Determining the uncompressed bit-rate of a video

415 views Asked by At

So recently, I experimented with encoding the default Windows 7 Wildlife sample video with the following handbrake settings: https://i.stack.imgur.com/GQhQv.png

I did this three times, with the only difference being the x264 Preset, which I set at Ultrafast, Medium and Placebo. Naturally, because the Ultrafast encoding was doing a much worse job of compressing the video down in a loss-less way, it had to reduce the video Bitrate before loss-less compression to comply to the set 1000kbps Bitrate of the final video. Medium and Placebo could afford to have a higher uncompressed Bitrate, and so look a lot better, as seen here: https://i.stack.imgur.com/YPmV8.png

The problem I'm facing, is that, from the MediaInfo details shown above, there doesn't seem to be a vast difference between the files at all, apart from reframes, and I doubt that they could have that much impact on the quality of a video (although feel free to prove me wrong).

So I'm wondering, as all the videos seem to have the same post-compression bitrate, how are you meant to determine the actual quality of a video without looking over it yourself? Is there a tool that allows you to view the uncompressed Bitrate?

1

There are 1 answers

0
szatmary On BEST ANSWER

Un compressed bitrate is width * heigh * bits per pixel. bbp is 12 for YUV420p, (the most common color format in video). All three final videos will have the same bitrate.

Bitrate is useless measure of quality. The best measure, is to watch the video (a subjective test). There are less accurate objective tests like PSNR and SSIM that can be automated.

The difference is in the speed setting is around how much time is spent in each phase of the encoder. For example, the motion compensation step of the encoding process. Basically the previous frames are searched for similar data to the current frame. If it is found, instead of encoding the pixels into the current frame, a reference is made to the pixeles in the previous frame. This means the current frame can be encoded using fewer bits. The more time you spend looking, the better of a match you will find. hence better quality at the same bitrate at the cost of CPU.