3Record of AhG meetings 3.1AhG Meeting on SAOC Sunday 1200-1700
Jeongil Seo, ETRI, presented
m15893
|
CE Report on Support for a Mastering Downmix
|
Jeongil Seo
Seungkwon Beack
Kyeongok Kang
Kwangki Kim
Minsoo Hahn
|
The CE proposes two new elements
-
Mastering Downmix Gain
-
Modified quantization scheme
The proposed means to support Mastering dowmix gain is the same as is used in MPEG Surround (Artistic Downmix Gain). However, since mastering downmix MDCT coefficients are biased slightly positive with respect to the mixing downmix, and new quantization scheme is proposed. This effect was observed for many tens of released CDs.
The contribution presented the results of a listening test from ETRI and also listening test results from LG. For both labs, the proposed system is significantly better (at the 95% level of significance) than using SAOC with unaltered mastering downmix signal.
There will be a break-out group (Jeongil Seo, Werner Oomen, Heiko Purnhagen) to discuss other means to accomplish this goal, which will report during the week.
Oliver Hellmuth, FhG, presented
m15828
|
Proposal for adoption of stereo-to-binaural scenario for the MPEG SAOC system
|
Oliver Hellmuth
Johannes Hilpert
Harald Mundt
Leonid Terentiev
Cornelia Falch
Heiko Purnhagen
Jonas Engdegård
Jeroen Koppens
|
This presented evidence on the operation of a “missing mode” in SAOC, the mode being stereo downmix to binaural presentation. Evidence of performance, as a listening test, was presented. Stereo to binaural performed similarly to mono to binaural.
It was the consensus of the AhG to recommend that the Audio Subgroup accept this technology into the SAOC FCD.
Oliver Hellmuth, FhG, presented
m15826
|
Core Experiment on Low Delay MPEG SAOC system
|
Johannes Hilpert
Maria Luis Valero
Markus Schnell
Oliver Hellmuth
Leonid Terentiev
Jonas Engdegård
Heiko Purnhagen
Erik Schuijers
|
This proposes a new core experiment which has the objective of reducing the one-way latency of the SAOC system thus making it better for teleconferencing applications. The reduction of latency is accomplished by replacing the SAOC QMF filterbank with a Complex Low Delay Filterbank (CLDFB) which has an analysis-synthesis latency of only 1.3 ms. This is particularly advantageous if the core coder uses the same filterbank, for example AAC-ELD. Example latencies are shown in the following table from m15826:
Dostları ilə paylaş: |