[Python] How to reverse-engineer an undocumented proprietary Raman spectroscopy binary file...

I’m working with several Raman spectroscopy data files that use a proprietary binary format with the extension .jws. The original software that could read these files is no longer available, and there is no public documentation describing the file structure.

I cannot share the actual files due to institutional restrictions, but here is what I know in general terms:

• 
  The files appear to contain a header of unknown length.
  
  
• 
  After the header, there seems to be a large binary block (possibly X–Y spectral data).
  
  
• 
  Some parts of the file look like sequences of 32-bit floats, but I cannot determine which block corresponds to the Raman shift axis and which block represents intensity.
  
  
• 
  There are no obvious markers, offsets, or metadata structures that match any known open formats.
  

My goal is simply to extract the Raman X-Y spectral data so that it can be analyzed with modern tools.

Since the format is undocumented, I’m looking for general advice:

What is the recommended workflow for reverse-engineering an unknown proprietary spectroscopy binary format?
Are there standard techniques, tools, or heuristics commonly used when dealing with legacy instrumental data?

Possible approaches I am considering:

• 
  Inspecting the binary with a hex editor
  
  
• 
  Searching for float32 / float64 regions
  
  
• 
  Identifying repeating patterns that could indicate arrays
  
  
• 
  Testing little-endian vs big-endian
  
  
• 
  Looking for integers that might be offsets or block sizes
  
  
• 
  Comparing multiple files to find structural similarities
  

Before I spend too much time manually decoding the files, I’d appreciate any general guidance or best practices from people who have reverse-engineered similar scientific data formats.

Thanks in advance.

Continue reading...