Stephan Schreiner, FhG, presented
This contribution discusses conventional practice in the audio industry. One step is mastering, which indicates application of e.g. compression, equalization or balance and level adjustment In one conventional practice, a stereo mix is produced and mastering is applied to that mix. In another, an intermediate step is introduced of producing stem mixes (“stems”), e.g. of similar sounds, mastering is applied to the stems and then the result is used to produce a final stereo signal. Post-processing is “mastering on the playback side” which can be done in the decoder or player. When considering this in practice, the contribution suggests that there may be an advantage to adopting the Stem-Mastering methodology.
The contribution notes opportunities for standardization might be:
-
metadata bitstream syntax
-
separator/discriminator data (i.e. re-create “stems”)
-
manipulators (e.g. compression functions)
-
profiles appropriate to classes of applications
Hyunkook Lee, LG, asked if it is necessary to transmit the stem signals if it is expected that the metadata would manipulate the stem signals. He further noted that mastering engineers would typically wish to use proprietary, state-of-the-art dynamics processors (e.g. compressors).
Pierrick, Philippe, Orange Labs, noted that one of the figures in the contribution looks surprisingly like the Interactive Music MAF, and asked if this work it is the intention to extend this MAF.
Kristofer Kjörling, Dolby, noted that a broadcaster that aggregates for broadcast will do significant post-processing (e.g. level adjustment, time-stretching, program insertion), and metadata should serve that purpose. He further noted that each National Body that has broadcasting members should try to submit information on current practice.
Kate Grant, Nine Tiles, noted that IEC 62379 specifies ways to support audio metadata in the broadcasting chain.
Mukta Kar, CableLabs, noted that metadata should have a low datarate, be linked and synchronized to the audio representation and be easily accessible (e.g. easy to extract from compressed bitstream or available as a readily accessible side-information stream). He noted that audio might learn about current practice in e.g. ATSC specification.
Dostları ilə paylaş: |