snr.m

matlab仿真通过的降噪程序

% ----------------------------------------------------------------------
%
%   Segmental Signal-to-Noise Ratio Objective Speech Quality Measure
%
%		    Robust Speech Processing Laboratory	
%			  Duke University, USA
%			   Copyright (c) 1998
%			  All Rights Reserved.
%
%  Description:
%
%     This function implements the segmental signal-to-noise ratio
%     defined on page 45 of [1] (see Equation 2.12).
%
%  Input/Output:
%
%     The input is a reference 8kHz sampled speech, and processed 
%     speech (could be noisy or enhanced).  
%
%     This function returns 2 parameters.  The first item is the
%     overall SNR for the two speech signals.  The second value
%     is the segmental signal-to-noise ratio (1 seg-snr per 
%     frame of input).  The segmental SNR is clamped to range 
%     between 35dB and -10dB see suggestions in [2].
%
%  References:
%
%     [1] S. R. Quackenbush, T. P. Barnwell, and M. A. Clements,
%	    Objective Measures of Speech Quality.  Prentice Hall
%	    Advanced Reference Series, Englewood Cliffs, NJ, 1988,
%	    ISBN: 0-13-629056-6.
%
%     [2] P. E. Papamichalis, Practical Approaches to Speech 
%	    Coding, Prentice-Hall, Englewood Cliffs, NJ, 1987.
%	    ISBN: 0-13-689019-9. (see pages 179-181).
%
%  Authors:
%
%     Bryan L. Pellom and John H. L. Hansen
%     Robust Speech Processing Laboratory, Duke University
%     Department of Electrical Engineeering
%
%  Last Modified:
%
%     July 22, 1998
%
% ----------------------------------------------------------------------

function [overall_snr, segmental_snr] = snr(clean_speech, processed_speech)

% ----------------------------------------------------------------------
% Check the length of the clean and processed speech.  Must be the same.
% ----------------------------------------------------------------------

clean_length      = length(clean_speech);
processed_length  = length(processed_speech);

if (clean_length ~= processed_length)
  disp('Error: Both Speech Files must be same length.');
  return
end

% ----------------------------------------------------------------------
% Scale both clean speech and processed speech to have same dynamic
% range.  Also remove DC component from each signal
% ----------------------------------------------------------------------

%clean_speech     = clean_speech     - mean(clean_speech);
%processed_speech = processed_speech - mean(processed_speech);

%processed_speech = processed_speech.*(max(abs(clean_speech))/...
%				      max(abs(processed_speech)));

overall_snr = 10*log10(sum(clean_speech.^2)/sum((clean_speech-processed_speech).^2));

% ----------------------------------------------------------------------
% Global Variables
% ----------------------------------------------------------------------

sample_rate =  8000;		   % default sample rate
winlength   =  240;		   % window length in samples
skiprate    =  60;		   % window skip in samples
MIN_SNR     = -10;		   % minimum SNR in dB
MAX_SNR     =  35;		   % maximum SNR in dB

% ----------------------------------------------------------------------
% For each frame of input speech, calculate the Segmental SNR
% ----------------------------------------------------------------------

num_frames = clean_length/skiprate-(winlength/skiprate); % number of frames
start      = 1;					% starting sample
window     = 0.5*(1 - cos(2*pi*(1:winlength)'/(winlength+1)));

for frame_count = 1:num_frames

   % ----------------------------------------------------------
   % (1) Get the Frames for the test and reference speech. 
   %     Multiply by Hanning Window.
   % ----------------------------------------------------------

   clean_frame = clean_speech(start:start+winlength-1);
   processed_frame = processed_speech(start:start+winlength-1);
   clean_frame = clean_frame.*window;
   processed_frame = processed_frame.*window;

   % ----------------------------------------------------------
   % (2) Compute the Segmental SNR
   % ----------------------------------------------------------

   signal_energy = sum(clean_frame.^2);
   noise_energy  = sum((clean_frame-processed_frame).^2);
   segmental_snr(frame_count) = 10*log10(signal_energy/noise_energy);
   segmental_snr(frame_count) = max(segmental_snr(frame_count),MIN_SNR);
   segmental_snr(frame_count) = min(segmental_snr(frame_count),MAX_SNR);
   start = start + skiprate;

end