|
Could anybody explain as to what "selective enhancement" is all about with regards to the MPEG4 FGS bit plane model. Also could somebody tell me whats the significance of the frequency weighing matrix. thanks and regards Arvind Raman __________________________________________________ |
|
|
MPEG4 related questions
Started by ●April 19, 2001
Reply by ●April 24, 20012001-04-24
|
hi, this may help you Necessity of FGS: The objective of FGS is to get optimized video quality over network at a given rate. We can get optimized video quality by possible three methods 1. Encode the video at lowest bit rate of all possible channel connections 2. Encode the video at highest bit rate of all possible connection (Video server has to transcode for different rates). 3. Generate multiple bit streams of multiple bit rates at the time of encoding. In multicasting network the multiple bit rates are transferred in the network and individual receiver takes one of the bit streams decodes it to reconstruct the video. He can shift from one bit rate to another bit rate to get the best video quality available when channel bit rate varies. For a single or unicast, multiple bit streams are stored in the video server, which comes in between encoder and channel and one of them are transmitted in the channel. According to the feedback the rate is negotiated between video server and decoding terminal. If the number of available bit rates are less then reconstructed video quality will not be good, so for less bit streams are generated we will not be able to cover bit rate with fine granularity. End users will not get the best video with their available channel connection. The optimization has to be done between two things - the number of bit rate generated and storage space required in video server. MPEG - 4 has addressed this problem of achieving FGS by three methods: a. Bit plane coding of the DCT residue b. Wavelet coding of the image residue c. Matching pursuite of the image residue After quantitative research the "Bit plane coding of the DCT residue" is chosen for working draft of MPEG - 4 FGS Amendment. FGS encoder generates two layers - base layer and enhancement layer, that enhancement layer can be truncated to any amount of bits with in a VOP(Video Object Plane). The remaining portion improves the quality of the VOP. The more enhancement bits are received better he quality of the received picture. No single bit rates needs to be given to the encoder, but only a bit rate range. The encoder generates a base layer to meet the lower bound of the bit rate range and an enhancement layer to meet the total bit rate constraint (upper bound of the bit rate range). FGS enhancement bit stream can be sliced into packets to satisfy the varying user bit rates. Selective Enhancement in FGS: For some frames of a video sequence, a part of the frame may be more important than other parts. Therefore we may use bit plane shifting method to put the bit planes of the region of interest earlier in the bit stream. More details can be given later if it of interest to you. regards atanu ----- Original Message ----- From: Arvind Raman <> To: image <> Cc: dsp related <> Sent: Thursday, April 19, 2001 1:28 PM Subject: [imagedsp] MPEG4 related questions > Could anybody explain as to what "selective > enhancement" is all about with regards to the MPEG4 > FGS bit plane model. > Also could somebody tell me whats the significance of > the frequency weighing matrix. > > thanks and regards > Arvind Raman > __________________________________________________ > > _____________________________________ > Note: If you do a simple "reply" with your email client, only the author of this message will receive your answer. You need to do a "reply all" if you want your answer to be distributed to the entire group. > > _____________________________________ > About this discussion group: > > To Join: > > To Post: > > To Leave: > > Archives: http://www.yahoogroups.com/group/imagedsp > > More Groups: http://www.dsprelated.com > > ">http://docs.yahoo.com/info/terms/ > |
