MPEG Audio Byte Stream Format

This specification defines segment formats for implementations of Media Source Extensions™ [MEDIA-SOURCE] that choose to support MPEG audio streams specified in [ISO11172-3], [ISO13818-3], and [ISO14496-3].

It defines the MIME-types (see 2. MIME-types) used to signal codecs, and provides the necessary format specific definitions for initialization segments, media segments, and random access points required by the Byte Stream Formats section of the Media Source Extensions™ specification. This document also defines extra behaviors and state that only apply to this byte stream format.

This section specifies the MIME-types that may be passed to isTypeSupported() or addSourceBuffer() for byte streams that conform to this specification.

• "audio/aac" for sequences of ADTS frames, as specified in [ISO14496-3].
• "audio/mpeg" for MPEG-1/2/2.5 Layer I/II/III streams, as specified in [RFC3003].

The "codecs" MIME-type parameter MUST NOT be used with these MIME-types.

The format of an MPEG Audio Frame depends on the MIME type used (see 2. MIME-types).

• If the "audio/aac" MIME-type is used, an MPEG Audio Frame is a sequence of bytes that conform to the adts_frame() syntax specified in Table 1.A.5 of [ISO14496-3].
• If the "audio/mpeg" MIME-type is used, an MPEG Audio Frame is a sequence of bytes that conform to the frame() syntax element specified in Section 2.4.1.2 of [ISO11172-3] or the corresponding definition in [ISO13818-3].

Since [ID3v1], [ID3v2] metadata frames, and Icecast headers are common in existing MPEG audio streams, implementations SHOULD gracefully handle such frames. Zero or more of these metadata frames are allowed to occur before, after, or between MPEG Audio Frame. Minimal implementations MUST accept, consume, and ignore these frames. More advanced implementations MAY choose to expose the metadata information via an inband TextTrack or some other mechanism.

The MPEG audio byte stream is a combination of one or more MPEG Audio Frame and zero or more metadata frames.

• Every MPEG Audio Frame is a random access point.
• Every MPEG Audio Frame header is an initialization segment.
• The coded audio in each MPEG Audio Frame is a media segment.

As well as sections marked as non-normative, all authoring guidelines, diagrams, examples, and notes in this specification are non-normative. Everything else in this specification is normative.

The key words MAY, MUST, MUST NOT, and SHOULD in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.