The compression algorithm of an Mp3.

In addition to the physiological structural properties of the human ear, the function of the brain also plays a very important role.

The pitch in the sound is determined by the fundamental tone, while the timbre is determined by the harmonics, and the human brain will automatically complete the fundamental tone, even if the fundamental tone does not exist. For example, the bandwidth of a telephone is only 300~3200 Hz, but when we listen to a man with a base tone of 120 Hz talking on the telephone, we can still hear his correct tone and will not confuse a boy with a girl. . .

We still don’t know how the brain uses complex calculations to reconstruct this non-existent tone.

PS Add a little visual easter egg, can you see what’s weird about this image?

(Please read the answer to the end)

…………………………………………………………………………………………………………………… ………… ……… …………………………………………………………………………………………………………………… ………………………………………………………………………………………………………………………………… ………………………………………………………………………………………………………………………………… ……………………………………………………………………………………………………………………………… ……………………………………………………………………………………………………………………………… ……………………………
_ Your vision~ amazing! The human body still has too many unknown magic eggs waiting to be excavated~~

The compression algorithm of an Mp3.

The birth of the MP3 compression algorithm is nonsense of human organs in the digital age. The whole algorithm is not improved around the math, but rather optimized around how to fool the human hearing organ.

So this algorithm is very curious, Baidu finally found information after a long time, and has a little understanding of the principle of it, so please record it.

basic principle
There is a special effect of shading effect on the human hearing model.
The role of the cochlea is as a spectrum analyzer, converting sound waves into signals of different frequencies. The villous cells at each specific location will be stimulated by a specific frequency, but when the basilar membrane leads to fluctuations, the villous cells around it will also be stimulated. That is, if there is a frequency with a high volume, and at the same time there is a relatively weak frequency near it, the sound of the relatively weak frequency will be covered by the relatively loud sound, and our human ears have no way to distinguish the sound There is another sound of a weaker frequency.

To the human ear, the perception characteristics of sound do not change on a linear frequency scale (human hearing is not that good), but can be expressed in a series of limited frequency bands called critical frequency bands. Simply put, the entire frequency band is divided into several segments, and in each frequency band the auditory perception of the human ear is the same, that is, the psychoacoustic characteristics are the same.
Then, according to this principle, the mp3 compression work can be simply divided into two parts:

The first step: dividing the original audio data into several subcritical frequency bands according to certain principles;

Step 2: Analyze the frequency spectrum according to the psychoacoustic model to find the masking effect curve. Then, according to this curve, each sub-frequency band is quantized separately, and finally the compression of the audio is below the masking effect curve.

In this way, mp3 compression is done. And it is surprising that mp3 is really compressed in the digital world, but it belongs to compression without distortion for human perception.