Read a binary file of format float

Hello, I have a binary file that holds a content of float array that wa written in c#. I want to read it into matlab to view that float array. I think I need to use the next command but it seems to fail:

myfloat=fread(fID,1,'float');

Answer 1

[Salaheddin Hosseinzadeh]

Hi Alon,

As a C# programmer you know that "it fails" is not enough!

have a look at fopen please

BTW, sometime when I want to generate a file specially on drive C, I have to open MATLAB as an administrator! You may wanna to this as well.

[Alon]

It shows 0 elements in the array

[Alon]

Hello, I don't have much knowledge with matlab or c#.. Do I use this command right?

myFloatArray=fread(fid,1,'float',0,'l');

[Salaheddin Hosseinzadeh]

Dear Alon,

Just answer James questions.

When you open the file using fopen, do you get a valid File ID?

A valid file ID is an integer.

fileID = fopen('myfile.xyz')

Most importantly, pleaseee post the error message

[Alon]

I do it that way:

fid = fopen('myRecordBin.bin', 'r');
fread(fid,inf,'float',4,'l');
fclose(fid);

Now, I can read it and matlab's workspace show 69222 elements and in the vector it shows: 69222x1 double.

But my original float array(and not double) in c# is twice: 138444 elements.

Answer 2

[James Tursa]

Did you open the file first to obtain a valid fID? Are you getting an error message? (If so, please post the entire message). Are you getting garbage for myfloat? (Indicating possible endian problem, or maybe header data not accounted for, or data is not really float, etc.)

[Alon]

Hello I know that I need to at liite endian format. But what do I have to write in the skip? 0?

[James Tursa]

Do you have the C# code that wrote the file? If so, please post the relevant parts that did the writing.

[Alon]

Hello, here is the c# code:

float[] samples32array = samples32.ToArray();

using(BinaryWriter writer = new BinaryWriter(File.Open("C:/Users/alon/Desktop/myRecordings/myRecordBin.bin", FileMode.Create)))
          {
              foreach (var value in samples32array)
              {
                  writer.Write(value);
              }
          }

[Alon]

Now, I can read it and matlab's workspace show 69222 elements and in the vector it shows: 69222x1 double.

But my original float array(and not double) in c# is twice: 138444 elements.

[James Tursa]

Are the values correct for the elements you did read? If so, are you sure there are 138444 floats in the file? What is the file size? To keep the values read in from fread as single use '*float' instead of 'float'.

[Alon]

Hello, the values are like in c# but placed in wrong index . Size of file is: 553772 bytes. ==> 138443 floats. I've tried to use *float but what does it give me?

[James Tursa]

"... placed in wrong index ..."

Please elaborate on the above phrase. Do the numbers match exactly but in the wrong index? If so, give an example of the first few numbers on the C# side and their locations on the MATLAB side to detail explicitly what the indexing problem is you are seeing.

Using '*float' vs 'float' should not have changed any of the values, so your last comment doesn't make sense to me.

'float' --> Read the numbers as float and convert to duoble

'*float' --> Read the numbers as float and keep them that way (i.e., single)

[Alon]

[Alon]

I'm sorry it didn't change the values but still they are placed wrong.

[Alon]

Hello! Thanks for you pleasing answers. I got my problem, here we go:

fid = fopen('myRecordBin.bin', 'r');
myarray=fread(fid,inf,'*float',0,'l');
fclose(fid);