calling a matlab function with scalar type

I am collecting data from a Simulink model into an array so I can manipulate it using a custom function in Matlab. The data is an array of dimensions 1601 x 1, single precision data. Now, I want to operate on one element of this data with my own function which is written in C and converted into a MEX function.

The function works fine with data entered like this:

x = 0.2345
myFunc(x)

...yet when I do this:

x = arr(100)
myFunc(x)

I get the error:

Error using myFunc
Input multiplier must be a scalar.

...however this

isscalar(x)

...yields this...

ans =
  logical
  1

...so is 'x' a scalar or not?

If not, how can I make it a scalar so I can call it with my function? Thanks.

[Geoff Hayes]

Richard - can you attach your myFunc so that we can see what it is doing? Perhaps the error message, Input multiplier must be a scalar. is incorrectly being written to the console and it should be some other error message instead.

Also, what is x equal to once you assign it as

 x = arr(100)

[Honglei Chen]

It would be helpful if you could share what's the result when you do

whos arr x

[Richard Poley]

Geoff, Thanks for your reply. The MEX function is attached inside the zip. It simply computes log2 with the input between 0 and 1. When I reference x as you wrote, I see:

x =

single

    0.0941

[Richard Poley]

Honglei, Thanks.

The result is as follows:

 Name           Size            Bytes  Class     Attributes

arr         1601x1              6404  single              
x              1x1                 4  single

[Geoff Hayes]

Richard - can you attach the C code instead? I would like to see that as opposed to the compiled Mex function. (Should be able to just attach the code rather than a zip file, even if you have to rename it with a txt extension.)

[Richard Poley]

Sorry, Geoff. It's proprietary so I'll need to check.

I'll post back when I know.

Answer 1

[Geoff Hayes]

No worries, Richard. One other thing you may want to consider trying is to cast your x to a double. The example

 x = 0.2345
 class(x)

shows the x is a double and since this works (as expected) then you may want to try that for the other. For example,

 x = double(arr(100));
 myFunc(x)

Now, you will be passing in the same data type. The error message might be misleading and so we see Input multiplier must be a scalar rather than the "true" error message which might be "Input multiplier must be a double rather than a single". This is just a guess though! :)

[Richard Poley]

Geoff. That looks like it works for me. Thanks very much for your help.