What is Digital Audio?

Digital Audio: Simple Definition

Digital audio is a continuous sound wave measured at regular intervals and stored as numbers. Each measurement is a sample, and playing the file back means reconstructing a wave from those samples and driving a speaker with it.

Two figures define the raw quality. Sample rate is how many measurements are taken per second, commonly 44,100 for CD audio. Bit depth is how precisely each measurement is recorded, commonly 16 bits, giving 65,536 possible levels per sample.

Channels multiply everything. Mono stores one stream, stereo two, and 5.1 surround six, and the file size scales accordingly. This is why converting a surround soundtrack to stereo shrinks it substantially, and why that conversion is a real downmix rather than a repackaging.

Sampling and Quantization

Converting sound to digital form involves two steps: sampling measures the waveform at regular intervals, and quantization rounds each measurement to a fixed number of bits.^[3] The sampling rate sets the highest reproducible frequency, by the Nyquist principle it must be at least twice that frequency, which is why 44.1 kHz comfortably covers the roughly 20 kHz limit of human hearing.^[3] The bit depth, commonly 16 bits for CD-quality audio, sets how finely each sample is measured and therefore the dynamic range between the quietest and loudest sounds.

Uncompressed, Lossless and Lossy

Audio can be stored uncompressed as raw PCM (as in WAV), losslessly compressed (FLAC) to shrink size with no quality change, or lossily compressed (MP3, AAC, Opus) to shrink it much further by discarding inaudible detail.^[4] Modern codecs such as Opus adapt well across both speech and music at low bitrates.^[4] The choice mirrors the wider lossy-versus-lossless trade-off: WAV and FLAC for archiving and editing where every detail matters, MP3 and AAC for everyday listening and sharing where small files matter more.

Bitrate and Perceived Quality

For lossy audio, the bitrate largely determines quality. Very low bitrates produce the watery, hollow artifacts of over-compression, while higher bitrates preserve more detail at the cost of size. As a practical guide, MP3 around 192 to 256 kbps and AAC around 256 kbps sound transparent to most listeners, meaning indistinguishable from the original in normal listening. Pushing higher gives diminishing returns for lossy formats; if true fidelity is required, a lossless format like FLAC removes the question entirely.

Channels and Containers

An audio stream may carry one channel (mono), two (stereo) or more for surround sound, multiplying the data accordingly.^[1] The encoded stream is then placed in a container, and tools such as FFmpeg can transcode between codecs or remux between containers.^[2]

How Digital Audio Works

The sample rate is not arbitrary. Capturing a frequency requires sampling at more than twice that frequency, and human hearing tops out around 20 kHz, which is why 44.1 kHz became the standard: it comfortably covers the audible range with margin for the filters involved.

Uncompressed audio is therefore predictable and large: sample rate times bit depth times channels gives roughly 10 megabytes per minute for CD-quality stereo. FLAC halves that losslessly by prediction; MP3 and AAC reduce it by an order of magnitude by discarding what the ear will not miss.

Higher sample rates are not automatically better. 96 kHz captures frequencies no human can hear, which is useful headroom while editing and pointless in a finished file. For delivery, 44.1 or 48 kHz is the sane choice, and the extra data buys nothing a listener can detect.

Examples of Digital Audio