pv2note.m

自动语音识别源码对非特定人的孤立词语音识别算法 进行研究。考虑到语音识别技术在嵌入式系统处理能力较弱、存储空间有限的问 题上进行算法的选择和优化

function note = pv2note(pv, timeUnit, pitchTh, minNoteDuration, plotOpt)
% pv2note: Note segmentation on a give pitch vector
%	Usage: note = pv2note(pv, timeUnit, pitchTh, minNoteDuration, plotOpt)
%
%	For example:
%		pv=[49.21841 48.33041 48.33041 47.76274 48.04425 48.04425 47.76274 48.33041 48.04425 48.04425 48.04425 47.48574 47.48574 47.48574 47.48574 47.48574 47.76274 48.04425 48.04425 47.76274 47.76274 47.76274 47.21309 47.21309 47.21309 47.21309 47.21309 51.48682 51.48682 51.48682 51.48682 51.83658 51.83658 51.83658 51.83658 51.83658 54.92247 54.09792 54.09792 54.09792 54.09792 54.09792 54.09792 54.92247 54.92247 54.92247 54.92247 54.50529 54.50529 54.92247 54.50529 54.50529 54.09792 54.50529 55.34996 54.92247 54.92247 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 53.31086 53.31086 53.31086 53.31086 56.23796 55.78827 55.78827 56.23796 56.69965 56.69965 56.69965 57.17399 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 57.17399 57.17399 57.17399 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 59.76274 59.76274 59.76274 59.21309 59.21309 59.21309 58.68037 58.68037 58.68037 58.68037 54.09792 54.09792 54.09792 54.50529 54.50529 54.09792 54.09792 54.50529 54.92247 54.50529 54.50529 54.09792 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 50.47805 51.48682 51.83658 52.19354 52.55803 52.19354 52.55803 51.83658 51.83658 52.55803 52.93035 52.93035 52.93035 52.93035 52.93035 51.14399 51.14399 54.50529 53.31086 52.55803 52.19354 52.19354 52.19354 52.55803 52.93035 54.09792 54.50529 54.92247 55.78827 56.23796 56.23796 55.78827 55.34996 54.09792 54.09792 54.09792 51.48682 50.15444 50.15444 50.80782 51.14399 51.14399 51.14399 51.14399 52.19354 52.19354 51.83658 51.83658 51.83658 51.48682 51.48682 51.48682 51.83658 51.83658 51.48682 51.48682 51.48682 51.48682 51.48682 50.80782 50.80782 52.55803 51.48682 51.14399 50.80782 51.14399 51.48682 51.48682 51.48682 50.80782 50.80782 50.80782 48.91732 48.62138 48.33041 48.33041 48.33041 48.04425 48.91732 48.91732 48.91732 49.21841 49.21841 48.91732 48.62138 48.33041 48.33041 48.33041 49.83678 48.62138 48.62138 48.62138 48.62138 48.62138 48.91732 49.52484 49.52484 48.91732 48.62138 48.33041];
%		timeUnit=256/8000;
%		pitchTh=0.8;
%		minNoteDuration=0.1;
%		plotOpt=1;
%		note=pv2note(pv, timeUnit, pitchTh, minNoteDuration, plotOpt);
%		fprintf('Hit return to hear the pitch vector...'); pause; fprintf('\n');
%		pvPlay(pv, timeUnit);
%		fprintf('Hit return to hear the segmented note...'); pause; fprintf('\n');
%		notePlay(note); 

%	Roger Jang, 20060527, 20070531

if nargin<1, selfdemo; return; end
if nargin<2, timeUnit=256/8000; end
if nargin<3, pitchTh=0.8; end
if nargin<4, minNoteDuration=0.1; end
if nargin<5, plotOpt=0; end

note=[];
sameNotePv=pv(1);
for i=2:length(pv)
	notePitch=median(sameNotePv);
	noteDuration=length(sameNotePv)*timeUnit;
	if abs(pv(i)-pv(i-1))>pitchTh & noteDuration>minNoteDuration		% If the pitch diff. is high and note duration is long...
		note=[note, notePitch, length(sameNotePv)*timeUnit];		% The time unit is 1 second
		sameNotePv=pv(i);
	else
		sameNotePv=[sameNotePv, pv(i)];
	end
end
% Process the last note
note=[note, median(sameNotePv), length(sameNotePv)*timeUnit];		% The time unit is 1 second

if plotOpt
	plot((1:length(pv))*timeUnit, pv, '.-g');
	hold on; noteplot(note, 1); hold off
end

% ====== Self demo
function selfdemo
pv=[49.21841 48.33041 48.33041 47.76274 48.04425 48.04425 47.76274 48.33041 48.04425 48.04425 48.04425 47.48574 47.48574 47.48574 47.48574 47.48574 47.76274 48.04425 48.04425 47.76274 47.76274 47.76274 47.21309 47.21309 47.21309 47.21309 47.21309 51.48682 51.48682 51.48682 51.48682 51.83658 51.83658 51.83658 51.83658 51.83658 54.92247 54.09792 54.09792 54.09792 54.09792 54.09792 54.09792 54.92247 54.92247 54.92247 54.92247 54.50529 54.50529 54.92247 54.50529 54.50529 54.09792 54.50529 55.34996 54.92247 54.92247 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 54.50529 53.31086 53.31086 53.31086 53.31086 56.23796 55.78827 55.78827 56.23796 56.69965 56.69965 56.69965 57.17399 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 57.17399 57.17399 57.17399 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 56.69965 59.76274 59.76274 59.76274 59.21309 59.21309 59.21309 58.68037 58.68037 58.68037 58.68037 54.09792 54.09792 54.09792 54.50529 54.50529 54.09792 54.09792 54.50529 54.92247 54.50529 54.50529 54.09792 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 53.69992 50.47805 51.48682 51.83658 52.19354 52.55803 52.19354 52.55803 51.83658 51.83658 52.55803 52.93035 52.93035 52.93035 52.93035 52.93035 51.14399 51.14399 54.50529 53.31086 52.55803 52.19354 52.19354 52.19354 52.55803 52.93035 54.09792 54.50529 54.92247 55.78827 56.23796 56.23796 55.78827 55.34996 54.09792 54.09792 54.09792 51.48682 50.15444 50.15444 50.80782 51.14399 51.14399 51.14399 51.14399 52.19354 52.19354 51.83658 51.83658 51.83658 51.48682 51.48682 51.48682 51.83658 51.83658 51.48682 51.48682 51.48682 51.48682 51.48682 50.80782 50.80782 52.55803 51.48682 51.14399 50.80782 51.14399 51.48682 51.48682 51.48682 50.80782 50.80782 50.80782 48.91732 48.62138 48.33041 48.33041 48.33041 48.04425 48.91732 48.91732 48.91732 49.21841 49.21841 48.91732 48.62138 48.33041 48.33041 48.33041 49.83678 48.62138 48.62138 48.62138 48.62138 48.62138 48.91732 49.52484 49.52484 48.91732 48.62138 48.33041];
timeUnit=256/8000;
pitchTh=0.8;
minNoteDuration=0.1;
plotOpt=1;
note=feval(mfilename, pv, timeUnit, pitchTh, minNoteDuration, plotOpt);
fprintf('Hit return to hear the pitch vector...'); pause; fprintf('\n');
pvPlay(pv, timeUnit);
fprintf('Hit return to hear the segmented note...'); pause; fprintf('\n');
notePlay(note, 1);