Why do things like confetti ruin the birate of videos and livestreams?
In: 12
Most codecs only change the pixels of a video when there is motion of some kind, whether it’s a person talking or a parade with confetti. All the places in the video that stay the same, the codec only has to transmit the pixel once until there’s a chance detected. This saves a lot of data transfer, but when things like confetti are in the video that means a lot of change and a lot of data transfer. When there’s so much data transfer sometimes the codec it stream can’t keep up and the pixels don’t change when they’re supposed to so you see the video break up.
Videos are often compressed such that only occasionally is an entire frame defined (called a “key” frame) and then subsequent frames only include information about what changed since the previous frame. An example would be if a student was sending a paper to someone for proofreading, it would be much less writing to just say what the editor changed from the original paper instead of sending back the entire paper with the edits.
This works pretty well with most kinds of video because what changes on the screen between frames is relatively minor. But things like confetti are very difficult to compress because there is a lot changing all across the frame, and those changes are significant differences. The bandwidth necessary to define those changes dramatically increases and because so much changes in a relatively short period of time the “only tell me what changed” method doesn’t work very well.
A blank canvas can be compressed down to one thing for the computer to remember and output later. If a black ball moved across the white background, it could still remember most of the white from before, but some of it would have to change to black. Now do that thousands of times with dozens of different colors. It’s a lot of rapidly changing information all at once. The computer can’t rely on remembering shortcuts, so it has to put in a lot of hard work to actually output the video.
Tom Scott has a great video explaining and demonstrating this: https://youtu.be/r6Rp-uo6HmI
Videos aren’t served to you in their original format.
They’re run through what’s called a codec. What that does is it tries to find all the sneaky ways it can save space in typical videos.
In typical videos, things don’t move around and change color much, so that’s where they focus – on the stuff that does change.
However, they have limited space/second for storage. So they try to do the most they can in their budget.
The issue is that stuff like confetti moves fast and changes shape, color, speed, and direction unpredictably.
What ends up happening is that only some of the detail in the video can be retained while staying in budget. Lots of changes between frames means that the budget gets spread really thin, resulting in that glitchy and blurry mess that comes with videos of confetti.
Latest Answers