Streaming Normalization

Abstract

A normalization for streaming digital audio signals applies a gain factor according to the maximum sample magnitude in a window of samples and compare the gain factor to prior gain factors to adjust the gain factor for the samples in the window of samples. Adaptation of the gain factor with rapid decreases but slow increases avoids saturation but allows quiet passages.

1 . A method of normalization for a sequence of digital signals, comprising: 
 (a) finding a first gain factor for a first block of digital signals of a sequence of digital signals;    (b) finding a second gain factor for a second block of digital signals; and    (c) adjusting said second gain factor depending upon said first gain factor.    
   
   
       2 . The method of  claim 1 , wherein: 
 (a) said adjusting of step (c) of  claim 1  (i) decreases said second gain factor when said second gain factor is greater than said first gain factor and (ii) increases said second gain factor when said second gain factor is less than said first gain factor.    
   
   
       3 . The method of  claim 2 , wherein: 
 (a) said adjusting of step (c) of  claim 1  (i) decreases said second gain factor by an amount equal to a first constant multiplied by the difference of said second gain factor minus said first gain factor when said second gain factor is greater than said first gain factor and (ii) increases said second gain factor by an amount equal to a second constant multiplied by the difference of said first gain factor minus said second gain factor when said second gain factor is less than said first gain factor, wherein said first constant is larger than said second constant.    
   
   
       4 . The method of  claim 1 , wherein: 
 (a) said finding a second gain factor of step (b) of  claim 1  uses the maximum digital signal value of the signals in said second block; and    (b) said finding a first gain factor of step (a) of  claim 1  uses gain factors of said first block plus other ones of said digital signals in said sequence of digital signals.    
   
   
       5 . The method of  claim 4 , wherein: 
 (a) said second gain factor is constrained to lie in the range 1/0 to K where K is a constant selected from a range of about 2 to 5.    
   
   
       6 - 8 . (canceled)  
   
   
       9 . A receiver for streaming signals, comprising: 
 (a) a playout buffer to receive input signals;    (b) a normalizer coupled to said playout buffer, said normalizer including a current gain circuit, a memory for a prior normalization, and a combiner circuit for updating a prior normalization with a current gain; and    (c) an output coupled to said playout buffer and said normalizer.