Efficient convertors between binary and decimal numbers

Alternatives to the built-in functions bin2dec & dec2bin, exhibiting a somewhat faster performance.

Description

These programs were developed for fast conversions between binary and decimal integers (there are other programs by fellow Matlab-users, dealing with fractions). They can be useful when the binary numbers are in vector form instead of strings.

b2d takes a binary number in the form of an array and calculates the equivalent decimal one. Nothing fancy, but quite effective as a component of a program that needs to be efficient in terms of computational overhead.

d2b is the symmetrical program, taking a decimal number and providing the equivalent binary one, in the form of an array.

Usage: y = b2d(x), y = d2b(x)

• MATLAB release: MATLAB 7.8 (R2009a)
• Other requirements: These programs are quite simple, so they can probably run smoothly on even the earliest releases of Matlab.

[Roberto Olmi]

A little bug for the trivial input value: x=1. Add these lines to fix:
if x==1
   y=1;
   return
end

[Ali]

More correct version:

function y = d2b(x)

% Convert a decimanl number into a binary array
%
% Similar to dec2bin but yields a numerical array instead of a string and is found to
% be rather faster

if x==1
   y=1;
   return
end

c = ceil(log(x)/log(2)) + 1; % Number of divisions necessary ( rounding up the log2(x) )
y(c) = 0; % Initialize output array
for i = 1:c
    r = floor(x / 2);
    y(c+1-i) = x - 2*r;
    x = r;
end

% If there is a preceding one, remove it.
if(y(1) == 0)
    y(1) = [];
end

[Marc Lalancette]

Actually, the bug was for any power of 2 (1, 2, 4, ...). Ali's modification works, but a more elegant fix is:
c = floor(log2(x)) + 1;
No need for any other modification from original code (no special case for 1, no need to remove preceding zeros).