Pulse-Code Modulation

脉冲编码调制 (Pulse-Code Modulation, PCM) is a method for converting an analog signal with continuous time and continuous amplitude into discrete digital values. The encoding process typically involves band-limiting, sampling at a fixed frequency, mapping each sample to a finite number of quantization levels, and representing the quantized results in binary form; during playback, the continuous signal is restored through digital-to-analog conversion and reconstruction filtering. PCM is not exclusive to any particular file format or optical disc; digital telephony, audio workstations, CDs, DVDs, Blu-ray discs, and computer audio can all utilize PCM with different parameters.

The sampling rate indicates the number of samples taken per second, while the bit depth indicates the number of quantization levels available for each sample. For uncompressed integer PCM, the raw data rate is calculated by multiplying the sampling rate, the number of effective bits, and the number of channels; for example, a stereo 44.1 kHz, 16-bit PCM CD-DA track has a data rate of 1,411.2 kbit/s. Increasing the sampling rate expands the representable frequency range, while increasing the bit depth reduces ideal quantization noise and increases the representable dynamic range; actual systems are also affected by analog circuits, clocks, filtering, jitter handling, and recording levels.

“PCM” describes the principles for representing sample values but does not fully specify byte order, integer or floating-point formats, sign representation, channel arrangement, or file encapsulation. Container formats such as WAVE, AIFF, CAF, and MXF can all store PCM; without external parameters, players typically cannot determine the sampling rate, bit depth, and number of channels based solely on the raw PCM data. Linear PCM (LPCM) uses equally spaced quantization levels and is the most common form in music and film production; A-law and μ-law in telephone systems first compress and expand signals according to nonlinear rules and also belong to the 脉冲编码调制 system.

PCM can be stored directly without data compression; it can also serve as input for lossless codecs such as FLAC, ALAC, and MLP, or act as a working signal before or after lossy codecs such as MP3, AAC, and Dolby Digital. Decoding lossy audio to obtain PCM does not mean that the originally discarded information has been restored; it merely indicates that the decoder has converted the compressed bitstream back into a sample sequence suitable for playback and processing.

Fixed-point PCM and floating-point PCM are also common in digital audio processing. The full-scale range of fixed-point formats is well-defined, and integer overflow can cause clipping; floating-point formats provide greater computational headroom within the workstation, facilitating mixing and plugin processing, but are typically converted to a specified integer format when finally output to most consumer interfaces or physical media. PCM parameters and data content should be evaluated separately; a 24-bit file may contain a signal with effective precision lower than 24 bits, and bit depth alone is not sufficient to prove recording quality.