As predicted in the foreword to AES41 in 2000, digital compression techniques now dominate the broadcast television environment. In addition to the problems foreseen relating to cascaded compression, new problems have arisen because of the use of loudness control and surround sound with those digital compression techniques.
Metadata within the compressed audio bit-stream is used to control loudness and the mixing down of multichannel surround sound to two-channel stereo. These metadata are usually known by terms such as "dialnorm", "prog_ref_level", and "down-mix coefficients".
Whilst this might seem unrelated to the original scope of AES41, dealing with bit allocations and scale factors, it is simply another form of data that can affect a later encoding of the audio: this time is it is more macroscopic than microscopic.
The metadata is lost when the bit-stream is uncompressed unless provision is made to transport it or store it somewhere. Existing methods rely on non-audio mechanisms to convey the metadata alongside the audio, for example a serial data link like RS-422 and serial digital video SMPTE 259M, or a "chunk" in an audio file (for metadata that does not change).
This revision extends AES41 to include data formats for carrying this loudness and down-mix metadata with the uncompressed PCM using the same transport mechanism as before. The metadata can therefore be carried in the audio to which it relates.
Posted: Monday, December 21, 2009
« Call for Comment on DRAFT AES59-xxxx, 25-way… | Main | AES3-2009; new multi-part revision published »
