🚀 go-pugleaf

In article <2468@njitgw.njit.edu> jfa0522@hertz.njit.edu (john f andrews ece) writes:
|>From: jfa0522@hertz.njit.edu (john f andrews ece)
|>Newsgroups: comp.dsp
|>Subject: Binary Data formats
|>References: 
|>Sender: 
|>Followup-To: 
|>Distribution: world
|>Organization: New Jersey Institute of Technology, Newark, N.J.
|>Keywords: 
|>
|>
|>
...
|>I am faced with the uncertain task of decyphering the data files created by
|>an instrument with missing documentation. The instrument stores the data to
|>DOS-compatible diskettes, for "later retrieval and analysis". Problem is the
|>data format is unknown and the manufacturer has been unable to provide details
|>(or unwilling...I am third party here... my thesis may depend on accessing
|>the data).
|>
|>Preliminary investigation by someone else has reveals the data is in a 
|>binary format. This individual questions my potential ability to do any
|>better than he has with the task over the past year, and has asked what I
|>know about binary data formats. Presumably if I know less than he, he will
|>not share the task with me.
|>
|>What I do know is the basic BCD and packed-BCD formats from my 360/370
|>assembly course.... so what I am looking for his some direction for learning
|>what the possibilities are for this type of data. Any ideas? Every suggestion
|>welcome and appreciated.
|>

From what I know about Biomed data acquistion gear, it is likely the *DATA*
portion of the information is stored in integer binary.  It could be one's
complement, but more likely two's complement.  A rare alternative is magnitude
and sign.

If you knew what the hardware and operating system were, you would be way ahead.
EMG signals are typically bandlimited and/or filtered at a few hundred cycles
per second  ---  and a BASIC program may have been used for acquisition.  In
this case, the format may well be floating point instead of integer.  The
problem is that there are few standards for PC BASIC floating point:  the
mantissas and exponents vary in bit size and offset.

Have you scanned the diskette with an octal dump program?  Can you see ASCII
characters anywhere?  Important because once you psych out the *DATA* format
you have the challenge of psyching out the *FILE* format, so that you can
tell where one file ends and the next begins  ---  and separate out any
extraneous file header information (if it exists within the file and not just
in a directory at the beginning of the diskette.)

In similar previous instances, I've found that a few phone calls to the
manufacturer (in Norway) can lead to identification of who put the product
together in the first place.  It may be just relabelled or put together in
a mountain retreat by elves  ---  and once you get the phone number of the
person who *really* built it, you will have your answer!

P.S.  Based on my 3 decades of involvement with biomedical graduate education,
I'd suspect you are getting a funny deal from whoever holds this data.  It
would help your career if you got into a project with *real* *usable* data
and a sound scientific hypothesis.  But, then again, I've been wrong more than
once ...  !

Rob Lake
BP Research
rbl@BP.COM

|>In case it makes any difference, the Instrument is a DANTEC electromyography
|>instrument, out of Norway, I think.
|>
|>Thanks for ay and all suggestions.
|>
|>
|>-----------------------------------------------------------------------------
|>john f andrews                SYSOP           The Biomedical Engineering BBS
|>    24 hrs                300/1200/2400               (201) 596-5679
|>-----------------------------------------------------------------------------
|>INTERNET jfa0522@hertz.njit.edu    LabRat@faraday.njit.edu    CIS 73710,2600
|>-----------------------------------------------------------------------------
|>
|>
|>
|>
|>
|>
|>