Audio Compression Techniques

Audio compression reduces the amount of data needed to store or transmit digital sound. The starting point is the uncompressed representation, pulse-code modulation (PCM), in which the sound wave is sampled at a fixed rate and each sample is stored as a number. PCM is simple and faithful but produces large files, so a stereo recording at common rates consumes a substantial amount of storage and bandwidth. Compression schemes aim to shrink this data while preserving as much perceived quality as possible.

Compression methods fall into two categories. Lossless techniques encode the same information more compactly, exploiting redundancy so that the original samples can be reconstructed exactly; the savings are modest. Lossy techniques achieve much greater reduction by discarding information judged unlikely to be heard. These rely on psychoacoustic models that account for how the ear masks quiet sounds near louder ones and is insensitive to certain frequencies, allowing detail to be removed where it will not be noticed.

Several lossy codecs are associated with optical media and broadcasting. AC-3, marketed as Dolby Digital, and DTS both carry multichannel surround audio on DVD and other formats, differing in their coding details and typical data rates. DSD (Direct Stream Digital) takes a different route, using a high-rate one-bit stream rather than multi-bit PCM samples. The choice of technique balances file size, sound quality, the number of channels carried, and the processing power required to encode and decode the stream.

Frequently asked questions

What is PCM?

Pulse-code modulation is the basic uncompressed digital audio format, storing each sampled amplitude as a number. It is faithful but produces large files, motivating compression.

How do lossless and lossy compression differ?

Lossless compression reconstructs the original samples exactly but saves little space. Lossy compression discards inaudible detail using psychoacoustic models, achieving much smaller files with some loss of fidelity.

What are AC-3 and DTS used for?

Both are lossy codecs that carry multichannel surround sound, commonly on DVDs and in broadcasting. They differ in their internal coding and typical data rates.