function p=gaussian_prob(x, m, C, use_log)
% p=gaussian_prob(x, m, C, use_log)
%
% Evaluate the multi-variate density with mean vector m and covariance
% matrix C for the input vector x.
% Vectorized version: Here X is a matrix of column vectors, and p is 
% a vector of probabilities for each vector.

if nargin<4, use_log = 0; end

d   = length(m);

if size(x,1)~=d
   x=x';
end
N       = size(x,2);

m       = m(:);
M       = m*ones(1,N);
denom   = (2*pi)^(d/2)*sqrt(abs(det(C)));
invC    = inv(C);
mahal   = sum(((x-M)'*invC).*(x-M)',2);   % Chris Bregler's trick

switch use_log,
case 2,
  p     = mahal;
case 1,
  p     = -0.5*mahal - log(denom);
case 0,
  numer = exp(-0.5*mahal);
  p     = numer/denom;
otherwise
    error('Unsupported log type')
end
    
    
% if use_log == 2,
%   p     = mahal;
% elseif use_log==1,
%   p     = -0.5*mahal - log(denom);
% else
%   numer = exp(-0.5*mahal);
%   p     = numer/denom;
% end