Data Compression


Survey Formats


Compression Methods

Data Formats


Audio Data

MPEG-1 Audio

MPEG-1 Layer I to III

Sampling and Bit Rate

Variable Bit Rates

Principle of the coding


Acoustic Model

Entropy Coding

Encoding (Diagram)

Image Data

Video Data




Acoustic Model

The converted data will be evaluated with the help of a psychoacoustic model. One takes advantage of different aspects of the human hearing. In an ideal scenario, only contents should be removed which is hardly or not at all audible. The following appoaches are offered:

Masking effects

Loud parts cover quiet ones. Frequently the example of the fly in the engine room of the Titanic is used. Generally it applies, that the human hearing focusses on loud tones and does not notice quiet tones of high frequency nearby.

Masking of successive tones

The human auditory can only be aware of the louder tone, if two tones immediately succeed each other.

Joint Stereo Effect

Only higher frequencies can be located perfectly in their direction. The reversal conclusion is, that humans are not able to perceive the direction of deep tones. This circumstance one utilises for dimensioning loudspeaker systems by assigning deep tones only to a single subwoofer. For data reduction, deep tones do not need to be encoded for both channels in parallel (in particular if these are passed to a single subwoofer afterwards).

 <   ^   > 

MPEG-1 Audio Filter Entropy Coding