语音信号处理12235135.docx

语音信号处理课程考核报告题目1:绘制同一段语音的宽带语谱图和窄带语谱图题目2:基于倒谱图判断浊音的基音周期学号:201212235135 姓名:黄子文班级:电信DB1204班2015年5月20日题目一：绘制同一段语音的宽带语谱图和窄带语谱图实验原理和部分代码首先是声音的录入和播放，我用的是wavrecord和wavplay。两个函数的基本功能如下： (1)wavrecord y=wavrecord(n,Fs) records n samples of audio signal,sampled at a rate of Fs Hz(samples per second). The default value for Fs is 11025Hz Fs为采样频率，一般默认为11025Hz。n为采样点数，一般是Fs的t倍，其中t为采样的时长。 (2)wavplaywavplay（y，Fs）plays the audo signal stored in the vector y on a PC-based audio output device. Fs is the integer sample rate in Hz(samples per second). The default value for Fs is 11025Hz.这个函数就是可以播放刚录入的音频信号，在这里我主要是检测一下录入是否成功。以下是录音和放音的程序：fs=16000; %è??ù?μ?êduration=2; %??ò?ê±??fprintf(Press any key to start %g seconds of recording...\n,duration);pause;fprintf(Recording...\n);y=wavrecord(duration*fs,fs); %duration*fs ê?×üμ?2é?ùμ?êyfprintf(Finished recording.\n);fprintf(Press any key to play the recording...\n);pause;wavplay(y,fs);将这个程序保存为record01.m绘制这段语音的宽带语谱图和窄带语谱图。这里，我定义了一个函数specgram_ex3p19(x, winlgh, frmlgh, sampling_rate)，这个函数的代码和注释如下：function [spec] = specgram_ex3p19(x, winlgh, frmlgh, sampling_rate); % specgram_hw3p20(x, winlgh, frmlgh, sampling_rate)%% function to compute a spectrogram %% x = input waveform % winlgh = window length in samples% frmlgh = frame length in samples% sampling_rate = samples/sec% PARAMETERS%sampling_rate = 10000;% sampling ratelfft = 1024; % FFT lengthlfft2 = lfft/2;%winlgh = 200; % (128) window length (in samples)%frmlgh = 10; % frame interval (in samples)noverlap = winlgh - frmlgh;% x = x(1:4500); x = 2.0*x/max(abs(x));etime = length(x)/sampling_rate;%---------------------------------------------------------------------------%spec = abs(specgram(x, lfft, sampling_rate, winlgh, noverlap));subplot(211);plot((1:length(x))/sampling_rate,x)xlabel(Time (s));title(SPEECH);axis([0 etime -2.5 2.5]);grid;subplot(212)imagesc(0:.010:etime, 0:1000:(sampling_rate/2), log10(abs(spec)));axis(xy)xlabel(Time (ms)),