MATLAB: 3.4gb .raw file in MATLAB

raw filesxyz

Hello

I'm trying to read in a 3D .raw file into MATLAB. The file was obtained from Volume Graphics studio max, a licensed program used for CT scans. We are only interested in the middle 2D plane of data, or in other words the middle 2D array. This may be a horizontal or vertical slice. There is no need to read in all of the data.

Is there a way to only read in one column of the data and then use that to find the middle value? The size is N x N x N but we do not know what N is.

This is all information I have. Any suggestions?

Best Answer

Unfortunately, Volume Graphics does not choose to put its manuals online. Fortunately someone else put one of their manuals online.

Page 250 of http://www.esrf.eu/files/live/sites/www/files/UsersAndScience/Experiments/Imaging/ID19/microtomography/vgstudiomax.pdf describes RAW files in section 5.2.2, indicating the data order has x vary most quickly, then y, then z. There is no indication of whether there is a file header; I would expect there to be one.

You would fopen() the file and read the header.

Then for a horizontal slice, you would fseek() forward by N * N * (bytes per element) * (zplane - 1), after which you would fread(fid, [N, N], PRECISION) where PRECISION is '*uint8' or '*uint16' or as required for your data.

For a vertical slice, you would fseek() forward by N * (bytes per element) * (yplane - 1), after which you would fread(fid, [1, N], PRECISION) and then you would fseek forward by N * (N-1) * (bytes per element) and fread() again, and keep repeating that until the entire slice was read.

If the slice is in the other direction, I would have to work it out but I suspect you could make use of the "skip" option.

I suspect, though, that what might be easiest is to memmapfile() and then index into it.

Related Solutions

MATLAB: Fread: how to read n bytes from bin files

Is it correct? The easiest way for you to find out is to try.

 fid = fopen('acc-angle.bin', 'rb')

There's no 'b' option for fopen permissions. 'r' is all that is needed for binary.

Otherwise, your code looks ok to me as long as your file has little endian encoding. A faster option would be to read all records at once using the skip parameter of fread, something like:

fid = fopen('acc-angle.bin', 'r');
time = fread(fid, Inf, 'int64', 12);
fseek(fid, 'bof', 8);
ax = fread(fid, Inf, 'single', 16);
fseek(fid, 'bof', 12);
ay = fread(fid, Inf, 'single', 16);
fseek(fid, 'bof', 16);
az = fread(fid, Inf, 'single', 16);
fclose(fid);

MATLAB: How to read the fits file content from gz file without extracting them.

There are some useful zip utilities in the File Exchange

https://www.mathworks.com/matlabcentral/fileexchange/56619-zipcontentsinfo

https://www.mathworks.com/matlabcentral/fileexchange/71434-zmat

https://www.mathworks.com/matlabcentral/fileexchange/6055-unzip-file-s

However,

"how can i get "fits" content from gz file without unzipping them"

With the above utilities, especially zmat, you could unzip into memory. However, it would be the entire file that was unzipped. It might happen that the file itself is just the result of fwrite() of a bunch of binary data that is all the same datatype; in such cases then a typecast() later you would have your array of data. But much more common is that the file itself contains headers and possibly various encoded data, and would have to be interpreted by an appropriate routine in order to extract the data. The appropriate routines probably are not expecting to process from memory as a source.

Best Answer

Related Solutions

MATLAB: Fread: how to read n bytes from bin files

MATLAB: How to read the fits file content from gz file without extracting them.

Related Question