binary process, biner to decimal

i have binary number 20 bit (uint8), i want to convert the binary number to decimal. i use bin2dec but before i must convert num2str so i can use bin2dec. 20 bit (1x20) --> num2str --> 1x58 char. when i start to dec2bin, matlab error because binary string must be 52 bits or less. how to overcome this problem ?

[Jan]

The text description of the input is in general less exact as a shart Matlab example. What exactly is "a binary number 20 bit (uint8)? Is this a [1 x 20] uint8 vector, or a [1 x 3] uint8 vector with 4 unused bits?

Answer 1

[Walter Roberson]

b = uint8([1 1 0 0 1 1 0 1 0 1 1 1 0 0 1 1 0 1 0 1]);
bin2dec( char(b + '0') )

[Internazionale]

it works, can you describe to me how it works ? especially when bin2dec (char(b+'0')), thanks

[Walter Roberson]

'0' is the character representation of zero. add 0 (the number) to it and you will have a value that is still the value for the character representation of '0'. Use char() to turn the value back into the character. So char(0 + '0') is going to be '0'.

Now, the character representation of '1' has a value that is exactly 1 greater than the value for the character representation '0'. ('1' - '0') is 1. So char(1 + '0') = '1'.

Therefore to turn 0 or 1 into '0' or '1', add '0' to the value and take char() of the result.

Answer 2

[Jan]

Guessing that your input is e.g.:

b = uint8([1 1 0 0 1 1 0 1 0 1 1 1 0 0 1 1 0 1 0 1]);

Then assuming, that the 1st element is the most significant bit:

d = 2 .^ (length(b)-1:-1:0) * double(b(:));

This is the dot-product of an [1xN] and [Nx1] vector, such that the sum is calculated implicitly.

[Internazionale]

and the output decimal ?