u8rw

Translate between raw binary data and MATLAB matrices.

Description

U8READ and U8WRITE are similar to the built-in functions FREAD and FWRITE except that they transfer data from and to a MATLAB uint8 matrix instead of a file.

A format (precision) argument controls the translation between MATLAB variables and the bytes stored in the uint8 matrix. Mixed precisions within a single call are supported. Big or little endian byte ordering can be specified.

These files are useful for formatting data for external devices (e.g. GPIB and serial ports) and for translating complex binary files (e.g. digital maps).

MEX-files for MacOS, MacOS X and PCWin are included plus source.

• MATLAB release: MATLAB 5.2 (R10)

[Jim Hunziker]

This was just what I was looking for when reading an array of big endian doubles from a DLL in Windows. It worked great!

[Tim Largy]

This tool performs misaligned memory accesses. Some systems allow this (e.g. Windows on x86) but other systems don?t; instead the program will crash. To get around this problem you?ll have to tell your compiler and linker to allow misaligned memory accesses. For example, when using the Sun C compiler on a modern SPARC-based system you must use the -xmemalign flag. Consult your compiler documentation then modify your mexopts.sh file to include the proper compiler and linker flags.

I applaud the author of this tool for including a test program. MATLAB has changed since this tool was written and the test program contains MATLAB calls that don?t exist in MATLAB 7. Fortunately they are related to displaying graphs and can be commented out.

[William Slater]

This is almost exactly what I was looking for. Two suggestions would have made it perfect.

1) Support to read VAX F_float, D_float, and G_float
2) Ability to write the decoded bytes into a structure to preserve data types of embedded data fields

Otherwise, a very beneficial piece of code.