Title: Audio Video Coding, Processing and Optimization
Technical Area: Multimedia
Audio Video Codecs are the key components of all the video cloud applications. Different encoder implementations could lead to very different quality and performance (speed and power) even though they may conform to the same standard, especially when modern video encoding/decoding technologies are becoming more and more complicated.
H.264/AVC and H.265/HEVC are the two widely adopted international video coding standards jointly developed by VCEG from ITU-T and MPEG from ISO/IEC. Their successor, H.266/VVC just had a Call for Proposal (CFP) and the answers from 30+ companies or research institutes look very promising. Meanwhile, there are also quite some other standard activities. For example, AVS from China has been accepted as the IEEE 1857 standard and AV1 from Alliance of Open Media (AOM) has been publicly released.
Besides audio video compression, pre-processing and post-processing are two other areas directly affecting the audio-visual qualities. Jointly optimizing pre- processing, compression and post-processing could result in the best audio-visual quality. Since there is no standard defining pre- or post-processing, the potential for further quality improvement is huge.
The proposed research work should focus on end-to-end audio video quality improvement without ignoring the speed and power constraints, since mobile platform will be one of the major applications.
The final research results should be evaluated with well accepted objective criteria to demonstrate the efficacy and efficiency. We encourage the research work to be published in prestigious journals or be presented in top tier conferences. We also encourage patent filing for innovative ideas.
Algorithm description, intermediate test data, software codes and any other related items should be shared with Alibaba group members.
Our engineers will work with the collaborators closely to provide real application data from our related cloud services if needed and help to integrate the algorithm into our products.
Related Research Topics
Some related research topics are listed as follows, but not limited to:
- Audio and video pre-processing and post-processing which may include, but not limited to, denoising, echo cancellation, super-resolution, frame rate conversion, deblurring, etc.
- New audio video encoding/decoding technology which may include, but not limited to machine learning based video encoding/decoding, H.266/VVC or AVS investigation and research, etc.
- Video encoder/decoder (H.264, H,265, H.266, AVS, AV1) optimization which may include but not limited to subjective quality improvement, speed and power optimization, etc.
- Subjective and objective audio video quality assessment.