How to perform bitwise xor operation for audio scrambling??

Hello,

I using DWT and make scrambling for mono type audio. In my research, i need to use bit wise xor function for audio coefficients and keys. First, I change the format of audio coefficients from negative float value to 64 bit binary value using "c=num2bin(q,c)" where q is quantizer and c is audio coefficients. The problem is i can't xor with audio coefficients and key. This error occur " Undefined function 'bitxor' for input arguments of type'char'. " How should I continue to get audio scrambled value?? PFA for errors

I would be grateful if someone help me in this matter :)

Answer 1

[Walter Roberson]

Subtract '0' (the character for the digit 0) from the char to get the numeric values.

However, in your previous post on this topic, it was recommended that you use other approaches: in particular typecast to uint64 and then bitget().

[Su Latt Sandi]

Thank alot sir. I tested using bitget and typecast but that is not okay for many number. I'll test to get numeric values by subtraction as you mention.

[Walter Roberson]

typecast() to an uint* type the same size, and then bitget() are guaranteed to work for all numeric items -- but you have to pay attention to the internal bit order.

Note that when you ask to get a particular bit of a floating point number, you are asking about the IEEE 754 representation of the number. For example the most significant bit of all non-nan floating point numbers is always the sign bit.

[Su Latt Sandi]

Hello Sir. I want to inform you that is bit wise xor operation is successfully done.I subtract "0" from num2bin values to get numeric as per your comment. Now, I want to change bin2num formt for bitwise xor values.But I can't get negative float number using bin2num. Please help me and thank you so much for helping me.

[Walter Roberson]

uint8() or uint16() or uint32() or uint64() the result of the bin2num() according to the size of what you num2bin (careful, if you using uint64 you might not get back the correct number with bin2num(). typecast() the result to the original data type of your data.

[Su Latt Sandi]

Sir, i tested with this commend "scramble=typecast(test,'double')". I'm not sure that commend is correct or not. When I run that commend this error occur The first input argument must be a vector.I need to get floating point numbers between -1 to +1 for wavelet reconstruction and audio write process.

[Walter Roberson]

What is test in this context? I can tell from the error message that whatever it is a scalar value that is not int64 or uint64

I would like to take this opportunity to again recommend that you do not use num2bin() and bin2num() for this work.

[sound, fs] = audioread('SomeFileName.wav');
sound_bits = reshape( (dec2bin(typecast(sound, 'uint8'), 8) - '0').', 1, []);

[Su Latt Sandi]

Thanks Sir, it's working properly, and as I'm beginner in Matlab so I'm sorry if i make mistake. And I want to know how to convert back to get values that range of -1 to +1 to write audio.