Math Science Chemistry Economics Biology News Search
Mp3 is a format of compression that has represented a real revolution in the musical field. Previously, the exchange of music tracks through the Net required long times. Now, thanks to the development of the sharing programs (like Napster, Winmx, Kazaa, etc.), few minutes are enough for having any track. In fact a format that reduced an audio track from 40 MB to 4 MB had never been used before.
The audio tracks of an audio CD, to be used on a computer, require specific programs to converts “*.cda” files into wave format. However the wave tracks, because of their excellent quality, possess a lot of information, so they have very big dimensions (one song could have an average size of 50-60 MB, corresponding approximately to 43 floppies!). So, the transfers of audio files into the Net is extremely difficult.
Mp3 means MPEG1 Layer3 and is essentially a compression format of audio data that allows to reduce the dimensions of an audio track a lot, without loss of quality. This happens because the compression algorithm eliminates parts of the "redundant" sounds, which can’t be perceived by the human ear. Currently the web sites that give tracks or players proliferate on the Net, the musical choice is very large and updated and the audio quality is really appreciable. The relation between dimensions and quality is what has carried the format Mp3 to an enormous success.
The technology of the Mp3 format is older than we could imagine. The scientific research began in 1987 in the Fraunhofer Institut Integrierte Schaltungen (Fraunhofer IIS-A) at Erlangen, in Germany, with the help of Dieter Seitzer, a teacher at the University of Erlangen. After about three years of studies in 1989 the new format was approved in Germany and, in 1992, it was approved also by the International Standards Organisation (ISO) and integrated in the "Motion Picture Experts Group's (MPEG) specification", a standard for video and audio compression. It was in the MPEG laboratory, where Leonardo Chiariglione works (one of the best well-known Italians over the Internet who has devoted himself to the development of the Mpeg formats) that the first experiments on MPEG format were carried out. Now the MPEG laboratories are working for the evolution of the MPEG -7 format; this isn’t a compression standard (like the others) but it is a description standard that will allow to seek an image recognizing its inside forms and colors (v. Xml). Then, the diffusion of Mp3 became fast, above all with the arrival of the AMP (Advanced Multimedia Products), the first Mp3 player developed by Tomislav Uzelac that was subsequently used like model for the famous WinAmp and MacAmp.
Obviously, the Mp3 format is based on audio studies, particularly on the Psychoacoustic science. The technology of Mp3 compression analyzes the sound, divides it, eliminates a lot of the sounds that don't correspond to the sound models, and maintains the sound that coincides with them. For the compression, the user can specify the number of bits allocated for each sound second (called bitrate): high bitrate gives better sound resolution. You can imagine a movie: with more frames per second the video will be fluid; the same, a higher bitrate will correspond to a better sound, faithful to the original.
The sounds are erased but not all of them. The signal is analyzed and the bit distribution is defined; afterwards, it is divided into little parts, elaborated separately by their algorithms. The frequencies of each frame are compared with the psychoacoustics models contained in the compressor. Then it determines which frequencies have to be elaborated, maintaining them and cutting the others. The disguise effects are very important: if there is a strong sound and a weak sound it is possible to eliminate the weak sound, calculating the milliseconds during which it won't be perceptible. In the same way, two sounds that are overlapped (because of the strong intensity) or static parts of the sound (silence, rustle of the sound) are masked intelligently.
In a Mp3 stereo of X kbps, X/2 for each channel.
For example: a 192 kbps music track the half of the bitrate for each channel will be used. (192:2=96 kbps).
Therefore a second of music at 128 kbps will occupy 16 kb. The best parts of the tracks are compressed with the CBR (constant Bitrate) method. In this method Mp3 will have the same bitrate during all the compression: if the sound is simple it will play well, but if the sound is complicated the bitrate could not be sufficient and the compressor will be forced to cut more sounds. The answer to this problem is the VBR (Variable BitRate), where the passages of the simple sound (like voice and few tools) will be encoded with fewer bits; whereas, in the complicated points, a better bitrate will be applied, with the purpose of maintaining high quality without any cut of frequencies. In the CBR method the bitrate is chosen while in the VBR method is chosen on the level of its quality.
Therefore if we cut the frequencies above the 18 KHz the compressor will have fewer frequencies to manage and it will use the bit at disposal for low frequencies, which contain the biggest part of the sound. So we will improve the quality without increasing of bitrate.
In the image below, the frequencies perceptible by human ears:
The Home-Stereo often applies a 20 KHz Cutoff filter to delete the high frequencies which contains rustle and useless sounds, and could cause distorsions in the amplifiers. Moreover, digital music contains a range of frequencies from 0 to 22 KHz (N. o. R.: to 44.1 KHz according to some sources), while in real life there are sounds of higher frequencies. For that reason there isn’t any acoustic difference, perceptible by our ear, between Cd Audio and Mp3 files.
Prof. Giusy Lipera.