International organisation for standardisation organisation internationale de normalisation
Download 3.13 Mb.
|
- Navigate this page:
- 2.2Approval of agenda and allocation of contributions
- 2.3Communications from the Chair
- 2.5Received National Body Comments and Liaison matters
- 2.6Task Groups
- 3.1MPEG-4 Audio Lossless Coding (Sunday 1300 – 1800)
1Opening of the meetingThe MPEG Audio Subgroup meeting was held during the 72nd meeting of WG11, April 18-22, 2005 in Busan, KR. The list of participants is given in Annex A. 2Administrative matters2.1Approval of previous meeting reportThe 71st Audio Subgroup meeting report had been previously distributed by e-mail and was approved. 2.2Approval of agenda and allocation of contributionsThe agenda and schedule for the meeting was discussed, edited and approved. It shows the documents contributed to this meeting and presented to the Audio Subgroup, either in the task groups or in Audio plenary. The Chair brought relevant documents from Requirements, Systems and MDS to the attention of the group. It was revised in the course of the week to reflect the progress of the meeting, and the final version is shown in . 2.3Communications from the ChairThe Chair summarised the issues raised at the Sunday evening Chair’s meeting, proposed task groups for the week, and proposed agenda items for discussion in Audio plenary. 2.4Joint meetingsThe joint meetings with Audio over the course of the week are listed here and are reported on below.
2.5Received National Body Comments and Liaison mattersThe NB Comments and Liaison documents for the meeting that require a response are as shown below.
2.6Task GroupsTask groups were convened for the duration of the MPEG meeting, as shown in Annex C. Results of task group activities are reported below. 3AhG meetingsIn order to obtain a compete view of the business of this MPEG meeting, discussion of contributions in AhG meetings are recorded in this section of the Audio report. 3.1MPEG-4 Audio Lossless Coding (Sunday 1300 – 1800)Eunmi Oh, Samsung, presented
She began with a refresh of topics covered at the last meeting, namely that
The presentation showed that one can uniquely determine on average more than 12 bits (and more than 13 symbols) even though many of the bits in the arithmetic decoding buffer are not known. This capability can actually be viewed as a form of additional “compression” when transmitting truncated bitstreams. Ralph Geiger, FhG, presented
This document confirmed the results presented in 11911. The consensus of the AhG is to recommend to add this technology to the specification text and RM code. It does not require any changes to the syntax. Rongshan Yu, I2R, presented
This document shows that there is some small improvement as a result of adding one bit in the syntax to explicitly signal an M/S stereo IntIMDCT. The consensus of the AhG is to recommend to add this technology to the specification text and RM code. Takehiro Moriya, NTT, presented
This is a cross-check on CE10, which proposed adaptive block switching in ALS. It shows that the adaptive block switching gives small but consistent improvement at all word lengths and sampling rate, and requires no increase in the decoder complexity. The consensus of the AhG is to recommend to add this technology to the specification text and RM code. Takehiro Moriya, NTT, presented
This proposes a harmonization of LTP and the adaptive block switching structure. The new LTP structure and syntax is very much simpler than that in the original LTP CE proposal and performance is comparable. As a bonus, encoder complexity is significantly reduced. Tilman Liebchen, TUB, presented
This is a cross-check on the LTP CE as it is harmonized with adaptive block switching. The document shows that LTP has it greatest improvement in performance when used in conjunction with low order prediction (e.g. K=15) and low sampling rates. It notes that at low order, LTP gives a significant performance improvement, while at high orders, block switching gives a significant performance improvement. The consensus of the AhG is to recommend to add the adaptive block switching and the harmonized LTP technology to the specification text and RM code. Tilman Liebchen, TUB, presented
This is a cross-check on the improved floating point compression, which proposes masked Lempel-Ziv to compress floating point data. In fact, it reports on the performance of all four combinations of the use (or not) of Approximate Common Factor Coding (ACFC) and Masked Lempel-Ziv (MLZ). When both tools are available, and enabled on a block-by-block basis, a performance improvement of more than 5% is realized on IEEE floating point signals. The consensus of the AhG is to recommend to add the ACF and MLZ technology to the specification text and RM code. The Chair notes that the group may want to consider putting this into a separate profile. Ralf Geiger, FhG, presented
This clearly showed that the SLS specification needs a normative deterministic realization of TNS, which is referred to as intTNS, in order to provide good performance when TNS is used in the AAC core. The consensus of the AhG is to recommend to add the intTNS realization to the specification text and RM code. Ralf Geiger, FhG, presented
This contribution shows how to significantly reduce table storage by storing only every 8th table value and interpolating intermediate values using a normative and deterministic interpolation method. Furthermore, it shows that the performance is not significantly affected by the proposed change, and that the accuracy of the IntMDCT is not significantly affected by the proposed change. Since the performance of the original tables and the interpolated tables is comparable, one can also achieve the computational complexity of the original system by pre-computing the interpolated values to build a table of the same size as in the original system (i.e. the original large tables prior to this proposed change). The consensus of the AhG is to recommend to add the proposed table compaction to the specification text and RM code. Takehiro Moriya, NTT, presented
This contribution contains two proposals. The first notes that some biomedical data may contain more than 256 channels, in which case the current 1 byte field is not adequate. Hence it proposes that this field be expanded to 2 bytes. The second proposal is to modify the inter-channel predictor to increase the number of taps from 1 to 3, and to add a second 3-tap predictor separated from the first by some moderately large sample lag. Tilman Liebchen, TUB, presented
This contribution showed some, but not significant, improvement in performance for stereo material. As the order of the short-term predictor increases, the increase in performance due to the inter-channel predictor decreases significantly. AhG recommends that this discussion be continued during the MPEG week. Tilman Liebchen, TUB, presented
The proposed bitstream revisions define a new ALSSpecificConfig(), which now carries all non-audio related payload information and is required for MPEG-4 systems support. Additionally, it proposes some modifications to make random access representations more compact. AhG recommends that this discussion be continued during the MPEG week. Xiaolin Wu, McMaster University, presented
This contribution proposes CBAC as a new entropy coder. It is modular and could easily be adapted into ALS, and has low complexity, both in terms of operations and storage. Tilman Liebchen, TUB, noted that, for medium to large predictor orders, any complexity savings in the entropy coder is overwhelmed by the complexity of the predictor itself. Thomas Wiegand, HHI, noted that CABAC is already rigorously specified as an MPEG tool, has been extensively studied by MPEG, and in addition, its application to audio coding has been investigated by a TUB Ph.D. thesis. This discussion will be continued Tuesday 11AM during the MPEG week. Directory: sc29 -> open -> 29view 29view -> International organisation for standardisation organisation internationale de normalisation 29view -> International organisation for standardisation organisation internationale de normalisation Download 3.13 Mb. Share with your friends:
|