In this project, we design and develop a new type of loudspeakers that project a focused sound beam to the human. This is made possible be exploiting the nonlinear acoustics interaction of finite amplitude ultrasonic wave to generate audible sound by using an array of ultrasonic transducer. Our core AudioBEAM technology offers a number of advantages over conventional sound reproduction technologies.

Organisations and countries (local and international) of collaborators:

• Texas Instruments (Singapore, USA) - sponsored their DSP hardware and development system and provide research funding for other items.
  
• Institute of Acoustics (IoA), Chinese Academy of Sinica (China) - develop Advanced acoustics control technologies by combining expertise of IOA on acoustics and DSP Lab (NTU) on digital signal processing

This is a SingTel/NTU joint applied R&D cooperation project. 

The main objective of the project is to develop an MPEG-4 system to test the performance of the MPEG-4 standard and to investigate its potential applications in the areas of video-on-demand, video transcoding, and video streaming.

This project is supported by the A-star under the SERC grant scheme.

In this project, research investigations will be fully in line with the major technical issues and can be grouped into two sub-fields:

• Multimedia Coding and Streaming - Focused on producing better signal decomposition fro representation, yield appreciably improved rate-distortion and quality-of-service (QoS) performance, while maintaining attractive computational complexity. Furthermore, useful signal processing tools will be created for scalability, transcoding, denoising and artifact removal, etc. The most prominent application would be transmitting wireless video for 3G and even future 4G wireless environment.
• Multimedia Information Indexing and Retrieval - Typical features are color, texture, shape, and motion trajectories of video objects (for video only). Applications that could be benefiting from the generic solutions for indexing and retrieval are vast, ranging from content-based video retrieval, security surveillance, video-on-demand, to traffic or safety monitoring.

This is a joint project with Institute for Infocomm Research

There are two important aspects in visual information presentation and organization:

• Information about the contents (e.g., which is where) - The focus is on the new techniques based on dominant regions, moving objects, text extraction and color information for content-base video processing. Information on dominant regions and color is important to detect scene changes; separation of moving objects and background is the base of individual control of coding scheme in different regions; while the text extraction is particularly useful in digital libraries. Variable frame rate can apply in scenarios where the available bandwidth or storage is a precious item. The rate control strategies range from a simple global motion analysis to more detailed content-based analysis.
• Perception of Information by the End Receiver (human eyes, mostly) - As the second objective of the project, a new quantitative performance assessment method based on the perceptual distortion metrics is to be devised. Such metrics can automatically assess the compressed video instead of carrying out subjective viewing tests from time to time. Investigation will be also extended to new way of image/video coding for better performance, where perception-based quality evaluation is to be employed.