V_PSYCDIGIT measures psychometric function using TIDIGITS stimuli
Usage:
(1)[m,v]=v_psycdigit([],[],'GMPWrn10',[],[],[],[],[],dfile);
% measure SRT using addditive white noise, repetitions allowed, data in .WAV format
(2)[m,v]=v_psycdigit(@v_specsub,[],'GMPn10',[],[],[],[],[],dfile);
% compare spectral subtraction with unprocessed noisy speech
Inputs:
proc processing function handle (e.g. @v_specsub) called as follows:
y=proc(x,fs,i,r) % process noisy waveform x through model i with parameters r
y=proc(x,fs,snr,i,r) % process clean speech degraded to snr through model i with parameters r
y=proc() % return comment text string describing algorithm (if 'c' option specified)
r parameters for proc (omitted from call if r=[])
mode string containing options (see below for list)
p,q,xp parameters passed on to v_psycest or v_psycestu
noise noise waveform or wav file containing noise
fn noise waveform sample frequency [16000]
dfile path to TIdigits folder
ofile output text file (see below for output file format)
Outputs:
m(2,n,3) estimated srt and slope of all models
m(i,n,k): i={1=srt, 2=slope}, n=model number, k={1=mean, 2=mode (MAP), 3=marginal mode}
v(3,n) estimated covariance matrix entries:
[var(srt) cov(srt,slope) var(slope)]' of n'th model
List of options for "mode" input:
a proc adds its own noise
b* base figure number for plotting results [100]
c y=proc() returns comment string
e/E* plot evolving psychometric functions *=1,2,3 for mean, mode, marginal mode (F=after each trial)
f/F* plot psychometric functions *=1,2,3 for mean, mode, marginal mode (F=after each trial)
g prompt with number of digits
G prompt with SNR
l* min token length (in digits)
L* max token length (if different from min)
m* use * external models [default 1]
M include an extra model with no processing
n* *=average number of trials per model
p/P plot pdf (P=after each trial)
r allow repetitions
s respond s to save the noisy stimulus as a .wav file
t/T* taper noise */10 seconds at start and end
v/V* plot srt/slope convergence (V= after each trial)
*=1,2,3 for mean, mode, marginal mode
x* add */10 seconds of noise to the front of the speech before processing
X* truncate */10 seconds from the start of the processed speech
W data is in microsoft WAV format
z omit tokens containing "oh"
Z omit tokens containing "zero"
Output file format
Each line starts 'x ' where x is one of the following
% Comment
V File version type
O mode options
P details about proc
C Comment returned by proc
M measurement0001 function [m,v]=v_psycdigit(proc,r,mode,p,q,xp,noise,fn,dfile,ofile) 0002 %V_PSYCDIGIT measures psychometric function using TIDIGITS stimuli 0003 % 0004 % Usage: 0005 % (1)[m,v]=v_psycdigit([],[],'GMPWrn10',[],[],[],[],[],dfile); 0006 % % measure SRT using addditive white noise, repetitions allowed, data in .WAV format 0007 % (2)[m,v]=v_psycdigit(@v_specsub,[],'GMPn10',[],[],[],[],[],dfile); 0008 % % compare spectral subtraction with unprocessed noisy speech 0009 % 0010 % Inputs: 0011 % proc processing function handle (e.g. @v_specsub) called as follows: 0012 % y=proc(x,fs,i,r) % process noisy waveform x through model i with parameters r 0013 % y=proc(x,fs,snr,i,r) % process clean speech degraded to snr through model i with parameters r 0014 % y=proc() % return comment text string describing algorithm (if 'c' option specified) 0015 % r parameters for proc (omitted from call if r=[]) 0016 % mode string containing options (see below for list) 0017 % p,q,xp parameters passed on to v_psycest or v_psycestu 0018 % noise noise waveform or wav file containing noise 0019 % fn noise waveform sample frequency [16000] 0020 % dfile path to TIdigits folder 0021 % ofile output text file (see below for output file format) 0022 % 0023 % Outputs: 0024 % m(2,n,3) estimated srt and slope of all models 0025 % m(i,n,k): i={1=srt, 2=slope}, n=model number, k={1=mean, 2=mode (MAP), 3=marginal mode} 0026 % v(3,n) estimated covariance matrix entries: 0027 % [var(srt) cov(srt,slope) var(slope)]' of n'th model 0028 % 0029 % List of options for "mode" input: 0030 % a proc adds its own noise 0031 % b* base figure number for plotting results [100] 0032 % c y=proc() returns comment string 0033 % e/E* plot evolving psychometric functions *=1,2,3 for mean, mode, marginal mode (F=after each trial) 0034 % f/F* plot psychometric functions *=1,2,3 for mean, mode, marginal mode (F=after each trial) 0035 % g prompt with number of digits 0036 % G prompt with SNR 0037 % l* min token length (in digits) 0038 % L* max token length (if different from min) 0039 % m* use * external models [default 1] 0040 % M include an extra model with no processing 0041 % n* *=average number of trials per model 0042 % p/P plot pdf (P=after each trial) 0043 % r allow repetitions 0044 % s respond s to save the noisy stimulus as a .wav file 0045 % t/T* taper noise */10 seconds at start and end 0046 % v/V* plot srt/slope convergence (V= after each trial) 0047 % *=1,2,3 for mean, mode, marginal mode 0048 % x* add */10 seconds of noise to the front of the speech before processing 0049 % X* truncate */10 seconds from the start of the processed speech 0050 % W data is in microsoft WAV format 0051 % z omit tokens containing "oh" 0052 % Z omit tokens containing "zero" 0053 % 0054 % Output file format 0055 % Each line starts 'x ' where x is one of the following 0056 % % Comment 0057 % V File version type 0058 % O mode options 0059 % P details about proc 0060 % C Comment returned by proc 0061 % M measurement 0062 0063 % Future mode options: 0064 % [ d score as single digits ] 0065 % [ i/I plot SRT improvement (I=after each trial) ] 0066 % [ j* scaling: 0=autoscale each token, 1=constant speech,2=const noise, 3=const noisy ] 0067 % [N* ignore the first * trials ] 0068 % [o/O write to output file, O write result of each probe] 0069 % [ S save all stimuli as wav files ] 0070 % [ u do not normalize the noise level ] 0071 % [ y* type of noise ] 0072 % 0073 % Bugs/Suggestions 0074 % (1) Add sounds to indicate error and end of test 0075 % (2) Routine could return a label to replace "SNR" in necessary 0076 % (3) Add an input option to "discard" this sample 0077 % (4) output file should include mode argument + date + IP address + computer name + r arguments + p/q values 0078 % (5) Quote filenames in output file or else never have anything else on the line 0079 % (6) silence detection doesn't work 0080 0081 % Copyright (C) Mike Brookes 2010 0082 % Version: $Id: v_psycdigit.m 10865 2018-09-21 17:22:45Z dmb $ 0083 % 0084 % VOICEBOX is a MATLAB toolbox for speech processing. 0085 % Home page: http://www.ee.ic.ac.uk/hp/staff/dmb/voicebox/voicebox.html 0086 % 0087 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 0088 % This program is free software; you can redistribute it and/or modify 0089 % it under the terms of the GNU Lesser General Public License as published by 0090 % the Free Software Foundation; either version 3 of the License, or 0091 % (at your option) any later version. 0092 % 0093 % This program is distributed in the hope that it will be useful, 0094 % but WITHOUT ANY WARRANTY; without even the implied warranty of 0095 % MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 0096 % GNU Lesser General Public License for more details. 0097 % 0098 % You can obtain a copy of the GNU Lesser General Public License from 0099 % https://www.gnu.org/licenses/ . 0100 % See files gpl-3.0.txt and lgpl-3.0.txt included in this distribution. 0101 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 0102 0103 % 25+4 m* use * external models [default 1] 0104 % 26 M include an extra model with no processing 0105 % 1 a proc adds its own noise 0106 % 27+5 n* *=average number of trials per model 0107 % 23+2 l* min token length (in digits) 0108 % 24+3 L* max token length (if different from min) 0109 % 7 d score as single digits 0110 % 37 s speech-shaped noise instead of white 0111 % 19+1 j* scaling: 0=autoscale each token, 1=constant speech, 2=const noise, 3=const noisy 0112 % 35 r allow repetitions 0113 % 41 u unimodal psychometric function 0114 % 32 P plot pdf 0115 % 31 p plot srt convergence 0116 % 13 g prompt with number of digits 0117 % 14 G prompt with SNR 0118 % 47+6 x* add */10 seconds of noise to the front of the speech 0119 % 48+7 X* truncate */10 from the start of the processed speech 0120 % input parameters 0121 persistent tok mtok tigflgp digitsp 0122 if nargin<10 0123 ofile=''; 0124 if nargin<9 0125 dfile='F:\home\dmb\data\old_speech\tidigits'; 0126 % dfile='Y:\Speech\TIDigits\Disc 1\TIDIGITS'; 0127 if nargin<8 0128 fn=16000; 0129 if nargin<7 0130 noise=[]; 0131 if nargin<6 0132 xp=[]; 0133 if nargin<5 0134 q=[]; 0135 if nargin<4 0136 p=[]; 0137 if nargin<3 0138 mode=''; 0139 if nargin<2 0140 r=[]; 0141 end 0142 end 0143 end 0144 end 0145 end 0146 end 0147 end 0148 end 0149 end 0150 0151 % parse the mode string 0152 0153 i=1; 0154 pv=[0 3 3 1 25 0 0 1 1 1 1 100]; % pv default values if option is unspecified 0155 px=[0 3 3 1 25 20 18 1 1 1 1 100]; % pv values if option is given without a value 0156 pvs='jlLmnxXvVfFb'; 0157 pf=zeros(1,52); 0158 mval=0; 0159 lmode=length(mode); 0160 while i<=lmode 0161 if i<lmode % read a following integer if it exists 0162 [v,nv,e,ni]=sscanf(mode(i+1:end),'%d',1); 0163 else 0164 nv=0; 0165 ni=1; 0166 end 0167 j=find(mode(i)==pvs,1); 0168 k=1+2*(double(lower(mode(i)))-'a')+(mode(i)<'a'); 0169 if k>=1 && k<=52 0170 pf(k)=1-pf(k); 0171 if ~isempty(j) 0172 if nv==0 0173 pv(j)=px(j); 0174 else 0175 pv(j)=v; 0176 mval=mval || j==4; % m has a value specified 0177 end 0178 end 0179 end 0180 i=i+ni; 0181 end 0182 if isempty(proc) 0183 pv(4)=0; % not allowed any processed models if not process is specified 0184 end 0185 % derived input parameters 0186 0187 varthr=20; % variance threshold for "silence" 0188 nxevo=30; % size of evloving pdf 0189 0190 if pf(23)>pf(24) || pv(2)>pv(3) 0191 pv(3)=pv(2); % Make max token length reasonable 0192 end 0193 if pf(24)>pf(23) 0194 pv(2)=1; % min =1 if only max is specified 0195 end 0196 zmodel=pv(4)+(pf(26)>0); % total number of models (including null model) 0197 ntrial=zmodel*pv(5); % total number of trials 0198 if pf(12)>pf(11) % 'F' specified but not 'f' 0199 pv(10)=pv(11); % copy across the average type selector 0200 end 0201 if pf(11)+pf(12)>0 && (pv(10)<1 || pv(10)>3) 0202 error('Invalid average type for option f/F'); 0203 end 0204 0205 % sort out p argument to v_psycest or v_psycestu 0206 unimode=pf(41); 0207 if isempty(p) 0208 p=0.75; % default recognition rate at threshold 0209 end 0210 if size(p,1)<3-unimode 0211 p(2-unimode,1)=0.04; % miss probability 0212 p(3-unimode,1)=0; % dummy entry for guess probability 0213 end 0214 if size(p,2)<(zmodel) 0215 p=p(:,1+mod(0:zmodel-1,size(p,2))); 0216 end 0217 0218 % first sort out the digit samples 0219 0220 tigflg=[pv(2:3) pf(51:52)]; 0221 if isempty(tok) || any(tigflg~=tigflgp) || ~strcmp(dfile,digitsp) 0222 disp('Scanning TIDIGITS files ...'); 0223 digitsp=dfile; 0224 tigflgp=tigflg; 0225 if any(dfile(end)=='/\') 0226 dfile(end)=[]; % remove a trailing separator 0227 end 0228 dirlist{1}=''; 0229 ntok=0; 0230 mtok=zeros(1,pv(3)); 0231 tok=cell(1,2); 0232 while ~isempty(dirlist) 0233 dd=dir([dfile dirlist{1}]); 0234 for i=1:length(dd) 0235 name=dd(i).name; 0236 if name(1)~='.' % ignore directories starting with '.' 0237 if dd(i).isdir 0238 dirlist{end+1}=[dirlist{1} '\' name]; 0239 elseif length(name)>4 && strcmpi(name(end-3:end),'.wav') 0240 digs=name(1:end-4); 0241 digz=upper(digs)=='Z'; 0242 digo=upper(digs)=='O'; 0243 digs(digo | digz)='0'; 0244 digs=digs(digs>='0' & digs<='9'); 0245 ndigs=length(digs); 0246 if ndigs>=tigflg(1) && ndigs<=tigflg(2) && ~(tigflg(3) && any(digo)) && ~(tigflg(4) && any(digz)) 0247 ntok=ntok+1; 0248 mtok(ndigs)=mtok(ndigs)+1; 0249 tok{ntok,1}=[dirlist{1} '\' name]; 0250 tok{ntok,2}=digs; 0251 end 0252 end 0253 end 0254 end 0255 dirlist(1)=[]; % remove this directory from the list 0256 end 0257 end 0258 ntok=size(tok,1); 0259 if pf(46) 0260 [s,fs]=v_readwav([dfile tok{1,1}]); % get the first speech token to set fs 0261 else 0262 [s,fs]=v_readsph([dfile tok{1,1}]); % get the first speech token to set fs 0263 end 0264 0265 % calculate guess probability assuming you get the correct number of digits 0266 0267 if any(mode=='d') 0268 p(3-unimode,:)=0.1; 0269 else 0270 p(3-unimode,:)=0.1.^(1:pv(3))*mtok'/ntok; 0271 end 0272 0273 % now initialize the models 0274 0275 if unimode 0276 [xx,ii,m,v]=v_psycestu(-zmodel,p,q,xp); % initialize all models 0277 [pact,qact]=v_psycestu(0); % save the actual parameters 0278 else 0279 [xx,ii,m,v]=v_psycest(-zmodel,p,q,xp); % initialize all models 0280 [pact,qact]=v_psycest(0); % save the actual parameters 0281 end 0282 if pv(4) && pf(5) 0283 x=[]; % set x=[] to force a description output 0284 ii=1; % for now, only do model 1 0285 if pf(1) % if process adds its own noise 0286 if isempty(r) % proc does not require any auxilliary parameters 0287 if mval 0288 procdesc=proc(x,fs,xx,ii); % process the noisy speech 0289 else 0290 procdesc=proc(x,fs,xx); % process the noisy speech 0291 end 0292 else 0293 if mval 0294 procdesc=proc(x,fs,xx,ii,r); % process the noisy speech 0295 else 0296 procdesc=proc(x,fs,xx,r); % process the noisy speech 0297 end 0298 end 0299 else % if process does not add its own noise 0300 if isempty(r) % proc does not require any auxilliary parameters 0301 if mval 0302 procdesc=proc(x,fs,ii); % process the noisy speech 0303 else 0304 procdesc=proc(x,fs); % process the noisy speech 0305 end 0306 else 0307 if mval 0308 procdesc=proc(x,fs,ii,r); % process the noisy speech 0309 else 0310 procdesc=proc(x,fs,r); % process the noisy speech 0311 end 0312 end 0313 end 0314 else 0315 procdesc=''; 0316 end 0317 0318 % now initialize the output file 0319 0320 if pf(29) || pf(30) % o/O output info 0321 nw=fix(datevec(now)); 0322 of=sprintf('psy%4d%02d%02d%02d%02d.txt',nw(1:5)); % filename includes date and time to nearest minute 0323 ofid=fopen(of,'wt'); 0324 if ~ofid 0325 error('Cannot write to %s',of); 0326 end 0327 fprintf(ofid,'%% %s evaluation on %s\n',mfilename,datestr(now)); 0328 fmfnm=[mfilename('fullpath') '.m']; 0329 dd=dir(fmfnm); 0330 fprintf(ofid,'%% %s = %d bytes = %s\n',fmfnm,dd.bytes,dd.date); 0331 fprintf(ofid,'V %d\n',2); % print file format version number 0332 fmfnm=which(func2str(proc)); 0333 dd=dir(fmfnm); 0334 fprintf(ofid,'P %s = %d bytes = %s\n',fmfnm,dd.bytes,dd.date); 0335 if numel(procdesc) 0336 fprintf(ofid,'C %s\n',procdesc); 0337 end 0338 fprintf(ofid,'O %s\n',mode); 0339 end 0340 0341 % now start testing 0342 0343 disp(['Testing ' procdesc]); 0344 % now do the main loop 0345 mnt=zeros(2+2*unimode,zmodel,3,ntrial+1); 0346 vnt=zeros(3+unimode,zmodel,ntrial+1); 0347 mnt(:,:,:,1)=m; 0348 vnt(:,:,1)=v; 0349 i=0; 0350 imax=0; 0351 quitit=0; 0352 while ~quitit 0353 i=i+1; 0354 isp=min(1+floor(rand(1)*ntok),ntok); % select a token 0355 if pf(46) 0356 [s,fs]=v_readwav([dfile tok{isp,1}]); % get the speech token 0357 else 0358 [s,fs]=v_readsph([dfile tok{isp,1}]); % get the speech token 0359 end 0360 s=[zeros(pv(6)*round(fs/10),1); v_activlev(s(:),fs,'n')]; % preappend zeros and normalize speech level 0361 if pf(1) && ii<=pv(4) % if process adds its own noise 0362 if isempty(r) 0363 if mval 0364 y=proc(s,fs,xx,ii); % process the noisy speech 0365 else 0366 y=proc(s,fs,xx); % process the noisy speech 0367 end 0368 else 0369 if mval 0370 y=proc(s,fs,xx,ii,r); % process the noisy speech 0371 else 0372 y=proc(s,fs,xx,r); % process the noisy speech 0373 end 0374 end 0375 else 0376 nn=randn(length(s),1); 0377 x=nn+s*10^(xx/20); % create the data 0378 if ii<=pv(4) 0379 if isempty(r) 0380 if mval 0381 y=proc(x,fs,ii); % process the noisy speech 0382 else 0383 y=proc(x,fs); % process the noisy speech 0384 end 0385 else 0386 if mval 0387 y=proc(x,fs,ii,r); % process the noisy speech 0388 else 0389 y=proc(x,fs,r); % process the noisy speech 0390 end 0391 end 0392 else 0393 y=x; % unprocessed for last model 0394 end 0395 end 0396 y=y(1+pv(7)*round(fs/10):end); % remove junk from the start 0397 if pf(13) 0398 prg=sprintf(' %d',length(tok{isp,2})); 0399 else 0400 prg=''; 0401 end 0402 if pf(14) 0403 prG=sprintf('SNR=%d dB, ',round(xx)); 0404 else 0405 prG=''; 0406 end 0407 if pf(35) 0408 prr=sprintf(', ''r'' to repeat'); 0409 else 0410 prr=''; 0411 end 0412 prompt=[prG 'enter' prg ' digits (''q'' to quit' prr '): ']; 0413 % ansr=-(var(y)>varthr); % meant to detect silences but doesn't work 0414 ansr=-1; 0415 say=1; 0416 while ansr<0 0417 if say 0418 tdel=0; 0419 tic; 0420 soundsc(y,fs); % *** probably shouldn't be ...sc 0421 say=0; 0422 end 0423 rv=input(prompt,'s'); 0424 tdel=toc; 0425 if ~isempty(rv) 0426 if lower(rv(1))=='q' 0427 quitit=1; 0428 ansr=2; 0429 elseif lower(rv(1))=='r' && pf(35) 0430 say=1; 0431 elseif lower(rv(1))=='s' && pf(37) % save the token 0432 ofn=input('File name: ','s'); 0433 if numel(ofn) 0434 v_writewav(y,fs,ofn); 0435 end 0436 elseif all(rv>='0') && all(rv<='9') && ( ~pf(13) || length(rv)==length(tok{isp,2})) 0437 ansr=strcmp(rv,tok{isp,2}); 0438 end 0439 end 0440 end 0441 quitit=quitit || i==ntrial; % mark as quiting if we have done all the trials 0442 jj=ii; % remember which model has just been updated 0443 xxold=xx; % and what the SNR was 0444 if ansr<2 % valid answer: update the pdfs 0445 if unimode 0446 [xx,ii,m,v]=v_psycestu(ii,xx,ansr); 0447 else 0448 [xx,ii,m,v]=v_psycest(ii,xx,ansr); 0449 end 0450 mnt(:,:,:,i+1)=m; 0451 vnt(:,:,i+1)=v; 0452 imax=i; 0453 end 0454 if pf(30) || quitit && pf(29) % 'O/o': output 0455 if ansr>1 0456 rv=num2str(double(rv(1))); 0457 end 0458 % could add in token name and length 0459 fprintf(ofid,'M %d %d %.3g %s %s %d %.1f',i,ii,xxold,tok{isp,2},rv,ansr,tdel); 0460 fprintf(ofid,' %.3g',m(:,jj,:),v(:,jj)); 0461 fprintf(ofid,'\n'); 0462 end 0463 if pf(32) % 'P': plot PDF: figures 1:m 0464 figure(jj); 0465 if unimode 0466 v_psycestu(jj); 0467 else 0468 v_psycest(jj); 0469 end 0470 elseif quitit && pf(31) % 'p': plot PDF: figures 1:m 0471 for jj=1:zmodel 0472 figure(jj); 0473 if unimode 0474 v_psycestu(jj); 0475 else 0476 v_psycest(jj); 0477 end 0478 end 0479 end 0480 if pf(12) || quitit && pf(11) % 'F/f': plot Psychometric function on figure 101 0481 figure(pv(12)+1); 0482 if unimode 0483 qqu.pk=m(1,jj,pv(10)); % peak position 0484 qqu.w=m(2,1,pv(10)); % peak width 0485 qqu.ll=m(3,1,pv(10)); % peak slope on low side 0486 qqu.lh=m(4,1,pv(10)); % peak slope on high side 0487 qqu.gu=pact(2); % guess rate 0488 qqu.la=pact(1); % lapse rate 0489 sw=qqu.w*qqu.ll*qqu.lh/(qqu.ll+qqu.lh); % normalized distance of inflections from peak 0490 xax=linspace(qqu.pk-(4+sw)/qqu.ll,qqu.pk+(4+sw)/qqu.lh,200); 0491 bs=(qqu.lh-qqu.ll)/2; 0492 bu=(qqu.lh+qqu.ll)/2; 0493 plot(xax,qqu.gu+(1-qqu.gu-qqu.la)*(1+2*exp(-sw)*cosh(bs*(xax-qqu.pk)+bu*abs(xax-qqu.pk))+exp(-2*sw)).^(-1)); 0494 else 0495 sd=(pact(1,:)-pact(3,:)).*(1-pact(2,:)-pact(1,:))./(m(2,:,pv(10)).*(1-pact(3,:)-pact(2,:))); 0496 md=m(1,:,pv(10))-sd.*log((pact(1,:)-pact(3,:))./(1-pact(2,:)-pact(1,:))); 0497 xax=linspace(min(md-3*sd),max(md+3*sd),100); 0498 for jj=1:zmodel 0499 plot(xax,v_psychofunc('',[pact(1,jj); m(:,jj,pv(10)); pact(2:3,jj); qact(10)],xax)); 0500 hold on 0501 end 0502 hold off 0503 end 0504 axis([xax(1) xax(end) 0 1]); 0505 xlabel('SNR (dB)'); 0506 ylabel('Recognition Probability'); 0507 title(sprintf('Intelligibility: %s',procdesc)); 0508 end 0509 if pf(10) || quitit && pf(9) % 'E/e': plot evolving Psychometric function on figure 103 0510 figure(pv(12)+3); 0511 psyevo=zeros(ntrial+1,nxevo); % space for evolving pdf 0512 if unimode % unimodal psychometric function 0513 qqu.pk=m(1,jj,pv(10)); % peak position 0514 qqu.w=m(2,1,pv(10)); % peak width 0515 qqu.ll=m(3,1,pv(10)); % peak slope on low side 0516 qqu.lh=m(4,1,pv(10)); % peak slope on high side 0517 qqu.gu=pact(2); % guess rate 0518 qqu.la=pact(1); % lapse rate 0519 sw=qqu.w*qqu.ll*qqu.lh/(qqu.ll+qqu.lh); % normalized distance of inflections from peak 0520 xax=linspace(qqu.pk-(4+sw)/qqu.ll,qqu.pk+(4+sw)/qqu.lh,nxevo); 0521 for iet=1:imax+1 0522 qqu.pk=mnt(1,1,pv(10),iet); % peak position 0523 qqu.w=mnt(2,1,pv(10),iet); % peak width 0524 qqu.ll=mnt(3,1,pv(10),iet); % peak slope on low side 0525 qqu.lh=mnt(4,1,pv(10),iet); % peak slope on high side 0526 sw=qqu.w*qqu.ll*qqu.lh/(qqu.ll+qqu.lh); % normalized distance of inflections from peak 0527 bs=(qqu.lh-qqu.ll)/2; 0528 bu=(qqu.lh+qqu.ll)/2; 0529 psyevo(iet,:)=qqu.gu+(1-qqu.gu-qqu.la)*(1+2*exp(-sw)*cosh(bs*(xax-qqu.pk)+bu*abs(xax-qqu.pk))+exp(-2*sw)).^(-1); 0530 end 0531 imagesc(xax,0:ntrial,psyevo); 0532 axis 'ij' % put trial 0 at the top 0533 else % monotonic psychometric function (not implemented) 0534 end 0535 xlabel('SNR (dB)'); 0536 ylabel('After trial'); 0537 title(sprintf('Intelligibility: %s',procdesc)); 0538 end 0539 if pf(44) || quitit && pf(43) % 'V/v': plot convergence on figure 102 0540 figure(pv(12)+2); 0541 if unimode % unimodal psychometric function 0542 sw=mnt(2,1,pv(10),1:imax+1).*mnt(3,1,pv(10),1:imax+1).*mnt(4,1,pv(10),1:imax+1)./(mnt(3,1,pv(10),1:imax+1).*mnt(4,1,pv(10),1:imax+1)); 0543 subplot(221) 0544 plot(0:imax,squeeze(mnt(1,1,pv(10),1:imax+1))); 0545 set(gca,'xlim',[0 ntrial]); 0546 xlabel('After trial'); 0547 ylabel('Peak position (dB SNR)'); 0548 subplot(222) 0549 plot(0:imax,squeeze(mnt(1,1,pv(10),1:imax+1)-sw.*mnt(3,1,pv(10),1:imax+1)),'-b',0:imax,squeeze(mnt(1,1,pv(10),1:imax+1)+sw.*mnt(4,1,pv(10),1:imax+1)),'-b'); 0550 set(gca,'xlim',[0 ntrial]); 0551 xlabel('After trial'); 0552 ylabel('Inflections (dB)'); 0553 subplot(223) 0554 plot(0:imax,squeeze(mnt(3,1,pv(10),1:imax+1))); 0555 set(gca,'xlim',[0 ntrial]); 0556 xlabel('After trial'); 0557 ylabel('Upwards slope (prob/dB)'); 0558 subplot(224) 0559 plot(0:imax,squeeze(mnt(4,1,pv(10),1:imax+1))); 0560 set(gca,'xlim',[0 ntrial]); 0561 xlabel('After trial'); 0562 ylabel('Downwards slope (prob/dB)'); 0563 else % monotonic psychometric function 0564 subplot(211); 0565 for jj=1:zmodel 0566 plot(0:imax,squeeze(mnt(1,jj,pv(10),1:imax+1))); 0567 hold on 0568 end 0569 hold off 0570 set(gca,'xlim',[0 ntrial]); 0571 xlabel('After trial'); 0572 ylabel('SRT (dB SNR)'); 0573 subplot(212); 0574 for jj=1:zmodel 0575 plot(0:i,squeeze(mnt(2,jj,pv(10),1:imax+1))); 0576 hold on 0577 end 0578 hold off 0579 set(gca,'xlim',[0 ntrial]); 0580 xlabel('After trial'); 0581 ylabel('Slope (Prob/dB)'); 0582 end 0583 end 0584 end % main loop for each probe value 0585 if pf(29) || pf(30) 0586 fclose(ofid); 0587 end