fix. all compression algorithms expand the size of the data on most inputs in order to make certain inputs with sufficiently regular p

there isn't necessarily a fix. all compression algorithms expand the size of the data on most inputs in order to make certain inputs with sufficiently regular patterns a lot smaller. it just so happens that most of the images/video/audio/text/etc people are interested in is fairly regular. with a lossy compression algorithm that is required to hit a certain compression ratio to keep the bitrate under some threshold, the only way to be able to guarantee that it is hit on uncompressible input is to introduce regularity/remove randomness, and in the case of video the easiest and least visually distorting way to do so is to pixelate it. this particular video is basically random noise which is incompressible by definition, which is why the result is so bad. if it was 1080p going in and the output size is required to be reduced by 16x, reducing the resolution to 480x270 is about as good you can do.