How to convert integer to 12 bit binary and vise versa

Hello, I want to convert integers in the range -400 to +800 to 12 bit binary and vice versa.

dec= -333;
a= decimalToBinaryVector(typecast(int16(dec),'uint16'),16);

str_x = num2str(a);
b=typecast(uint16(bin2dec(str_x)),'int16')

The above code gives me 16 bit Binary value. I want to convert the integer to 12 bit binary and vice versa.

[Walter Roberson]

Why? Your code at https://www.mathworks.com/matlabcentral/answers/508682-how-to-pass-binary-values-to-mex already does 12 bit conversion.

[A R]

Yes Walter, the code does 12 bit conversion. It works well with positive integers, but with negative numbers, I get wrong results while converting from binary to decimal.

b=-70.199;
Y = round(b,1);
dec= Y*10;
temp = dec;
mask = temp < 0;
temp(mask) = 2^12 + temp(mask) ;  %temp = 3394
a=decimalToBinaryVector(temp, 12);  %a= [1 1 0 1 0 1 0 0 0 0 1 0]

% binary to decimal
str_x = num2str(a);
b=typecast(uint16(bin2dec(str_x)),'int16')  % b = 3394

The output of b must be -70.199 but I get 3394 as the decimal value.

Answer 1

[Walter Roberson]

mod(typecast(int16(dec), 'uint16'), 4096)

and convert to binary.

Or

bitget(int16(dec), 12:-1:1)

which does the binary conversion. You might want to double() the output for your purposes.

[A R]

Walter, the decimal to binary conversion works well, but I am not getting correct value of the decimal number .

b=-70.199;
Y = round(b,1);
dec= Y*10;
a1=mod(typecast(int16(dec), 'uint16'), 4096)  %a1 = 3394
a=decimalToBinaryVector(a1);  %a1= [1 1 0 1 0 1 0 0 0 0 1 0]
% binary to decimal
str_x = num2str(a);
b1=typecast(uint16(bin2dec(str_x)),'int16')  % b = 3394

The output of b1 is 3394 but the correct output is -70.199

[A R]

b=-70.199;
Y = round(b,1);
dec= Y*10;
a= decimalToBinaryVector(typecast(int16(dec),'uint16'),16);
str_x = num2str(a);
b=typecast(uint16(bin2dec(str_x)),'int16');

The above code does the correct conversion to decimal number. But the decimal to binary value is in 16bit, I want that in 12bits

[Walter Roberson]

b = int16(sum(bitset(0,12:-1:1,a)));
if b > 2047; b = b - 4096; end

[A R]

Hi Walter, your code is the perfect fit to my problem. Thanks a lot.

Answer 2

[Steven Lord]

There's no int12 data type in MATLAB. Depending on what you want to do with this data, if you have Fixed-Point Designer you could store it as an fi object. See this documentation page for the basics of how to work with fi objects.

>> A = sfi(10, 12, 0)
A = 
    10

          DataTypeMode: Fixed-point: binary point scaling
            Signedness: Signed
            WordLength: 12
        FractionLength: 0
        
>> bin(A)
ans =
    '000000001010'

[A R]

Hello Steven, Thanks a lot for your response. Gonna learn something new today, about Fixed point designer. Will look into the documentation page and how it fits my project and will update you..Thanks again.