De Grafe Silver UC/F

There’s a trove of documents here and here (also available here) which might provide some clarification. (Does not include a detailed hardware architecture.)

One of the brochures indicates that the main graphics memory has ports to 7 supplementary processors, of various kinds:

1 Like

(Looking at the block diagram) Wow, quite a beast…

Funny thing is that most of the scanned documents on the echosiences site came from me. So I also tried to post a couple of them on my site. Still have to do the brochures though. So many things to do and so little time. Currently working on the PC interface. Borrowed an old pentium system with a ISA slot and finally got dos (first FreeDOS now MSdos6.22) installed on it.
The system works and it detects a silver card. Unfortunately the Silver still tells me “card not detected”.
So it is quite a puzzle where it goes wrong.

1 Like

So, I’m back again.
I have been working on and off with the silver and other equipment the last weeks to get the PC-AT function working and basically get a picture from the silver saved on a pc.
in short, I had to get a pc with a 8bit ISA slot, install dos 6.22 on it and go through all my old software to find the right disk(image) that I saved all those years ago.
Then I had to order new 60pin flat cable and connectors since the old seemed to give errors (that was not the case, but hey).

I found the custom made software of our company, but better yet, I fount the original executable that was sold with the Silver and I got that last one working (the other software keeps giving me a link error). Via some other Silver user I have had contact with the original developer and he even does not have this anymore.

To get further with documenting this I now have a saved picture file and the screen prints of the software and I want to write something for that on the earlier mentioned webpage.

So i was wondering. is there someone here that has some graphical (file) knowledge that could look at the file and write something about it. just a short text to describe the technical graphical specs of this picture file.

2 Likes

What about posting the file here in the hope of some broader collaboration on this?

Yeah, no problem. I’ll try to place it here. Wordpress won’t accept the file since it has no idea what it is :slight_smile: . The upload function here also won’t accept it.
So back to the good old FTP it to a folder on the site and copy-past a link here.
Link-to-picture-file
The picture itself is a really artistic (hmmmm) version of the word “Silver”. Sorry, I’m not really an artist :face_with_diagonal_mouth:
IMG_3803

However, it would probably do as a signage for an artsy coffee shop… :slight_smile:

1 Like

Regarding the image, this sems to be just a plain bitmap of RGB channels stored sequentially.
There seems to be nothing specific about this (there is no magic number indicating a file format, header or meta data, or the like). It even does not appear to have any dimensions encoded (unless I’m missing something). I guess, it’s always a full screen?

I first thought this may be 64-bit color vectors, but this is more likely one complete (8-bit) channel stored after the other. (E.g., in the first third of the data, there are many ajecent bytes with the same value, some of which seem to correspond to red values in the picture.)

For anyone willing to participate:

The file length is 216722 (0x34E92) bytes.
“Natural” color encodings of TV, I can think of, are

  • 8-bit Studio R’G’B’
  • Y’PbPr in 8-bit, 10-bit and 12-bit

Assuming 8-bit channels, this is 72240 x 3 + 2 extra bytes, which would suggest an image format of 210 x 344, 258 x 280, or 301 x 240 pixels (or, rather PELs).

Any ideas?

The netpbm tools may well be enough - see here for example. (I was a great netpbm fan before I eventually found and gradually moved over to Imagemagick - that can probably do the job too.)

So it seems to be nothing too obvious (as anything like horizontal lines of color vectors or channels stored sequentially as arrays of lines). Maybe something line oriented?

Hmm, I’m seeing 210496

Even more hmm: The hex dump (via BBEdit) gives me 216722 bytes (0x000000 … 0x034E91), while the OS reports 210496, the same number you have.

So all my efforts are futile.

Nice info. Thanks.
I also found a small conversion program from those days called pic2pcx.
When I run that on the .pic file it (duh) creates a .PCX file with the output text:

Wegschrijven silver2.pcx (368x268 x 256 ) 8863 bytes.

(Where “Wegschrijven” off course means “Writing”. It was a dutch company)

So the .PCX file that was written is (I presume) resolution of 368x368 in 256 colors.

Link to the PCX file

EDIT:
I remembered that “in the days” we used a command line image converter called alchemy, and yes this was also on my disks. Alchemy cannot read the .PIC file, but it can read and convert the .PCX file.
This also tells that the Width x Height = 368 x 268 and the number of colors is 256 (and a RAW size of 105248 and actual size of 9760 )

BTW, a new hex dump (using another method), correctly gives me 210496 bytes (0 … 0x3363F).
Please ignore any previous comments on this file from side, as these were based on false data (no idea, how a simple hex dump could have failed).

Found some interesting art (see below) while trying to render individual channels at various parameters, but not what we’re looking for.

Having a look at the structure, there may be blocks involved.

That PCX file was accepted by ImageMagick’s convert

convert SILVER2.PCX SILVER2.png

…which reproduces your original…

image

… so it must be the well-known PCX:

Files of type .pic could be a number of different raster formats back in the day… I remember running into them. But that linked file didn’t have any header or other indication of the actual format, as far as I can tell, so I can’t understand how that pic2pcx tool could possibly convert it to anything. Are you sure that the linked .pic file is the actual one converted by pic2pcx? I mean, it’s probably not difficult to fiddle around with that file a little and convert it to something else, but it would be necessary to know or guess the dimensions, at least.

Sorry, pretty busy at the moment. So fiddling around has to fit into the times in-between. But I am still working on it, and still am very happy with the comments .

The .pic file mentioned is the actual one saved from the silver with the original silver software on the pc. The pcx file is the output of the pic2pix.exe software when used with the original .pic file.

@JustSomeOne So this is obviously solved? And the 4 b/w pics of NoLand was a try to read the pic file?
@Tor. I don’t know that tool. Some tools use the file extension for identification or maybe check all formats they know.

There are several early formats without header (and without extension). There are special binary viewers like VELES that are handy. And even if you don’t know the dimensions or other details like color depth, you can try.
GIMP can open raw images (rename to .data). Then you can fiddle. I usually first try 2, 4 and 8bit grayscale. I manged to read most images from the rare Edinburgh IFF format (with and without headers), I posted on YouTube and explained it.
There are special databases on the web with most file formats. Also many open source tools. Some formats are rare. Or if you have a partial or renamed file.