THE MAGAZINE FOR COMPUTER APPLICATIONS
Circuit Cellar Online offers articles illustrating creative solutions
and unique applications through complete projects, practical
tutorials, and useful design techniques.

MPEG AND DSP INTEGRATION

by Priyesh Surati & David Austin

Start ý MPEG Audio Encoding and Decoding ý Encoder Technology ý Decoder Technology ý Affect on the Music Industry ý Current Uses ý Sources and PDF

ENCODER TECHNOLOGY

The encoder is a computer algorithm that uses psychoacoustic models, also known as perceptual models, to compress the raw digital audio file into an MP3 file (see Figure 4).

Figure 4ýHere you can see the various blocks of the encoder and how they interact with each other.

The compression algorithm uses psychoacoustic models to reduce the size of the raw audio file, a uniform quantizer, and encoding. Using a Fast Fourier Transform (FFT), the original audio is transformed from the time domain to the frequency domain to provide amplitude for every frequency component. [4]

The psychoacoustic model uses the fact that the human ear can only hear sounds of certain amplitudes and only frequencies between 15 and 20,000 Hz. Knowing this, the size of the audio signal is reduced.

Take, for example, an audio signal that has a loud bang (uses a large number of frequencies) and a soft sound (uses a small number of frequencies) at the same time. The compression algorithm eliminates the soft sound because it is unlikely that anyone would be able to hear it. This technique is called sound masking. If a strong signal appears, the weaker signal behind it is not perceivable. The MPEG algorithm removes this weaker signal (see Figure 5). This is advantageous because the information that is removed no longer needs to occupy hard disk space or Internet bandwidth.

Figure 5ýSignal 1 and Signal 2 would be masked, and Signal 3 would be used by the MPEG algorithm.

The compression algorithm reduces the bandwidth of the original signal because, in most cases, frequencies at the extreme of our perception (i.e., 15 Hz and 20,000 Hz) cannot be heard, so they are filtered out. Although these examples are simplified, these basic ideas are used to compress the digital audio signal into the MP3 file format.

STREAMING THE MP3 FILE

One of the essential features of an MP3 file, after it has been encoded, is its ability to be "streamed." Streaming is the process of sending a segment of bits to a computer via the Internet. Streaming MP3 songs makes it possible to have CD-quality radio stations on the Internet.

With older sound file formats like WAV, AIFF, and AU, the sound files are large. To play these file types, you have to wait for the entire file to download over the Internet before you can start listening.

With streaming audio you can start listening right away. In other words, the sound starts playing without waiting for the entire sound file to finish downloading. Streaming audio is only possible with highly compressed sound file formats or fast modems.

In the past, most on-line radio stations sounded like an AM radio station because the signalýs bandwidth had to be reduced in order to be streamed to a standard 56-kbps modem. Now, with the use of cable modem technology, these same radio stations can use MP3 technology to broadcast high-quality sound over the Internet.

PREVIOUS • NEXT