Inverse Modulus operator (Mod)
Mike Malkum
Is there any way to calculate the inverse of the mod operator in Matlab or simulink?
I appreciate any hint.
Regards,
Mike
Doug Schwarz
"Mike Malkum" <e6c...@yahoo.com> wrote:
I hope you mean the additive inverse, -mod, or the multiplicative
inverse, 1./mod, because if you want to unmod something I'm afraid
you're out of luck.
Suppose you want to know if a number is even or odd,
B = mod(A,2);
Now B is either 0 or 1. Do you think there could be any way to get back
the original number, A, knowing only whether it is even or odd?
--
Doug Schwarz
dmschwarz&ieee,org
Make obvious changes to get real email address.
John D'Errico
> In article <gut0nl$cl5$1...@fred.mathworks.com>,
> "Mike Malkum" <e6c...@yahoo.com> wrote:
>
> > Hello everyone,
> > Is there any way to calculate the inverse of the mod operator in Matlab or
> > simulink?
> >
> > I appreciate any hint.
> >
> > Regards,
> > Mike
>
> I hope you mean the additive inverse, -mod, or the multiplicative
> inverse, 1./mod, because if you want to unmod something I'm afraid
> you're out of luck.
>
> Suppose you want to know if a number is even or odd,
>
> B = mod(A,2);
>
> Now B is either 0 or 1. Do you think there could be any way to get back
> the original number, A, knowing only whether it is even or odd?
>
Actually, there are some things one can do,
although the original question was not specific
enough to answer it.
For starters, is the OP asking for an additive
inverse or a multiplicative inverse? Both CAN be
done, within limits. An additive inverse always
exists in modular arithmetic. Thus 4 is the
additive inverse of 3, mod 7.
A multiplicative inverse is more difficult, since it
does not always exist. Here we are looking for
x such that mod(x*A,p) == 1. You can find a
tool for the modular (multiplicative) inverse in
my vpi toolbox, called minv.m. It uses the
extended Euclidean algorithm.
>> a = 35;
>> p = vpi(2^31 - 1);
>> ainv = minv(a,p)
ainv =
1656630242
>> mod(a*ainv,p)
ans =
1
HTH,
John
Mike Malkum
Thank you very much Guys, I really do appreciate your contributions.
My question came from a simulink block uses this equation:
wt = mod(A, 2*pi)
Now, if I know wt and pi, can I get the value of A?
John, if I use your vpi toolbox, can it calculate the value of A?
Many thanks
Doug Schwarz
"Mike Malkum" <e6c...@yahoo.com> wrote:
Not exactly, no. Think about it, suppose wt is 0.1. You can't know if
A is 2*pi + 0.1 or 4*pi + 0.1 or 6*pi + 0.1, etc.
However, if you're trying to unwrap phase and you have a vector of phase
values then take a look at the function unwrap.
Matt Fig
>
> Thank you very much Guys, I really do appreciate your contributions.
> My question came from a simulink block uses this equation:
>
> wt = mod(A, 2*pi)
>
> Now, if I know wt and pi, can I get the value of A?
>
> John, if I use your vpi toolbox, can it calculate the value of A?
>
> Many thanks
I don't think you are THINKING about what the mod function does. Look closely at this example:
A = [1:20]*pi;
wt = mod(A, 2*pi)'
Study wt. Now suppose I randomly chose one element from A, and only showed you the corresponding element of wt. Could you tell me which element I chose from A with certainty?
Mike Malkum
Doug Matt, Thank you very much.
In fact, what I'm trying to find is the phase from the wt vector. (w=2*pi*f, and t is time). unfortunately when used d(unwrap (wt))/dt, I couldn't get the right phase of the signal. That is why I'm trying with the inverse of mod !
Matt, you're right, I didn't think much about the mod operator! with your example, it is now clear that impossible to give you the element from A you chose. But do you have any suggestions about the phase!
Once again, many thanks,
Mike