More about Audio Streams and Metadata

Created on 2012-11-08 17:18:00

Metadata has been defined as “data about other data”. Some digital audio files and digital transport streams carry metadata that describe aspects of the stream and the audio content.


Audio streams that include metadata


This industry standard transport stream carries audio samples, embedded clock and metadata. For each subframe (or channel) of audio, there is a synchronizing/identification Preamble, followed by audio data, followed by a Validity Bit, a User Bit, a Channel Status Bit (normally just called a Status Bit) and a Parity Bit. The User Bits and Status Bits are collected over 192 frames, providing metadata in the form of a block of 384 user bits and Subframe A and Subframe B blocks of 192 channel status bits.

Coded audio streams

Low-bit rate coded audio such as Dolby Digital (AC3) and DTS Digital Surround can be embedded in the IEC60958 (etc.) transport stream as defined in the standard IEC61937. In such cases, the Validity Bit is set, indicating a non-PCM signal. Additionally, the embedded IEC61937 audio carries its own metadata, specific to the coding format.

HDMI: Stream metadata

The HDMI transport stream carries a great deal of data and metadata. Although in audio test and measurement we are not concerned with much of the HDMI signal, the video formats and high-level stream metadata do affect a receiver’s ability to correctly accept embedded audio, and these metadata settings and readings are exposed in APx500.

HDMI: InfoFrame

HDMI carries some metadata in InfoFrames, and APx500 can set and read Audio InfoFrame data.

HDMI: Embedded audio stream metadata

Although audio on HDMI is packetized and multiplexed into a high-bit rate video stream, the format of the audio mimics IEC60958 audio and includes user bits, status bits and metadata related to coded audio.


HDMI also carries EDID, which is not directly related to the audio but affects how a downstream device may react to the transport stream.


HDMI streams are often encrypted using HDCP encryption, indicated by a metadata flag.


Audio streams that do not include metadata

Analog audio does not carry any digital metadata. Also, pure PCM sampled audio data is simply a pulse train of samples, and does not carry any metadata. I2S, TDM and other chip-level protocols are of this type; thus, the signals passing in or out of the DSIO ports have no metadata.