reply to post by Violater1
I will try to explain it as clear as possible.
When a video codec compresses a video stream, it looks for areas that doesn’t change. When it finds an area that doesn’t change, then it doesn’t
send a new copy of that area out every new frame. When it detects a change in that area, then it sends the updated information on down the video
stream.
Now here is the problem. You usually have a hard limit on the maximum data rate that the codec is allowed to send down the stream.
Lets say that you have a video that has a lot of motion in it. So much that to keep up with all the changes in the image, it would have to send data
faster than it is allowed to.
There is two options…..
#1 Cut frames where there is less frames per second, but keep the image quality above a specified percent on each frame.
#2 keep the frame rate up, but allow a lower image quality on every frame. As more motion happens, you allow the smaller defects in the image to go
unfixed.
It depends on the manufacture of the compression device, but most of the go for option 2.
The frame rate of the video was maintained, but there was blocks of the video that the codec couldn’t fill in because it didn’t have any more
bandwidth available to send the data.
For a hands on example. Download a copy of virtualdub, or any other video compression program that allows you to manually set the limits for the
codec.
www.virtualdub.org...
Set the slider control on the video codec to smoothest. Set the bandwidth limit to some ridiculous low value. And try to compress a video with a lot
of motion in it.
You will see moving object disappear and faces turn into jumbled blocks.
I have seen it so many times when I have been compressing video to fit in a limited space/bandwidth that it doesn’t even take a second look. I know
what is happening.