How Is A Variable Bit Rate Actually Produced?
In: Technology
Variable bit rates apply to lossy codecs for video and audio.
Lossy codecs compress video/audio by throwing away the parts that you’re least likely to notice (or can’t even notice) are missing. They work on some variant of decompose the audio/video into components and then only keep the data to describe the most significant components.
The bit rate defines how many components you can keep. In constant bit rate, every unit of time, there are a constant amount of components that must be kept. This means when the video/audio is very complex, it’s not described so well by the data kept and when the audio/video is very simple, excessive components are kept. Variable bit rate allows the algorithm to keep more components when there are lots of significant ones and fewer when there are fewer significant components. The algorithm will generally try to target a certain average bit rate.
As an example of components, imagine an audio snippet from some music, it is probably made from some complex wave shape. You can transform the signal and instead of representing it as some complex wave, you can define it as the sum of lots of simple sine waves with different frequencies. If you have enough of these simple sine waves, you can construct the original signal perfectly by simply adding them together. Now you take these sine waves of different frequency, and you see which ones are basically insignificant relative to the others. If one wave is sound level 10000, and another is sound level 3 – there’s no way your ears can discern the second one, so there’s no point storing that information, you just throw it away.
Latest Answers