• Please review our updated Terms and Rules here

Vintage dos/windows software to create .IMG floppy images?

RodCastler

RodCastler

Member
Joined
Oct 25, 2024
Messages
11
The title "Software to RESTORE .IMG floppy images" would be more suitable for this...

Hello everyone,
I have been giving this battle for days now, and I came to request for help: A file named DOC.IMG I created back in Dec/1995 I can't seem to open.

Facts I have gathered so far:
  • The file was created on an IBM Aptiva 486 box.
  • The file header begins with "DF_IMAGE__" then lots of garbled information.
  • I am not sure of how much compression there is because there are also tons of plain, readable, consecutive text from MSOffice files (Word and Excel). This leads me to think that complete documents are within this very same image, and are the ones I want to restore. Again, too much plain text to be a fully compressed image.
  • The file size is 1.478.656 bytes, which makes me believe it's a floppy disk image, although the IMG file itself won't fit in a floppy disk.
  • The 6th line in the file says "Non System Diskette" then "Microsoft Backup Diskette" then "Replace and press any key when ready".
  • Further below the text "NORTON Ver 1E" can be read as well at different points in the file.
What I have tried, with no positive results (errors such as "image format not recognized", "can't open", etc...)

At first, I believed it was a Microsoft backup file.
  • Aptiva Backup software that came with the machine refuses to open the image. The software doesn't seem to work with *.img files at any point either.
  • Also tried Windows 95 backup software, Windows 3.1 backup, and DOS restore. No avail.
After multiple attempts, I abandoned the Microsoft backup approach and began the 1.44 floppy disk image approach:
  • Virtualbox won't accept the image as a floppy image worth of being mounted.
  • WinImage refuses to open it.
  • Rawrite refuses to open it (either running under MSD DOS 6.22 or IBM Dos 6.3 or Windows 95).
  • Rawrite for Windows won't open it.
  • Norton Utilities 6.01 refuses to open it.
  • HD-COPY.exe refuses to open it.
  • PKZIP for DOS and Windows won't open it.
The file doesn't seem to be corrupt. I found it in a directory filled with other files with no sign of corruption.

Question: Any image floppy imaging software for Windows 95 or Windows 3.1 I should Try and I haven't yet?

Thank you so much
 
Last edited:
@RodCastler welcome to the forums! Your description of the file has me wondering if it's just a bunch of files merged together by a "copy /b" command? Some of the things you mention don't jive with that, but thought I'd bring up the idea.

By copy /b command I mean maybe you ran:
copy /b *.* DOC.IMG
Click to expand...
Which would create a single file, containing all the files in that directory, merged into one.

Although I don't know why you would have haha. Maybe you were just experimenting :)
 
Thank you @VeryVon for the welcome! I've been around vintage computer hobby for a long time, but never noticed that VCF had forums until now.

Your idea is not that bad. Who knows what was I experimenting with back then. But wouldn't the copy command just stitch files one after another? That wouldn't be consistent with all the garbled text.

Here's the very first portion of the file for your reference:

Code: 
DF_IMAGE__���P(������$������&$[y��������\�>P�u�����fG��Ǡx-O�
�N�A�+�rʴ!k��41�5�:C�z}�7ө�(�8��|�Z1ݎi�gsU&퓙5w�Oͮ'���nQ-}��ƪ.#��%\��C�+ph�ǹ`��'7$�c��o�/�l'�-��%/�Y�/���c_185'>RL(�OkN�5�����\����ގ�L����7e}�2�i�+btl^�tKyKc��c1�>0W�Ny�Y±�{|S�n]�y3&���S�K�lA.�!�e�Z��b;��׺����  �H�q��e9�
AUD<)��'?��uTCD$�r�128 B res. for description�������������������������������������������������������������������������������������������������������;�'`e/.IHC����@�  �����������������������������������3��Ў���|�Ȏؾ�|��i��<�t
������������
Non System Diskette
Microsoft Backup Diskette
Replace and press any key when ready... ������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������@�`���  ��P
�����@`�����! #@%`'�������/�1 3@5��7�9�;�=�?�A �OE`G�I�K�M�O�Q S@U`W�Y�[�]�_�a c@e`g�i�k�m�o�q s@u`w��y�{�}��� �O�`�����������  �  �@ ���� �� �� �� ��
�
 
I'm afraid libdsk is unlikely to help here - I haven't seen any disk image format that begins DF_IMAGE.

If you're lucky the data may be uncompressed with a header at the beginning. There's what looks like a Windows 9x boot sector (that IHC looks very characteristic) so if you extract 1474560 bytes starting 8 bytes before the IHC, you may end up with something readable by WinImage.
 
Thank you John and Chuck. Per John's suggestion, I've been attempting to remove the header, but for some reason the file gets reduced in size to only 1.470.131. I'm using nano on linux to do this. Shall I use a particular tool for manipulating the file?

By the way, I made a mistake in the original post. The correct size of the original file is 1.475.044 and not 1.478.656 as stated earlier.
 
OK.

I believe we are indeed getting somewhere.
I counted 8 characters before the IHC and chopped off all of what was before it, and named the new file "modded DOC.IMG", shown to the left in the image below (1.470.130 in size)

As a control sample, I took a totally unrelated backup test set using the windows backup tool, which creates a filename named compat.002 on a floppy disk.
I then recorded an image of that backup test and named it WIN_BACKUP_SAMPLE.IMG (the one to the right in the image below)

Comparing this two headers, they are VERY similar. Of course the contents that come after the header differ substantially. The files are also different in length. Hence, Winimage will happily open the backup sample which was recently created, but will refuse to open modded DOC.IMG (the original file minus the header) which is slightly shorter.


1730247985690.png
 
Last edited:
RodCastler said:
Thank you John and Chuck. Per John's suggestion, I've been attempting to remove the header, but for some reason the file gets reduced in size to only 1.470.131. I'm using nano on linux to do this. Shall I use a particular tool for manipulating the file?

By the way, I made a mistake in the original post. The correct size of the original file is 1.475.044 and not 1.478.656 as stated earlier.
Click to expand...
I tend to use dd for this. For example, to remove the first 256 bytes from a file, I would use:

dd if=infile of=outfile bs=1 skip=256

When making hex dumps I would suggest

od -A x -t x1 filename

to render the output in bytes rather than words.

If the file you are feeding into winimage is too short, you could pad it to the correct size by creating a block of zeroes:

dd if=/dev/zero of=zeroes.bin bs=1 count=<count>

and then append that to your disk image:

cat shortimage zeroes.bin > diskimage

However if the file without its header is less than 1474560 bytes, there may be some form of compression involved, in which case simply removing the header wouldn't be enough to make it readable as a raw disk image.
 
Thank you so much, John. I found a software named HexCmp that alows me to compare, and I have found very interesting similarities between the IMG file without the header and a regular Microsoft Backup file. I'll use your suggestions with dd, od and cat commands to fit and adjust some things and see if I can make it to a point where either Winimage or Virtualbox can actually open the img file. That's going to take me a few days probably. wish me luck, and thank you!
 
In your original post you mentioned the string "NORTON Ver 1E" was visible.

Although its possible this is just a part of the data that was on the floppy disk rather than pertinent to the format of the image, it might also be the case that it was a Norton utility that created the image.

You said Norton Utilities 6.0 refused to open it, but maybe a different version might? You can experiment with these: https://winworldpc.com/product/norton-utilities/
Given "Ver 1E" and you mentioned this was created in '95 I might try "1.0 (95) for Win".

I tried to find a Norton tool that might have the initials "DF" but came up empty handed. I know there was "Norton Disk Doctor" but maybe it used "DF" for floppy disk images? Just a guess.
 
RodCastler said:
Question: Any image floppy imaging software for Windows 95 or Windows 3.1 I should Try and I haven't yet?
Click to expand...

I've used ANADISK to create images. Sounds like you want to go the opposite direction? Maybe there is something in ANADISK to do what you want to do? It's an old DOS program created by Sydex.

Roger
 
Divarin said:
In your original post you mentioned the string "NORTON Ver 1E" was visible.

Although its possible this is just a part of the data that was on the floppy disk rather than pertinent to the format of the image, it might also be the case that it was a Norton utility that created the image.
Click to expand...

I thought exactly the same.. but I didn't think of trying other versions. Using Norton for Windows sounds like something I might have done as well, and I haven't tried yet! So I'll give it a shot and will report back.
 
Sydex? Who dat?

What did you use to get this data? I ask because the first line reads:
Code: 
0000000 3beb 2790 6560 2e2f 4849 2043 0102 2001
Which looks like a byte-swapped (big endian) version of a boot sector. That is, it should read:
Code: 
0000000 eb3b 9027 6065 2f2e 4948 4320 0201 0120


If you're using hexdump, you might be best off adding the -C switch.

How about posting the raw binary somewhere ? You might even get it posted here if you ZIP it up first.
 
Last edited:
Chuck(G) said:
Code: 
0000000 3beb 2790 6560 2e2f 4849 2043 0102 2001
Which looks like a byte-swapped (big endian) version of a boot sector. That is, it should read:
Code: 
0000000 eb3b 9027 6065 2f2e 4948 4320 0201 0120
Click to expand...
Good catch. Looks like that dump was made with od -x (dump 16 bit words in hex, using native machine endianness) rather than od -tx1 (dump 8-bit bytes in hex), which would read:

Code: 
0000000 eb 3b 90 27 60 65 2f 2e 49 48 43 20 02 01 01 20
 
So the thing starts off with a boot-sector image. I haven't bothered to warp my mind unraveling the remainder, but this could be something that simply uses a non-conventional disk geometry and is otherwise a disk image.
 
OK.
RodCastler said:
I counted 8 characters before the IHC and chopped off all of what was before it, and named the new file "modded DOC.IMG", shown to the left in the image below (1.470.130 in size)
Click to expand...
How far into the file was that? Exactly 4096 bytes? Because that .IMG file is exactly 4096 bytes longer than a straight binary dump of a 1.44 MB diskette.

Try dd if=infile of=outfile bs=1 skip=4096 and see if that gives you an MS-DOS floppy image. (I'm assuming that hex dump you gave earlier was not the start of your original file but the start of that new file you created with the first part trimmed off, but that file was broken because you loaded and saved in nano instead of trimming with dd.)
 
Chuck(G) said:
Sydex? Who dat?

What did you use to get this data? I ask because the first line reads:
Code: 
0000000 3beb 2790 6560 2e2f 4849 2043 0102 2001
Which looks like a byte-swapped (big endian) version of a boot sector. That is, it should read:
Code: 
0000000 eb3b 9027 6065 2f2e 4948 4320 0201 0120


If you're using hexdump, you might be best off adding the -C switch.

How about posting the raw binary somewhere ? You might even get it posted here if you ZIP it up first.
Click to expand...
I'm a bit of a newbie on these matters. I just went for a "hexdump -n 256 'DOC.IMG' > chuck.txt" with no other parameters / switches.
I'd love to post the entire file and be of better help. the only problem is that the files i'm attempting to rescue may have other people's information and I am working on the careful side.
 
You must log in or register to reply here.
Back
Top