How to convert floating point to bits in MATLAB?

[testing test]

Hello,

I am reading a .wav file using wavread command and then selecting the first column to select the left channel only. But the values of the channels are in floating point like 0.000349 but I need these values to be in bits. How to convert these values to bits? Is there any built-in command or any algorithm that you developed?

Thank you.

>> x=wavread('new2.wav');
>> y=x(:,1);

 

[JoannesPaulus]

I do not have a .wav file with me but according to matlab help:

[Y,FS,NBITS]=WAVREAD(FILE) returns the sample rate (FS) in Hertz
and the number of bits per sample (NBITS) used to encode the
data in the file.

You just need to multiply your vector by 2^(NBITS-1)

 

[testing test]

I do not have a .wav file with me but according to matlab help:

[Y,FS,NBITS]=WAVREAD(FILE) returns the sample rate (FS) in Hertz
and the number of bits per sample (NBITS) used to encode the
data in the file.

You just need to multiply your vector by 2^(NBITS-1)

What is NBits here? Also, if i give you a wav file, can you give me a sample code?

 

[JoannesPaulus]

NBITS is one of the 3 outputs of the wavread function. You need to write the following code:

[x,fs,nbits]=wavread('new2.wav');
y=floor(2^(nbits-1)*x(:,1));

 

[testing test]

It just converts to positive and negative decimal numbers.

 

[JoannesPaulus]

Oh... now I see what you mean... Add the following:

y=dec2bin(floor(2^(nbits-1)*x(:,1))+2^(nbits-2));

 

[exfiber]

Oh... now I see what you mean... Add the following:

y=dec2bin(floor(2^(nbits-1)*x(:,1))+2^(nbits-2));

That works for me. Much appreciated.

 

[testing test]

Oh... now I see what you mean... Add the following:

y=dec2bin(floor(2^(nbits-1)*x(:,1))+2^(nbits-2));

Thanks. It was helpful.

 

[testing test]

Can you please elaborate what you are trying to do in this line?

 

[gen0357]

There is a "quantize" function available in MATLAB which will directly convert the floating point numbers into fixed-point number of desired word length.

 

[mowgli]

There is a "quantize" function available in MATLAB which will directly convert the floating point numbers into fixed-point number of desired word length.

True,
use the following Matlab instructions

q= quantizer[x,y] % set the Fixed point format to X bit with Y decimal digits
h=num2hex(q, number) % h is the string which represents the hex value in FP

example
q = quantizer ([16,15]); %FP 16 bit Q15 format
h=num2hex(q,-0.5);

returns the string h = c000

regards