What Is Audio Streaming?

Audio streaming is the delivery of sound over a network as a continuous flow that a receiver decodes and plays while the transmission is still in progress, rather than downloading a complete file before playback begins. The defining constraint is that the client must maintain uninterrupted output from a buffer holding only a few seconds of material, on a network whose available throughput and delay vary from moment to moment. Everything characteristic of the field follows from that constraint: low-latency codecs, packet loss concealment, adaptive bitrate selection, and buffer control algorithms that trade startup delay against the risk of a stall.

Audio streaming is a branch of streaming media generally, but its engineering profile differs from video in useful ways. Bitrates are one or two orders of magnitude lower, which makes bandwidth adaptation easier, while human sensitivity to gaps and glitches in sound is far sharper than to a dropped video frame, which makes continuity requirements stricter.

Compression and Coding

Perceptual audio coding makes streaming practical by discarding signal components the ear cannot resolve. Encoders apply a filter bank or modified discrete cosine transform, estimate a masking threshold from a psychoacoustic model, and allocate quantization noise so that it stays beneath that threshold in each band. MP3 and Advanced Audio Coding established the approach for stored music; AAC-LC and HE-AAC remain widespread in broadcast and on-demand services. Interactive use added a second requirement, low algorithmic delay, since a codec that needs hundreds of milliseconds of lookahead cannot support conversation. Opus, specified in RFC 6716, addresses this by combining a linear prediction coder for speech with a transform coder for music, switching or hybridizing between them and operating from 6 kbit/s narrowband up to 510 kbit/s fullband at frame sizes as short as 2.5 milliseconds. Lossless formats such as FLAC and ALAC serve archival and high-resolution services where no perceptual loss is acceptable.

Transport Protocols

Two protocol families dominate. The first uses the Real-time Transport Protocol defined in RFC 3550, which carries media over UDP with sequence numbers and timestamps while its companion control protocol reports loss, jitter, and round-trip time back to the sender. Session control for this model is typically handled by the Real-Time Streaming Protocol version 2.0 or by SIP, and the combination underpins voice over IP, WebRTC, and professional contribution links. Losing a packet is preferable to waiting for retransmission, so receivers apply packet loss concealment, interpolating the missing interval from surrounding audio.

The second family runs over HTTP. Media is cut into short segments listed in a manifest, and the client requests them in sequence using ordinary web infrastructure, which passes through firewalls and caches in content delivery networks without special handling. HTTP Live Streaming, specified in RFC 8216, and MPEG-DASH, standardized as ISO/IEC 23009-1, both follow this pattern. Latency is higher, on the order of seconds, but reliability and scale are better, which suits music services, podcasts, and internet radio. Low-latency extensions using chunked transfer and partial segments have narrowed the gap.

Adaptation and Quality Control

Adaptive bitrate logic selects among alternate encodings of the same content based on measured throughput, buffer occupancy, and device capability. Throughput-based estimators react quickly but overshoot on bursty links; buffer-based controllers are steadier but slower to recover headroom, and most deployed players combine both. Jitter buffers absorb variation in packet arrival time, adjusting depth dynamically and, in speech applications, stretching or compressing silence to change buffer depth without audible artifacts. Objective quality is assessed with perceptual models such as PEAQ and POLQA, calibrated against listening tests.

Applications

Audio streaming has applications in a range of fields, including:

• Music and podcast services and internet radio
• Voice over IP, conferencing, and WebRTC communication
• Live event and sports broadcast contribution and distribution
• In-vehicle infotainment and connected home speakers
• Game voice chat and interactive audio
• Telemedicine, remote interpretation, and assistive listening systems