i.tape.other extraction question

[Cleavy McKnight]

I have used i.tape.other to read some imagery files.  Since I don't have a
tape drive, I read them from disk files, which works fine.  My question is
about the SIZE of the created imagery files.

The original data is 4 bands of Landsat MSS data, 2340 lines by 3469 pixels
(or samples) per line.  The resulting number of rows and columns in the new
GRASS imagery file SHOULD be

2340 x 3469 = 8117460 total samples = 8117460 bytes of data per band

What I get as output from i.tape.other, however, are files of these sizes:

-rw-r--r--  1 grass40   4933175 Nov  3 17:39 congo_mss_a.1
-rw-r--r--  1 grass40   6192527 Nov  3 17:40 congo_mss_a.2
-rw-r--r--  1 grass40   7874113 Nov  3 17:42 congo_mss_a.3
-rw-r--r--  1 grass40   7119734 Nov  3 17:43 congo_mss_a.4
-rw-r--r--  1 grass40   8129165 Nov  3 17:50 congo_mss_a.754

Question 1:  Why isn't each file 8117460 bytes in size?

Question 2:  Even more interestingly, why aren't the files AT LEAST all the
             SAME size, since they contain identical amounts of data and were
             obtained from the same source file???

             The file named "congo_mss_a.754" is a 3-band color composite
             image, which SHOULD be 3 x 8117460 bytes = 24352380 bytes in size.

Question 3:  How are these files formatted?  When I try to read them (the
             OUTPUT files from i.tape.other) with another software package
             (IDIMS), reading them as lines and pixels (= rows & columns)
             of data, they look like garbage.  Do they contain header
             information, or what?

Thanks for any help you can provide!

---Cleavy L. McKnight
   Department of Geology
   Baylor University

   cleavy2 at earthlogic.baylor.edu