readme.txt

我认为很不错的语音处理的ｍａｔｌａｂ源代码

Last Modified on 19 January 2009
Created by Edward L. Zechmann



The main program is Main_sound 



Main_sound prompts the user to input all of the information 
to identify and characterize impulsive noise from time recordings.  
Additionally a summary table of impulisve noise metrics can be written to a file.

Impulsive_Noise_Meter controls the flow of data for processing all of the files.  
make_summary_impls_stats_table makes the summary table.   


These programs were written by Edward L. Zechmann for the purposes of 
identifying impulsive noise peaks and characterizing them with impulsive 
noise metrics.  William Murphy and Chucri Kardous wrote the first version
of this program.  


local_peaks.m is the impulsive peak finder.

plot_peaks.m plots the time record and indicates the impulsive peaks with black circles.

snd_peak_metrics cacualtes all of the impulsive noise metrics.  


The following list of programs can be run separately.

Leq_all_calc.m		% Calculates the A-weighted level and other data
Aweight_time_filter.m	% applies the A-weighting filter to the time record
Cweight_time_filter.m	% applies the C-weighting filter to the time record
Nth_oct_time_filter2.m  	
m_round	

abcd_durations.m	% Calculates the A, B, C, and D durations
A_duration.m		% Calculates the A-duration 
B_duration.m		% Calculates the B-duration 
C_duration.m		% Calculates the C-duration 
D_duration.m		% Calculates the D-duration 
kurtosis2.m		% Calculates the kurtosis metric

data_outliers3.m  	% Identifies and removes outliers and calculates 
			% descriptive statistics of the metrics.
			% This would be difficult to run but possible. 



Four examples show how to use the Main_Sound and Impulsive_Noise_Meter.  
Main_Sound can also calculate a summary table of metrics 
see make_summary_impls_stats_table.m.


% It saves data to several files
default_data_file.mat is a matlab data file storing the sound data 
default_data_file.wav is wave file storing a sound data record
fileout.txt stores the sttistical analysis of the impulsive noise metrics 
fileout.mat stores the processed data 

Copy below this line to the next break line and paste in the matlab command line
% % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Example='1';

Fs_SP=100000; fc=1000; td=1; tau=0.01; delay=0.1; A1=8; A2=21;
[SP1, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
Fs_SP=100000; fc=1000; td=1; tau=0.1; delay=0.1; A1=2; A2=22;
[SP2, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
Fs_SP=100000; fc=1000; td=2; tau=0.4; delay=0.1; A1=2; A2=17;
[SP3, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
Fs_SP=100000; fc=2000; td=1; tau=0.1; delay=0.1; A1=8; A2=15;
[SP4, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
SP=[[0.9*SP4 SP1 0.3*SP3 SP2 0.5*SP4 0.2*SP1 SP3 5*SP2];...
    [SP1 0.7*SP4  10*SP2 SP3 0.4*SP2 0.5*SP3 SP1 1*SP4];...
    [SP3 0.2*SP1  0.5*SP4 2*SP2 0.4*SP2 0.7*SP3 0.3*SP1 10*SP4];...
    [SP1 0.7*SP2  0.3*SP3 SP4 0.4*SP1 10*SP2 SP3 1*SP4]];
%
% simulate 8 impulses in 4 microphones with different time records
%
% The process of making the example data added jump discontinuities to the
% data and the dicontinuities can be removed with the following code

[buf, SP]=sub_mean(SP, Fs_SP, 2000);

% % save the SP data variable to a matfile data file

save('default_data_file.mat', 'SP', 'Fs_SP');

% % save the data to a wav file

cf=1/(10*max(max(max(abs(SP)))));
[pp, buf, rdr]=sd_round(cf, 3);
numstr=eval(['sprintf( ', ' ''%9.', num2str(rdr), 'f''', ', pp )']);
fprintf(1, '%s\r', ['When Prompted use the calibration factor ', numstr]);

wavwrite(cf*SP', Fs_SP, 16, 'default_data_file.wav');

%
% Run the Main_Sound just respond to the prompts!
%

Main_Sound;


% % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%



% Example='2';


%
% Try to run the Impulsive_Noise_Meter form the Command Line 
% 
Fs_SP=100000; fc=1000; td=1; tau=0.01; delay=0.1; A1=8; A2=21;
[SP1, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
Fs_SP=100000; fc=1000; td=1; tau=0.1; delay=0.1; A1=2; A2=22;
[SP2, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
Fs_SP=100000; fc=1000; td=2; tau=0.4; delay=0.1; A1=2; A2=17;
[SP3, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
Fs_SP=100000; fc=2000; td=1; tau=0.1; delay=0.1; A1=8; A2=15;
[SP4, t]=analytic_impulse(Fs_SP, fc, td, tau, delay, A1, A2);
SP=[[0.9*SP4 SP1 0.3*SP3 SP2 0.5*SP4 0.2*SP1 SP3 5*SP2];...
    [SP1 0.7*SP4  10*SP2 SP3 0.4*SP2 0.5*SP3 SP1 1*SP4];...
    [SP3 0.2*SP1  0.5*SP4 2*SP2 0.4*SP2 0.7*SP3 0.3*SP1 10*SP4];...
    [SP1 0.7*SP2  0.3*SP3 SP4 0.4*SP1 10*SP2 SP3 1*SP4]];
%
% simulate 8 impulses in 4 microphones with different time records
%
% The process of making the example data added jump discontinuities to the
% data and the dicontinuities can be removed with the following code

[buf, SP]=sub_mean(SP, Fs_SP, 2000);

% % save the SP data variable to a matfile data file

save('default_data_file.mat', 'SP', 'Fs_SP');

% % save the data to a wav file

cf=1/(10*max(max(max(abs(SP)))));
[pp, buf, rdr]=sd_round(cf, 3);
numstr=eval(['sprintf( ', ' ''%9.', num2str(rdr), 'f''', ', pp )']);
fprintf(1, '%s\r', ['When Prompted use the calibration factor ', numstr]);

wavwrite(cf*SP', Fs_SP, 16, 'default_data_file.wav');

filenamesin={'default_data_file.mat', 'default_data_file.wav'};
fileout_txt='fileout.txt';
fileout_struct='fileout.mat';
save_struct=1;
peak_interval=0.5;
make_plot=1;
save_plot=1;
fig_format=[1];
Tool_Name='Hammer';
same_ylim=1;
min_peak=30;
sar=1;
plot_str={'Protected', 'Unprotected'};
Align_peaks=2;
portrait_landscape=1;
peak_alignment_tol=0.25;
percent1=[];
exhaust_cycle=[];
sod=0;

[s]=Impulsive_Noise_Meter(filenamesin, fileout_txt, fileout_struct, save_struct, peak_interval, make_plot, save_plot, fig_format, Tool_Name, same_ylim, min_peak, sar, plot_str, Align_peaks, portrait_landscape, peak_alignment_tol, percent1, exhaust_cycle, sod );




% % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%



% Example='3';
% 
% Try to analyze the Guns_example.wav using the command line with 
% the Impulsive_Noise_Meter
%
% When prompted use a calibration factor of 0.0001
%
% Note that mic 2 does not have adequate dynamic resolution

filenamesin={'Guns_example.wav'};
fileout_txt='fileout.txt';
fileout_struct='fileout.mat';
save_struct=1;
peak_interval=0.1;
make_plot=1;
save_plot=1;
fig_format=[1];
Tool_Name='Guns';
same_ylim=1;
min_peak=30;
sar=1;
plot_str={'Unprotected', 'Protected'};
Align_peaks=1;
portrait_landscape=1;
peak_alignment_tol=0.25;
percent1=[];
exhaust_cycle=[];
sod=0;


[s]=Impulsive_Noise_Meter(filenamesin, fileout_txt, fileout_struct, save_struct, peak_interval, make_plot, save_plot, fig_format, Tool_Name, same_ylim, min_peak, sar, plot_str, Align_peaks, portrait_landscape, peak_alignment_tol, percent1, exhaust_cycle, sod )


% % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%



% Example='4';
%
% Run the Main_Sound and open the Guns_example.wav this time!
% 

Main_Sound;




% % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%