1. ## Binary file grouping

I have a binary computer file that contains a bunch of Datablocks that are raw image files. The images can be put together to form a larger image. I need to figure out the positional information for the individual Datablocks so that I can create a program to automatically put them together.

In the Datablock Header, there is information for X and Y size, as well as IDs and groups, or something. What I am looking for is grouping information as well as individual ID information. The most I have assertained thus far is that they are grouped in multiples of 14* or 16*. aka (14*14=196, 16*16=256), however I have not found a fool proof way of grouping them. Even my above observations fall apart towards the middle/end of the list.

I have compiled these Datablock Headers into an HTML file and MS Excel file and have done basic analysis on them. I converted 2 byte blocks to Shorts and 4 byte blocks to integers. I then performed a MIN, MAX, and UNIQUEness calc on them all.

Here is the data:
----------
HTML Format (20MB)

Office 2003 Format (25MB)

2. Originally Posted by Jerick
I have a binary computer file that contains a bunch of Datablocks that are raw image files. The images can be put together to form a larger image. I need to figure out the positional information for the individual Datablocks so that I can create a program to automatically put them together.

In the Datablock Header, there is information for X and Y size, as well as IDs and groups, or something. What I am looking for is grouping information as well as individual ID information. The most I have assertained thus far is that they are grouped in multiples of 14* or 16*. aka (14*14=196, 16*16=256), however I have not found a fool proof way of grouping them. Even my above observations fall apart towards the middle/end of the list.

I have compiled these Datablock Headers into an HTML file and MS Excel file and have done basic analysis on them. I converted 2 byte blocks to Shorts and 4 byte blocks to integers. I then performed a MIN, MAX, and UNIQUEness calc on them all.

Here is the data:
----------
HTML Format (20MB)

Office 2003 Format (25MB)
If you know the file type, you could do an internet search to find the file format -- which would give the the parameters you need.

My guess is that you have a discrete cosine transform.

3. Originally Posted by aidan
If you know the file type, you could do an internet search to find the file format -- which would give the the parameters you need.

My guess is that you have a discrete cosine transform.
That's what I would recomend, in fact if it is a standard file type there should be data in the headers saying what the file type is (file extension might also be a bit of a clue).

If it is something like a jpg or png not only will you find the file format on line but code to read and write such files.

CB

4. I'm pretty sure they're a proprietary raw image format. You can rename the data files as a .raw and load them into Photoshop CS4 as 128x128 with a 26 byte header.

What I'm trying to do though, is find a pattern in the datablock headers (the .xls and .htm files) that will let me create a program to automatically stitch all of the images together, since there's about 21,000 of them.