Method and apparatus for improving the average image refresh rate in a compressed video bitstream

Abstract

An apparatus and method for digital video encoding is disclosed. The disclosed system provides for a way of improving video quality for a given video coding system design.

1 . A method of quality-improvement of a digitally-encoded video sequence, wherein the video sequence comprises information representing a sequence of encoded frames, each encoded frame comprising one or more encoded macroblocks, the method comprising: 
 determining one or more processing capabilities of a decoder that will decode the video sequence;    encoding macroblocks of a first image;.    encoding macroblocks of subsequent images, wherein some macroblocks are skipped; and    increasing video quality as a function of a fraction of macroblocks that are skipped to take advantage of decoder processing capability that would otherwise be unused as a result of the skipped macroblocks.    
   
   
       2 . The method of  claim 1  wherein the step of determining one or more processing capabilities of a decoder comprises having prior knowledge of the decoder type  
   
   
       3 . The method of  claim 1  wherein the step of determining one or more processing capabilities of the decoder comprises receiving processing capability information from the decoder.  
   
   
       4 . The method of  claim 1  wherein the step of determining one or more processing capabilities of the decoder comprises determining the number of macroblocks that can be decoded in a given interval if all macroblocks are skipped.  
   
   
       5 . The method of  claim 4  wherein the step of increasing video quality comprises determining the maximum frame rate in accordance with the following expression:  
     
       
         
           
             MaxFrameRate 
             = 
             
               1 
               
                 
                   
                     N 
                     coded 
                   
                   MaxMBPS 
                 
                 + 
                 
                   
                     N 
                     skipped 
                   
                   MaxSKIPPED 
                 
               
             
           
         
       
     
     where N coded  is the number of coded macroblocks per frame, N skipped  is the number of skipped macroblocks per frame, MaxMBPS is the maximum number of macroblocks that can be decoded in a given interval, and MaxSKIPPED is the maximum number of macroblocks that can be decoded in a given interval if all macroblocks are skipped.  
   
   
       6 . The method of  claim 1  wherein the step of increasing video quality comprises increasing a video frame rate.  
   
   
       7 . The method of  claim 1  wherein the step of increasing video quality comprises increasing a video picture size.  
   
   
       8 . The method of  claim 1  wherein the step of increasing video quality further comprises increasing a video frame rate as a function of a computational cost of the decoder to decode various types of macroblocks.  
   
   
       9 . The method of  claim 1  wherein the step of increasing video quality further comprises increasing a video picture size as a function of a computational cost of the decoder to decode various types of macroblocks.  
   
   
       10 . The method of  claim 1  further comprising: 
 taking account of a number of coefficients included in the encoded macroblocks and a computational requirement of the decoder as a function of this number.    
   
   
       11 . The method of  claim 10  wherein the step of increasing video quality comprises increasing a video frame rate.  
   
   
       12 . The method of  claim 10  wherein the step of increasing video quality comprises increasing a video picture size.  
   
   
       13 . The method of  claim 10  wherein the step of increasing video quality further comprises increasing a video frame rate as a function of a computational cost of the decoder to decode various types of macroblocks.  
   
   
       14 . The method of  claim 10  wherein the step of increasing video quality further comprises increasing a video picture size as a function of a computational cost of the decoder to decode various types of macroblocks.  
   
   
       15 . A video conferencing terminal adapted to produce encoded video including a sequence of encoded frames, each encoded frame comprising one or more encoded macroblocks, the video conferencing terminal comprising: 
 one or more image processing engines adapted to encode a video signal, wherein some macroblocks are skipped; and    a communication interface adapted to determine one or more processing capabilities of a decoder that will decode the encoded video and further adapted to increase video quality as a function of a fraction of macroblocks that are skipped to take advantage of decoder processing capability that would otherwise be unused as a result of the skipped macroblocks.    
   
   
       16 . The video conferencing terminal of  claim 15  wherein the processing capability of the decoder is determined as a function the number of macroblocks that can be decoded in a given interval if all macroblocks are skipped.  
   
   
       17 . The video conferencing terminal of  claim 16  wherein a maximum frame rate is determined in accordance with the following expression:  
     
       
         
           
             MaxFrameRate 
             = 
             
               1 
               
                 
                   
                     N 
                     coded 
                   
                   MaxMBPS 
                 
                 + 
                 
                   
                     N 
                     skipped 
                   
                   MaxSKIPPED 
                 
               
             
           
         
       
     
     where N coded  is the number of coded macroblocks per frame, N skipped  is the number of skipped macroblocks per frame, MaxMBPS is the maximum number of macroblocks that can be decoded in a given interval, and MaxSKIPPED is the maximum number of macroblocks that can be decoded in a given interval if all macroblocks are skipped.  
   
   
       18 . The video conferencing terminal of  claim 15  wherein video quality is increased by increasing a frame rate.  
   
   
       19 . The video conferencing terminal of  claim 15  wherein video quality is increased by increasing an picture size.  
   
   
       20 . The video conferencing terminal of  claim 18  wherein the frame rate is further determined as a function of a computational cost of the decoder to decode various types of macroblocks.  
   
   
       21 . The video conferencing terminal of  claim 19  wherein the picture size is further determined as a function of a computational cost of the decoder to decode various types of macroblocks.  
   
   
       22 . A method of quality-improvement of a digitally-encoded video sequence, the method comprising: 
 determining one or more processing capabilities of a decoder that will decode the video sequence; and    increasing video quality as a function of an encoder model of decoder processing load to take advantage of decoder processing capability that would otherwise be unused.    
   
   
       23 . The method of  claim 22  wherein the step of determining one or more processing capabilities of a decoder comprises having prior knowledge of the decoder type.  
   
   
       24 . The method of  claim 22  wherein the step of determining one or more processing capabilities of the decoder comprises receiving processing capability information from the decoder.  
   
   
       25 . The method of  claim 22  wherein the step of increasing video quality comprises increasing a video frame rate.  
   
   
       26 . The method of  claim 22  wherein the step of increasing video quality comprises increasing a video picture size.  
   
   
       27 . A video encoder for generating an encoded video sequence, comprising: 
 one or more image processing engines adapted to: 
 encode a video signal;  
 determine one or more processing capabilities of a decoder that will decode the encoded video sequence; and  
 increase video quality as a function of an encoder model of decoder processing load to take advantage of decoder processing capability that would otherwise be unused.  
   
   
   
       28 . The video encoder of  claim 27  wherein the processing capabilities of the decoder are determined as a function a number of macroblocks that can be decoded in a given interval if all macroblocks are skipped.  
   
   
       29 . The video encoder of  claim 28  wherein a maximum frame rate is determined in accordance with the following expression:  
     
       
         
           
             MaxFrameRate 
             = 
             
               1 
               
                 
                   
                     N 
                     coded 
                   
                   MaxMBPS 
                 
                 + 
                 
                   
                     N 
                     skipped 
                   
                   MaxSKIPPED 
                 
               
             
           
         
       
     
     where N coded  is the number of coded macroblocks per frame, N skipped  is the number of skipped macroblocks per frame, MaxMBPS is the maximum number of macroblocks that can be decoded in a given interval, and MaxSKIPPED is the maximum number of macroblocks that can be decoded in a given interval if all macroblocks are skipped.  
   
   
       30 . The video encoder of  claim 27  wherein video quality is increased by increasing a frame rate.  
   
   
       31 . The video encoder of  claim 27  wherein video quality is increased by increasing an picture size.  
   
   
       32 . The video encoder of  claim 30  wherein the frame rate is further determined as a function of a computational cost of the decoder to decode various types of macroblocks.  
   
   
       33 . The video encoder of  claim 31  wherein the picture size is further determined as a function of a computational cost of the decoder to decode various types of macroblocks.