用于 wav 文件频谱密度分析的单位存在问题（db 人为低）

Question

我有以下代码，它改编自我在这里找到的其他代码。 它运行良好，plot 看起来和预期的一样，除了 db 级别太低了 ~80db。 请注意，预期单位为 db re 1uPa。 非常感谢有关解决此问题的任何意见，我认为这与需要 re 1uPa 有关，但我无法追溯问题，我只是一个试图学习编码的可怜的生物学家！

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% FFT parameters
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
NFFT = 96000;    % 
NOVERLAP = round(0.75*NFFT);
w = hanning(NFFT);
% spectrogram dB scale
spectrogram_dB_scale = 100;  % dB range scale (means , the lowest displayed level is XX dB below the max level)
gain = 20
sensitivity = - 165
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% load signal
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
[data,Fs]=audioread('sample.wav'); %(newer matlab)
samples = length(data);
dt = 1/Fs;
t = (0:dt:(samples-1)*dt);
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% display 1 : averaged FFT spectrum
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
[sensor_spectrum, freq] = pwelch(data,w,NOVERLAP,NFFT,Fs);
% convert to dB scale (ref = 1)
sensor_spectrum_dB = 20*log10(sensor_spectrum) - sensitivity - gain ;


figure(1),semilogx(freq,sensor_spectrum_dB);grid
title(['Averaged FFT Spectrum  / Fs = ' num2str(Fs) ' Hz / Delta f = ' num2str(freq(2)-freq(1)) ' Hz ']);
xlabel('Frequency (Hz)');ylabel('Amplitude (dB (L))');

Answer 1

1. 无论如何，您想要使用10*log10的 plot 和 pwelch 的pwelch ，因为您正在处理功率谱。

2. 如果不了解更多关于您的数据的信息，将很难提供帮助……这是可靠的数据吗？ 生理？ 什么是灵敏度单位？ 什么是gain ？？

3. 缩放 PSD