MPEG-1 Audio Layer II
MPEG-1 audio Layer 2 (MP2 ) is a known as a Musicam and standardized by the Moving Picture Experts Group Audio method for data reduction of digital audio data. The abbreviation is also used as a file extension for files containing audio data of this type.
While it was in the PC and the Internet field replaced by MP3 essentially, it remains a dominant standard for digital broadcasting as part of the DAB and DVB standards. It also comes sporadically in the DVD for use.
Emerging from MPEG -1 Audio Layer 2 is the free, popular in the audiophile range Musepack.
For details and a brief historical introduction to MP2, MP3, see.
Functioning of coding with MP2
At the entrance is located on an uncompressed 48 kHz sampled and quantized with 16-bit audio signal. This is then divided into 32 sub-bands of 750 Hz to exploit a weakness of the human ear, thus saving bandwidth. You can now namely per subband determine so-called scale factors (ie, the loudest sound) describe the highest sound pressure level in that subband. If a few milliseconds before or after the loud sound quieter sounds on this need not be mitcodiert. They can delete or because the human ear is not able due to the time masking to perceive these quieter sounds. It has to be transferred to the loud sound.
FFT with 1024 samples
In addition, MP2 makes use of the fact the simultaneous masking effect by the input signal with an FFT ( Fast Fourier Transformation ) transformed and then sampled 1024 times. Since the original signal has a sampling frequency of 48 kHz, the maximum bandwidth of the signal can be 24 kHz according to the sampling. At 1024 samples, this means that all the 20 Hz the FFT signal is sampled. Now you can find out using the psychoacoustic model, if a loud sound quieter sound from the same or any adjacent subbands obscured ( hearing threshold ). If this is the case, the soft tones need not be also coded. This results for each sub-band a different quantization as in the simplest case the subbands may indeed be completely covered, and thereby the quantization is of such a sub-band 0 then bits. Of course, here also the temporal masking of signals must be taken into account.
In the last step the user data (such as music ) are still subjected to a so-called bit -packing and then provided with additional data that contain the volume progression or the number of bits used for example.
MPEG Multichannel - multichannel extension of MP2
See main article MPEG -2 multichannel