Mean absolute difference prediction for video encoding rate control

Abstract

Mean absolute difference (MAD) prediction for video encoding may be provided. Upon receiving a video stream comprising a plurality of quality layers, a first quantization parameter (QP) may be selected for a first frame of the video stream according to a second QP associated with a second frame and a third QP associated with a third frame. The first frame may then be encoded according to the selected first QP.

What is claimed is: 
     
         1 . A computer implemented method for providing Mean Absolute Difference prediction, the method comprising:
 receiving a plurality of video frames associated with a bitstream;   predicting a mean absolute difference of a current frame according to a first mean absolute difference associated with an immediately previous temporal frame and a second mean absolute difference associated with a similar frame that has a similar rate distortion characteristic to the current frame; and   encoding the current frame according to the predicted mean absolute difference.   
     
     
         2 . The computer implemented method of  claim 1 , further comprising selecting a quantization parameter of the current frame according to a bandwidth constraint and the predicted mean absolute difference of the current frame. 
     
     
         3 . The computer implemented method of  claim 2 , wherein the quantization parameter of the current frame is determined by a quadratic rate-quantizer model using the predicted mean absolute difference of the current frame. 
     
     
         4 . The computer implemented method of  claim 1 , further comprising predicting the mean absolute difference of the current frame according to a linear regression model, wherein the first mean absolute difference associated with the immediately previous temporal frame and the second mean absolute difference associated with the similar frame comprise regressors associated with the linear regression model. 
     
     
         5 . The computer implemented method of  claim 1 , wherein the similar frame is identified by:
 computing a rate distortion characteristic of each of the plurality of video frames; and   comparing the rate distortion characteristic of each of the plurality of video frames to a current rate distortion characteristic of the current frame.   
     
     
         6 . The computer implemented method of  claim 1 , wherein the bitstream comprises a base layer and at least one enhancement layer. 
     
     
         7 . The computer implemented method of  claim 6 , wherein the immediately previous temporal frame is associated with a different layer than a layer of the current frame. 
     
     
         8 . The computer implemented method of  claim 6 , wherein the immediately previous temporal frame is a previous frame in a highest frame rate layer. 
     
     
         9 . A computer-readable storage device which stores a set of instructions which when executed performs a method for providing quantization parameter prediction, the method comprising:
 receiving a plurality of video frames associated with a bitstream;   predicting a quantization parameter for a current frame according to a quantization parameter associated with an immediately previous temporal frame and a quantization parameter associated with a similar frame that has a similar rate distortion characteristic to the current frame; and   encoding the current frame according to the quantization parameter.   
     
     
         10 . The computer-readable storage device of  claim 9 , wherein the quantization parameter of the current frame is selected according to a bandwidth constraint. 
     
     
         11 . The computer-readable storage device of  claim 9 , further comprising selecting the quantization parameter of the current frame according to a predicting a mean absolute difference of the current frame according to a mean absolute difference associated with the immediately previous temporal frame and a mean absolute difference associated with the similar frame that has the similar rate distortion characteristic to the current frame. 
     
     
         12 . The computer-readable storage device of  claim 10 , wherein the quantization parameter of the current frame is determined by a quadratic rate-quantizer model using the predicted mean absolute difference of the current frame. 
     
     
         13 . The computer-readable storage device of  claim 10 , further comprising predicting a mean absolute difference of the current frame according to a linear regression model, wherein the mean absolute difference associated with the immediately previous temporal frame and the mean absolute difference associated with the similar frame comprise regressors associated with the linear regression model. 
     
     
         14 . The computer-readable storage device of  claim 9 , wherein the similar frame is identified by:
 computing a rate distortion characteristic of each of the plurality of video frames; and   comparing the rate distortion characteristic of each of the plurality of video frames to a current rate distortion characteristic of the current frame.   
     
     
         15 . The computer-readable storage device of  claim 9 , wherein the bitstream comprises a base layer and at least one enhancement layer. 
     
     
         16 . The computer-readable storage device of  claim 9 , wherein the immediately previous temporal frame is associated with a different layer than a layer of the current frame. 
     
     
         17 . The computer-readable storage device of  claim 16 , wherein the immediately previous temporal frame is a previous frame in a highest frame rate layer. 
     
     
         18 . A system for providing quantization parameter prediction, the system comprising:
 a memory storage; and   a processing unit coupled to the memory storage, wherein the processing unit is operative to:
 receive a plurality of video frames associated with a bitstream; 
 predict a mean absolute difference of a current frame according to a mean absolute difference associated with an immediately previous temporal frame and a mean absolute difference associated with a similar frame that has a similar rate distortion characteristic to the current frame; 
 select a quantization parameter for the current frame according to the predicted mean absolute difference of the current frame; and 
 encode the current frame according to the quantization parameter. 
   
     
     
         19 . The system of  claim 18 , wherein the quantization parameter of the current frame is selected according to a bandwidth constraint and the predicted mean absolute difference of the current frame. 
     
     
         20 . The system of  claim 18 , wherein the immediately previous temporal frame is a previous frame in a highest frame rate layer.