Turning binary data into speaker output.
显示 更早的评论
Hi I'm trying to come up with a code that turns binary data into an audio loudness signal for a presentation on air gapping.
I'm not entirely sure how but I would like to do it simply using Matlab to send a distinct loudness change output over time to the computer's speaker. Basically like morse code.
Any white noise audio file can be used as a base. I will be using a recording of a computer fan.
I am aiming for the following translation with 1 second between each:
1 - 80 dB (at your discretion)
0 - 50 dB (at your discretion)
00000 would be 5 seconds of the file at 50 dB
11111 would be 5 seconds of the file at 80 dB
Example:
Text: I love MATLAB :)
Binary: 01001001 00100000 01101100 01101111 01110110 01100101 00100000 01001101 01000001 01010100 01001100 01000001 01000010 00100000 00111010 00101001
采纳的回答
Walter Roberson
2020-7-17
编辑：Walter Roberson
2020-7-17
MATLAB cannot control the loudness of a sound. It can control the relative loudness, but not the absolute loudness. With Windows, you can make ActiveX calls to control the volume settings of your output device... but that cannot control the absolute loudness, just the electrical signal level being fed into your amplifier.
80 dB compared to 50 dB is 1000 times louder. The audio systems typically are 16 bit, so for the 50 dB sound you would have to restrict your samples to the +/- 0.001 range while your 80 dB would be the whole +/- 1.0 range.
[whitenoise, Fs] = audioread('WhiteNoiseSource.mp4');
noise5 = ifft( fft(whitenoise), Fs*5 );
n5L = length(noise5);
Text = 'I love MATLAB :)';
Binary = logical(reshape(dec2bin(Text, 8).', 1, []) - '0');
RelVol = 0.001 * ones(size(Binary));
RelVol(Binary) = 1;
BL = length(RelVol);
shaped_noise = repelem(RelVol, n5L) .* repelem(noise5, BL);
sound(shaped_noise, Fs)
6 个评论
Matthew Speijer
2020-7-17
编辑：Matthew Speijer
2020-7-17
Walter Roberson
2020-7-17
Ah, your noise file is not mono.
noisefile = '/WashingMachine-16-44p1-stereo-10secs.wav';
[whitenoise, Fs] = audioread(noisefile);
nchan = size(whitenoise, 2);
noise5 = abs(ifft( fft(whitenoise, [], 1), Fs*5 ));
n5L = size(noise5, 1);
Text = 'I love MATLAB :)';
Binary = logical(reshape(dec2bin(Text, 8).', [], 1) - '0');
RelVol = 0.001 * ones(size(Binary));
RelVol(Binary) = 1;
RelVol = repmat(RelVol, 1, nchan);
RL = size(RelVol,1);
shaped_noise = repelem(RelVol, n5L, 1) .* repmat(noise5, RL, 1);
audiowrite('shaped_noise.wav', shaped_noise, Fs);
On my system, using my built-in speakers, the 50dB version is not audible, and just sounds like silence for that period.
The noise5 computation I do in the code has the effect of resampling the frequency. A more refined code would do something like repeating the noise, or chopping off the appropriate amount of noise. Such as
if size(whitenoise, 1) > Fs*5
noise5 = whitenoise(1:Fs*5, :);
else
noise5 = abs(ifft( fft(whitenoise, [], 1), Fs*5 ));
end
Matthew Speijer
2020-7-17
编辑：Matthew Speijer
2020-7-18
Walter Roberson
2020-7-18
00000 would be 5 seconds of the file at 50 dB
11111 would be 5 seconds of the file at 80 dB
Was your intention to group every 5 bits and use that as a fraction to interpolate between 50 and 80 dB? Because I just read it as indicating that every bit should result in 5 seconds of output. Perhaps every bit should result in 1 second of output?
My interpretation was that each bit should result in the sound (re-) playing for 5 seconds. Did you intend that instead the sound play through as many times as needed with the current bit indicating the loudness for 1 second, a loudness modulation rather than a repetition?
Matthew Speijer
2020-7-18
编辑：Matthew Speijer
2020-7-18
更多回答（0 个）
类别
在 帮助中心 和 File Exchange 中查找有关 Audio Processing Algorithm Design 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
