Mehmet Reha Civanlar

Optimal pulse shape design using projections onto convex sets

Recent development in digital signal processing (DSP) hardware has made it possible to use almost any pulse shape for digital data transmission, resulting in a search for design algorithms capable of constructing pulse shapes matching the properties of a given channel. The technique of projection onto convex sets is particularly suitable for the solution of this problem because of its flexibility in modeling a variety of constraints. The use of the technique for the optimal pulse design problem is demonstrated through by an example from power line communications.<<ETX>>

Scalable video coding in frequency domain

Scalable video coding is important in a number of applications where video needs to be decoded and displayed at a variety of resolution scales. It is more efficient than simulcasting, in which all desired resolution scales are coded totally independent of one another within the constraint of a fixed available bandwidth. In this paper, we focus on scalability using the frequency domain approach. We employ the framework proposed for the ongoing second phase of Motion Picture Experts Group (MPEG-2) standard to study the performance of one such scheme and investigate improvements aimed at increasing its efficiency. Practical issues related to multiplexing of encoded data of various resolution scales to facilitate decoding are considered. Simulations are performed to investigate the potential of a chosen frequency domain scheme. Various prospects and limitations are also discussed.

Three-dimensional data visualization using DSP meshes

This paper presents a parallel scheme for three dimensional data visualization at interactive rates. The scheme is particularly suitable for multiprocessor systems with distributed frame buffers and is currently implemented on an AT&T Pixel Machine, a parallel computer based on mesh connected digital signal processors with a distributed frame buffer. Nearly linear performance increase with the number of processors in the mesh is obtained by partitioning the original three dimensional data into sub-blocks and processing each sub-block in parallel. The approach is very flexible in implementing a variety of visualization techniques, such as volume compositing (translucent models), binary-class and percentage mixtures and surface based volume rendering.

Image processing performance evaluation for DSP based parallel computers with distributed frame buffers

Performance evaluation criteria are presented to provide a useful tool for comparing current digital signal processing (DSP)-based parallel machines used in image processing (IP) applications. Additionally, these criteria can be considered as a step toward a unified computation model for IP which may influence the design of future parallel architectures using distributed frame buffers. As an example, it is shown how to determine the criteria for a parallel machine with a mesh connected architecture and 2-D interleaving scheme. Benchmarks are derived as a function of key architectural parameters (e.g. the number of processors) and image parameters. The use of the criteria in determining task performance is explained.<<ETX>>

Lessons learned on reliable transmission of real-time MPEG-2 streams over ATM

This paper describes a system that has been designed and built at AT&T Bell Labs for studying transmission of real- time MPEG-2 video over ATM networks for multi-cast applications. The set-up comprises a hardware real-time MPEG-2 video, audio and system encoder, an ATM network adaptation module for MPEG-2 transport over AAL-5, and ATM switch, a software system decoder and a hardware elementary stream decoder. The MPEG-2 transport stream has been characterized in terms of robustness to errors. This preliminary study showed the higher importance of the structural information of the stream (PES packet headers TS headers, sequence, picture headers, etc.) with respect to the coded video data (motion vectors, DCT coefficients, etc.). A brief study of the current MPEG-2 hardware decoding architectures allowed us to better understand the effects of bit-stream errors on the resulting video quality. In our experiments, while the loss of some structural data such as picture start codes led the hardware decoder to loose synchronization or to freeze, the loss of video data only affected the image quality. Furthermore the recovery times from a loss of synchronization were orders of magnitude higher than the recovery from some video data loss. An error-resilient real-time software transport stream decoder has been developed. In multiplex-wide operations (i.e. operations on the entire transport stream) it takes advantage of ring buffers and manages the timing information appropriately. In video-stream specific operations it uses resynchronization mechanisms at the picture level which exploit the redundancy of the PES and transport stream syntax. Furthermore time data transfers between the system decoder and the elementary stream decoder are employed. Experiments show that proper use of these methods can significantly improve the system performance.

Workstation-PC multipoint video teleconferencing implementation on packet networks

The increasing availability of high speed local area networks (LANs) and the recent developments in image coding hardware that follow the international standardizations have made multipoint desktop video teleconferencing feasible using mostly off-the-shelf hardware and software. We have implemented a multipoint video teleconferencing system using workstations and personal computers (PCs) on a fiber distributed data interface (FDDI) LAN. The system is based on motion JPEG image coding and uses commercially available hardware and software wherever possible. In this paper, we outline our design experiences.