ePic

Posted on by Thor Leave a comment

Image compression has been around for awhile. It seems everyone took a crack at making better algorithms to improve quality and size. Some chose to invent new ways and others chose to use existing methods but with their own flare. Kodak tried this with their PhotoCD, but there was a couple other photo processing options that popped up in 90’s. One was Seattle FilmWorks and another was Konica PC PictureShow. Both of which used “proprietary” formats to deliver developed film on disk.

Seattle FilmWorks later called PhotoWorks, used an image format with the extension SFW and was based on BMP and JPG, but with their own twist. The same goes for the format used by Konica’s PC PictureShow.

Konica PC PictureShow Disk

If you took your film in to be developed at one of Konica’s photo labs, you could could have those images put on a diskette or later a CD-R. The disks came with software to view your photos called PC PhotoShow. The images stored on disk where in another proprietary format with the extension KQP. The KQP format was actually licensed from another company called Pegasus Imaging Corporation, later known as Accusoft. They developed their own way to compress a JPEG file which they called an ePic. An SDK called PICTools was offered for many years, but seems not to be available anymore.

ePIC (Proprietary)
  • Supports PIC format compression, replacing the JPEG Huffman encoder with the proprietary ELS entropy encoder for 15% more compression.
  • Can be losslessly converted back to JPEG format using Op_RORE.

A search on the internet for Konica KQP shows quite a few people over the years wondering what to do with their old disks and converting the old format to JPG, only to find a lack of information and available tools to do so. One such person used python to edit the file and making the file renderable as a JPG. While the method worked well for their KQP files, it might not work for all of them. Let’s look closer and understand why.

hexdump -C Sample.PIC | head
00000000  42 4d 00 00 00 00 00 00  00 00 42 04 00 00 44 00  |BM........B...D.|
00000010  00 00 34 08 00 00 24 fa  ff ff 01 00 18 00 4a 50  |..4...$.......JP|
00000020  45 47 00 00 00 00 00 00  00 00 00 00 00 00 fc 00  |EG..............|
00000030  00 00 ec 00 00 00 2c 00  00 00 18 00 00 00 00 00  |......,.........|
00000040  00 00 02 00 00 00 08 00  00 00 01 00 00 00 01 00  |................|
00000050  00 00 60 00 00 00 00 00  60 00 00 60 00 00 00 00  |..`.....`..`....|
00000060  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|

At first glance the file appears to be a Bitmap (BMP), and it does have a Bitmap header claiming to have JPEG compression, but if we look a little further into the file.

identify -verbose Sample.PIC   
identify: length and filesize do not match `Sample.PIC' @ error/bmp.c/ReadBMPImage/950.
identify: unrecognized compression `Sample.PIC' @ error/bmp.c/ReadBMPImage/1019.

hexdump -C Sample.PIC      
00000000  42 4d 00 00 00 00 00 00  00 00 42 04 00 00 44 00  |BM........B...D.|
00000010  00 00 34 08 00 00 24 fa  ff ff 01 00 18 00 4a 50  |..4...$.......JP|
00000020  45 47 00 00 00 00 00 00  00 00 00 00 00 00 fc 00  |EG..............|
00000030  00 00 ec 00 00 00 2c 00  00 00 18 00 00 00 00 00  |......,.........|
00000040  00 00 02 00 00 00 08 00  00 00 01 00 00 00 01 00  |................|
00000050  00 00 60 00 00 00 00 00  60 00 00 60 00 00 00 00  |..`.....`..`....|
00000060  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000400  00 00 60 00 00 00 00 00  60 00 00 60 00 00 00 00  |..`.....`..`....|
00000410  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000440  00 00 ff d8 ff e0 00 10  4a 46 49 46 00 01 02 02  |........JFIF....|
00000450  00 00 00 00 00 00 ff e1  00 0a 50 49 43 00 01 19  |..........PIC...|
00000460  1e 01 ff c0 00 11 08 05  dc 08 34 03 01 11 00 02  |..........4.....|

We find a JPG marker, in fact almost the whole jpg file is included, except the quantization tables for luminance and chrominance which are needed to properly display the image. This is the area the Pegasus company thought they could encode better to further compress the image. Their method was to use a new algorithm called ELS (Entropy Logarithmic-Scale). This new method was used by the PICTools software to make a Pegasus PIC file while Konica used it for their KQP format. They are identical. By choosing the luminance and chrominance values during compression, you could make a highly compressed image, but required specific software to render.

Pegasus also made use of a special custom APP marker (PIC) within the JPEG structure of the PIC/KQP and also any JPG compressed using their software. This marker which takes up around 8 bytes holds the luminance and chrominance values. Take the above sample for instance, it is compressing the image with a Luminance of 25 and a Chrominance of 30, these are integer values and in hex they would be “19” and “1E” respectively. 

hexdump -C Sample.PIC      
00000440  00 00 ff d8 ff e0 00 10  4a 46 49 46 00 01 02 02  |........JFIF....|
00000450  00 00 00 00 00 00 ff e1  00 0a 50 49 43 00 01 19  |..........PIC...|
00000460  1e 01 ff c0 00 11 08 05  dc 08 34 03 01 11 00 02  |..........4.....|
00000470  11 01 03 11 01 ff c4 00  51 00 01 00 03 01 00 00  |........Q.......|

So in theory one could strip out any part of the file before the JPG beginning of file magic bytes (FF D8 FF E0), locate the APP marker, use the values to generate the two quantization tables, insert them in the appropriate spot and save out a JPG file.

This may be the case for the first few versions of the ePic format, but later versions got more complicated. It seems a “PIC2” version replaced the earlier versions and this format is a little more complicated. 

hexdump -C Sample.KQP | head
00000000  50 49 43 32 01 08 00 00  00 64 00 01 00 b9 3e 00  |PIC2.....d....>.|
00000010  00 05 08 00 00 00 4a 50  47 45 03 00 00 00 16 24  |......JPGE.....$|
00000020  00 00 00 43 6f 6d 70 72  65 73 73 69 6f 6e 20 62  |...Compression b|
00000030  79 20 50 65 67 61 73 75  73 20 49 6d 61 67 69 6e  |y Pegasus Imagin|
00000040  67 20 43 6f 72 70 2e 06  68 3e 00 00 ff d8 ff e0  |g Corp..h>......|
00000050  00 10 4a 46 49 46 00 01  01 00 00 01 00 01 00 00  |..JFIF..........|
00000060  ff e1 00 16 50 49 43 00  03 00 00 01 00 00 00 00  |....PIC.........|
00000070  00 00 00 00 00 00 00 00  ff db 00 84 00 0f 0a 0a  |................|
00000080  0a 0a 06 0f 0a 0a 0a 0f  0f 0f 0f 14 1e 14 14 14  |................|
00000090  14 14 28 1e 1e 19 1e 2d  28 32 32 2d 28 2d 2d 32  |..(....-(22-(--2|

Instead of the Bitmap (BMP) header, a proprietary PIC2 header is used, still containing a JPG in the JFIF format along with a the PIC APP marker, but encoded in a way that the simple method of adding a quantization table may not work. With the original format the JPG and the PIC/KQP were approximately the same size, this new version significantly reduces the size of the PIC/KQP in comparison with the JPG.

The ELS compression technology used in the ePic format seems to be patented by Pegasus and Accusoft, but is not entirely hidden as the libavcodec library includes a ELS decoder. Might be a fun project to use the code to decode the PIC/KQP formats fully.

In the meantime, a signature identifying the two versions should be added to PRONOM. Check out my proposal on my GitHub. If you need to convert your KQP or PIC files back to JPG here are a few links:

Konica PC PictureShow Version 4 (PIC2)

Accusoft PICTools Apollo Demo (Windows 7 Compatible)

Konica PC PictureShow for Macintosh

Canvas

Posted on by Thor Leave a comment

When it comes to design software there were many options over the years, many being released with a lot of hype and others disappearing not long after they released. There are few which lasted long enough to not be gobbled up by big names such as Adobe. One of those is Canvas by Deneba Systems.

First released in 1987, it is still available over at Canvas GFX. It’s amazing it was never bought by one of the big names, Adobe, Corel, Aldus, etc and remained under Deneba Systems until 2003 when it was bought by ACD Systems, but kept the name Deneba Canvas for a time. The later versions were not popular to all, and Mac support was dropped, but the software continued. Awhile back I was looking through a few of my old ZIP disks and found some software my father used in the mid 1980’s. He had a copy of Canvas version 2 for Macintosh. At that time I was more interested in playing games on our family’s Macintosh 128k than using design software.

Over the years I have come across many Canvas documents. With each version released, changes were made to the file format used to store the drawings and artwork. There were many file format changes as well as the extensions used with each version. Some are easily identifiable and others have some confusing structures. Lets look into it.

VersionPlatformExtensionDescription
Canvas 1-3 & artWORKSMacintoshnoneno strong pattern
Canvas 3.5Mac & WindowsCVSSimilar to v1-3
Canvas 5Mac & WindowsCV5CANVAS5 string
Canvas 6-8Mac & WindowsCNVCANVAS6 string
Canvas 9-XMac & WindowsCVXSimilar to 6-8
Canvas DrawMacCVDDifferent than others
Canvas Image FileCVIDAD5PROX

The first three versions of Canvas were Macintosh only and in those early days there was no extension, just a Type / Creator indicating to the Finder how to open them. Deneba Systems used the Creator codes DAD2, DAD5, through DADX.

The first versions are quite frustrating. I have gathered samples from Version 2, 3, 3.5 and artWORKS version 1. Even with numerous samples, there are no patterns I can discern from them. I even reached out to the current CanvasX technical support for answers. They wanted to be helpful, but their answers didn’t offer much help.

With “CVS” or ‘drw2’ for mac, the header contains ranges inside a structure, and other data like if it was compressed. When we see if it’s a valid file we check the ranges. There is no easy way to determine what hex values would be written because of flipping, Intel vs (PPC or 68K). Unfortunately, the research needed to identify the Hex value will require the original code for version 3.5 which we do not have access to easily. Canvas 3.5 code is 16 bit… this would also be an issue.

Let’s take a look at a couple samples:

hexdump -C Canvas2.1-Sample | head
00000000  00 00 03 06 00 00 3d 9c  00 00 00 2a 00 00 00 0a  |......=....*....|
00000010  00 00 00 76 00 00 00 36  00 00 00 2e 00 00 00 1e  |...v...6........|
00000020  00 00 00 12 00 00 00 42  00 00 00 1a 00 00 00 82  |.......B........|
00000030  00 00 00 3c 00 66 00 01  00 00 3d 9c 00 48 00 00  |...<.f....=..H..|
00000040  40 02 90 00 00 00 00 00  00 00 00 00 00 00 00 00  |@...............|
00000050  00 01 00 00 01 00 00 00  00 20 00 40 00 60 00 80  |......... .@.`..|
00000060  00 c0 01 40 01 80 01 c0  02 40 02 80 00 00 00 00  |...@.....@......|
00000070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 05  |................|
00000080  00 00 00 00 00 01 00 10  00 00 00 01 00 03 3f fc  |..............?.|
00000090  80 00 00 00 00 00 00 00  00 07 00 01 00 01 00 0b  |................|

hexdump -C Canvas2-s02 | head
00000000  00 00 03 b2 00 00 07 ec  00 00 00 2a 00 00 00 0a  |...........*....|
00000010  00 00 00 76 00 00 00 36  00 00 00 2e 00 00 00 1e  |...v...6........|
00000020  00 00 00 12 00 00 00 42  00 00 00 1a 00 00 00 82  |.......B........|
00000030  00 00 00 3c 00 66 00 01  00 00 07 ec 00 48 00 00  |...<.f.......H..|
00000040  40 02 90 00 00 00 00 00  00 00 00 00 00 00 00 00  |@...............|
00000050  00 01 01 00 01 00 00 00  00 20 00 40 00 60 00 80  |......... .@.`..|
00000060  00 c0 01 40 01 80 01 c0  02 40 02 80 00 00 00 00  |...@.....@......|
00000070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 05  |................|
00000080  00 00 00 00 00 01 00 10  00 00 00 01 00 03 3f fc  |..............?.|
00000090  80 00 00 00 00 00 00 00  00 07 00 01 00 01 00 0b  |................|

hexdump -C Canvas3.04 | head
00000000  00 00 02 5a 00 00 00 1c  00 00 00 2a 00 00 00 0a  |...Z.......*....|
00000010  00 00 00 76 00 00 00 36  00 00 00 2e 00 00 00 1e  |...v...6........|
00000020  00 00 00 12 00 00 00 42  00 00 00 1a 00 00 00 82  |.......B........|
00000030  00 00 00 3c 00 68 00 02  00 00 00 1c 00 48 00 00  |...<.h.......H..|
00000040  40 02 90 00 00 00 00 00  00 00 00 00 00 00 00 00  |@...............|
00000050  00 01 01 00 01 03 00 00  00 20 00 40 00 60 00 80  |......... .@.`..|
00000060  00 c0 01 40 01 80 01 c0  02 40 02 80 00 00 00 00  |...@.....@......|
00000070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000080  00 01 00 00 00 01 00 10  00 00 00 01 00 03 3f fc  |..............?.|
00000090  80 00 00 00 00 00 00 00  00 07 00 01 00 01 00 0b  |................|

hexdump -C Canvas5-3.5-Sample1.CVS | head
00000000  00 00 01 58 00 00 01 30  00 00 00 2a 00 00 00 00  |...X...0...*....|
00000010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000030  00 00 00 00 00 69 00 02  00 00 01 30 00 48 00 00  |.....i.....0.H..|
00000040  40 02 90 00 00 00 00 00  00 00 00 00 00 00 00 00  |@...............|
00000050  00 01 01 01 00 00 00 00  00 20 00 40 00 60 00 80  |......... .@.`..|
00000060  00 c0 01 40 01 80 01 c0  02 40 02 80 00 00 00 00  |...@.....@......|
00000070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000080  00 01 00 00 00 01 00 10  00 00 00 01 00 03 3f fc  |..............?.|
00000090  80 00 00 00 00 00 00 00  00 07 00 01 00 01 00 01  |................|

hexdump -C C3-5-S01.CVS | head
00000000  78 11 00 00 10 00 00 00  2a 00 00 00 0a 00 00 00  |x.......*.......|
00000010  26 00 00 00 26 00 00 00  26 00 00 00 26 00 00 00  |&...&...&...&...|
00000020  96 00 00 00 2a 00 00 00  2e 00 00 00 32 00 00 00  |....*.......2...|
00000030  00 00 00 00 01 6b 01 00  50 14 00 00 28 00 00 00  |.....k..P...(...|
00000040  6e 00 00 00 5b 00 00 00  01 00 04 00 00 00 00 00  |n...[...........|
00000050  e8 13 00 00 12 0b 00 00  12 0b 00 00 00 00 00 00  |................|
00000060  00 00 00 00 00 00 00 00  00 00 80 00 00 80 00 00  |................|
00000070  00 80 80 00 80 00 00 00  80 00 80 00 80 80 00 00  |................|
00000080  c0 c0 c0 00 80 80 80 00  00 00 ff 00 00 ff 00 00  |................|
00000090  00 ff ff 00 ff 00 00 00  ff 00 ff 00 ff ff 00 00  |................|

In the version 2 & 3 samples you can see some patterns, which I thought would allow for proper identification, but looking at more samples I found differences. One pattern I was hopeful might be consistent was the hex values “002000400060008000C00140018001C002400280”, but there are some which don’t match this pattern. If the file is truly compressed, it will be hard to know which values would be consistent among all files. I have over 8,000 samples and have a signature that only excludes around 20, so it will have to do for now.

When we start with Version 5 we get into some more identifiable headers, there is some oddness with some samples. But with an ascii string like “CANVAS5”, it should be easy, right? Not so fast, in version 5 you can compress the file structure. This removes the easily identifiable “CANVAS5” string. But some have a small string at the tail end, but others do not. 

hexdump -C Canvas5-Sample1.CV5 | head
00000000  02 00 00 80 00 00 00 00  00 00 00 4e 96 00 00 4e  |...........N...N|
00000010  96 18 02 00 00 00 0e a8  da 43 41 4e 56 41 53 35  |.........CANVAS5|
00000020  00 01 00 00 00 00 00 05  03 00 00 00 00 00 00 00  |................|
00000030  00 00 00 00 00 21 00 00  00 21 00 00 00 79 00 00  |.....!...!...y..|
00000040  00 03 00 00 01 6b 00 00  00 03 00 00 00 01 ff ff  |.....k..........|
00000050  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|

hexdump -C Canvas5-Sample3-cmp.CV5 | head
00000000  02 00 00 80 00 00 00 00  08 00 00 80 00 00 00 03  |................|
00000010  5c ff ff ff ff 00 00 40  22 00 00 03 50 10 00 89  |\......@"...P...|
00000020  07 60 bd 0f f0 00 00 10  03 04 10 56 00 20 05 00  |.`.........V. ..|
00000030  e0 18 02 10 35 04 30 4e  05 30 72 07 f0 a8 0d a1  |....5.0N.0r.....|
00000040  17 11 81 19 05 50 5c 00  60 0f 00 10 80 02 90 80  |.....P\.`.......|
00000050  03 f0 56 05 50 55 05 b0  75 12 51 29 05 e0 55 05  |..V.PU..u.Q)..U.|

hexdump -C Canvas5-Sample3-cmp.CV5 | tail
00001ff0  00 00 00 01 08 a5 ab c0  00 00 00 00 3f 89 2c 58  |............?.,X|
00002000  00 00 00 00 08 a5 ab 80  00 00 00 00 ff d4 11 e4  |................|
00002010  00 00 00 00 08 a5 ab 90  00 02 3e d8 ff d3 12 cc  |..........>.....|
00002020  00 00 00 00 00 00 00 00  00 02 3e d8 00 01 00 09  |..........>.....|
00002030  00 00 00 00 00 00 00 00  00 00 00 00 08 a5 ab f8  |................|
00002040  00 00 00 00 43 4e 56 35                           |....CNV5|

Canvas 6 uses a new extension, but has a similar structure to the file format. With compression as an option. But some of the compressed files on Windows has a reversed string, “5VNC“. So many Canvas 5 compressed look identical to Canvas 6 compressed, complicating identification.

hexdump -C Canvas6-Sample.CNV | head
00000000  01 00 80 00 00 90 07 cd  07 00 80 00 00 00 80 00  |................|
00000010  00 17 01 00 00 59 f5 0e  00 43 41 4e 56 41 53 36  |.....Y...CANVAS6|
00000020  00 01 00 00 00 00 06 00  00 00 00 00 00 00 00 00  |................|
00000030  00 00 00 00 00 21 7a 00  00 00 7a 00 00 00 03 00  |.....!z...z.....|
00000040  00 00 6e 01 00 00 03 00  00 00 01 00 00 00 ff ff  |..n.............|
00000050  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|

hexdump -C Canvas6-Sample1-c.CNV | head
00000000  01 00 80 00 00 58 ea 2b  00 c2 1d 00 00 d0 09 00  |.....X.+........|
00000010  00 00 00 0f 2e 00 00 0b  07 00 00 09 c4 10 00 01  |................|
00000020  00 00 03 00 20 04 00 70  ff 00 80 05 00 c0 06 06  |.... ..p........|
00000030  50 20 03 00 0f 06 10 6b  00 a0 12 01 00 48 07 20  |P .....k.....H. |
00000040  6d 07 30 40 06 40 11 06  00 0b 05 00 10 00 10 71  |m.0@.@.........q|
00000050  01 40 21 00 00 59 01 00  0f 05 10 00 00 e1 14 00  |.@!..Y..........|

hexdump -C Canvas6-Sample1-c.CNV | tail
000016a0  00 00 00 12 f6 00 00 c0  f0 12 00 3c d0 80 7c 58  |...........<..|X|
000016b0  2f 14 00 00 00 00 00 bc  f4 8d 00 0f 00 00 00 00  |/...............|
000016c0  f1 12 00 7f 00 00 00 f8  2e 14 00 bc f4 8d 00 1c  |................|
000016d0  f2 12 00 04 f3 12 00 fc  d1 80 7c 09 04 00 00 00  |..........|.....|
000016e0  00 00 40 00 f2 12 00 ff  ff ff ff 00 f1 12 00 1c  |..@.............|
000016f0  f1 12 00 bc f4 8d 00 00  00 00 40 35 56 4e 43     |..........@5VNC|

While most have the “CANVAS6” string near the beginning, quite a few are missing the CNV5/5VNC string at the end. Instead, many have the string “%SI-0200” near the end, which I use in my signature suggestion. This structure remained the same from version 6 to 8. 

hexdump -C Canvas8-S01.CNV | head
00000000  02 00 00 80 00 00 12 b8  80 00 00 11 19 00 00 11  |................|
00000010  19 18 02 00 00 00 0e f5  59 43 41 4e 56 41 53 36  |........YCANVAS6|
00000020  00 01 00 00 00 00 00 08  01 00 00 00 00 00 00 00  |................|
00000030  00 00 00 00 00 21 00 00  00 00 00 00 00 00 00 00  |.....!..........|
00000040  00 03 00 00 00 00 00 00  00 03 00 00 00 01 00 00  |................|
00000050  00 01 ff ff ff ff 00 00  00 02 00 00 00 02 00 00  |................|

But…….. There are plenty without these strings, just the “%SI-0200” near the end.

hexdump -C TELEGRPH.CNV | head
00000000  02 00 00 80 00 00 00 00  08 00 00 80 00 00 00 3d  |...............=|
00000010  f2 ff ff ff ff 00 00 75  76 00 00 3d e6 10 00 ff  |.......uv..=....|
00000020  00 00 b3 0d 90 a9 03 b0  8a 07 f0 98 07 60 80 08  |.............`..|
00000030  d0 35 01 c0 58 01 e0 59  04 80 b8 03 90 38 02 f0  |.5..X..Y.....8..|
00000040  e2 00 20 0b 03 70 1d 03  20 36 0f 30 00 01 80 09  |.. ..p.. 6.0....|

hexdump -C TELEGRPH.CNV | tail
00006850  2b 2c f9 ae 30 00 00 00  20 00 00 00 01 00 00 00  |+,..0... .......|
00006860  0f 00 00 00 10 00 00 00  1e 00 00 00 07 00 00 00  |................|
00006870  64 65 6e 65 62 61 00 00  00 00 01 4c 25 53 49 2d  |deneba.....L%SI-|
00006880  30 32 30 30 6d 61 63 00  00 00 00 00 00 00 00 00  |0200mac.........|
00006890  00 00 00 00                                       |....|

In version 9 and forward we have an extension change to CVX, but the format is similar with the “CANVAS6” string, but is a slightly different offset. It is still used with the current version of Canvas X.

hexdump -C Canvas9-Sample1.cvx | head
00000000  00 00 00 00 00 00 00 00  00 00 02 00 00 80 00 07  |................|
00000010  d1 84 d0 00 00 80 00 00  00 80 00 18 02 00 00 00  |................|
00000020  0f b7 ef 43 41 4e 56 41  53 36 00 01 00 00 00 00  |...CANVAS6......|
00000030  00 09 00 00 00 03 34 00  00 00 04 00 00 00 00 00  |......4.........|
00000040  00 00 00 3c 42 45 47 49  4e 5f 50 52 45 56 49 45  |...<BEGIN_PREVIE|
00000050  57 5f 54 41 47 3e 21 00  00 00 75 00 00 00 79 00  |W_TAG>!...u...y.|
00000060  00 00 03 00 00 01 6b 00  00 00 03 00 00 00 01 ff  |......k.........|
00000070  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|

hexdump -C Canvas9-Sample1-compressed.cvx | tail
00004090  00 00 e0 20 00 57 80 00  00 00 00 00 0a 13 00 09  |... .W..........|
000040a0  00 00 04 00 00 00 00 01  00 00 00 00 bf ff e0 80  |................|
000040b0  bf ff e0 40 01 8c 5e 00  02 4a 22 d0 00 00 01 60  |...@..^..J"....`|
000040c0  bf ff e0 40 00 5c 08 18  00 00 00 00 00 0d 84 80  |...@.\..........|
000040d0  43 61 6e 76 61 73 39 2d  53 61 6d 70 6c 65 31 2d  |Canvas9-Sample1-|
000040e0  63 6f 6d 70 72 65 73 73  65 64 2e 63 76 78 00 18  |compressed.cvx..|
000040f0  bf ff e0 70 0a 12 6a a0  02 43 22 b4 00 0c aa 9c  |...p..j..C".....|
00004100  bf ff e0 80 00 00 00 01  00 00 00 00 00 0d 84 80  |................|
00004110  bf ff e0 b0 43 4e 56 35                           |....CNV5|

hexdump -C CanvasX2019-S01.cvx | head
00000000  00 00 00 00 00 00 00 00  00 00 01 00 80 00 00 00  |................|
00000010  6e ab 03 00 80 00 00 00  80 00 00 17 01 00 00 ef  |n...............|
00000020  b7 0f 00 43 41 4e 56 41  53 36 00 01 00 00 00 00  |...CANVAS6......|
00000030  09 00 00 4d 01 00 00 eb  4c 00 00 41 00 00 00 31  |...M....L..A...1|
00000040  52 45 56 03 00 00 00 01  00 00 00 00 00 00 00 00  |REV.............|
00000050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|

This collection of file formats is very hard to make sense of. Some really great consistent patterns on many samples, with lots of exceptions. Super confusing. This software has had a long run, with the latter years staying pretty stagnate in terms of new development. It is worth defining and creating a signature for the consistent patterns, then we can dial in the variants over time?

The signatures I have built miss about 23 files in versions 1-3 out of the ~9000 samples I have and for Canvas 5, only some of the compressed files are currently not identified. But so far all my CNV and CVX files identify correctly, so probably good for now.

CanvasX dropped supported for the Macintosh, but did release an entirely different product called Canvas X Draw, which does support the Macintosh. Here is what a CVD file looks like:

hexdump -C CanvasXDraw7-Sample1.cvd | head
00000000  25 43 61 6e 76 61 73 43  56 44 09 31 2e 30 25 bb  |%CanvasCVD.1.0%.|
00000010  54 48 65 61 64 65 72 00  00 00 00 00 00 00 00 00  |THeader.........|
00000020  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000030  00 bb 52 4d 61 63 4f 53  56 65 72 73 69 6f 6e 20  |..RMacOSVersion |
00000040  31 30 2e 31 33 2e 36 20  28 42 75 69 6c 64 20 31  |10.13.6 (Build 1|
00000050  37 47 31 34 30 34 32 29  31 30 2e 32 33 30 34 08  |7G14042)10.2304.|
00000060  00 00 00 70 6c 61 74 66  6f 72 6d 0a 73 00 00 00  |...platform.s...|
00000070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000080  00 00 00 00 00 05 00 00  00 02 00 00 00 00 00 00  |................|
00000090  00 08 00 00 00 6f 73 0a  73 00 00 00 00 00 00 00  |.....os.s.......|

There is also the matter of a Canvas Image, which the User Guide calls proxy images. They are Raster images used in placements within Canvas Documents. Should be easy to identify.

hexdump -C Canvas5-Sample1.CVI | head
00000000  00 00 00 01 44 41 44 35  50 52 4f 58 00 00 09 99  |....DAD5PROX....|
00000010  00 00 00 11 00 00 00 2d  00 00 00 03 00 00 00 08  |.......-........|
00000020  00 48 00 00 00 00 00 06  00 03 00 08 00 00 00 11  |.H..............|
00000030  00 00 00 2d 00 03 00 03  00 48 00 00 00 48 00 00  |...-.....H...H..|
00000040  00 00 00 00 00 00 00 00  00 00 00 11 00 00 00 2d  |...............-|
00000050  00 00 00 02 00 00 00 08  00 00 00 01 00 00 00 11  |................|
00000060  00 00 00 2d ff ff ff ff  ff ff ff ff ff ff ff ff  |...-............|
00000070  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|

Phew, if you held on for this whole post you must really like confusing file format structures. This format has been on my mind on and off for about 6 years. Hopefully these signatures will work for the vast majority of the Canvas files found in archives and personal systems. As always here is my GitHub with the signatures I am proposing and a few samples to get you confused.

FlashPix

Posted on by Thor Leave a comment

Is there a perfect raster image format? TIFF has been around quite some time and is generally accepted as a preferred preservation format. There have been a few attempts to have a single file contain multiple resolutions with the purpose of providing resolutions for different uses, lower-resolution for web and higher-resolution for print. Even the semi popular JPEG2000 added multiple resolutions to improve the JPEG format. Kodak came up with a few ideas to do this as well. The Kodak PCD, PhotoCD or Image PAC files was one that was used for awhile before it was abandoned. Another was FlashPix.

I briefly mentioned FlashPix on an earlier post about the Microsoft Picture It! format. They are extremely similar. Both. have the same basic structure in a Compound Object format. Some of the FlashPix files generated by Picture It! even have the same identifiers in the CompObj header.

FlashPix was supposed to be the answer to all the problems with storing bitmap image data and how we view the web. Kodak partnered with some big names, Microsoft Corporation, Hewlett-Packard Company and Live Picture, Inc, were among them. Kodak marketed the format and even included it as a native file format to some of its new digital cameras. The format was made official in June of 1996, with a Whitepaper explaining all the benefits and architecture. There was a lot of hype, some even calling it, “Not your Grandma’s format“. Many graphics software started to include support for the new format, including Adobe Photoshop. So what happened, why didn’t the format catch on? Some say it was the size of storing multiple resolutions in one file, others believe it was the complicated Compound Object structure that lead to its demise. Either way, the format had a lot of hype in the late 1990’s, but by the year 2000, it had gone silent and all the websites went away.

FlashPix did have a big impact, and there were many software and hardware devices which were made compatible. There are a few stories left behind of those who scanned all their photos to the FlashPix format only to find a few years later it was unsupported on more modern computers. There was also a few early digital camera’s which could capture directly to the format. Take my Kodak DC260 zoom camera, circa 1998. Changing the Capture Preferences, I can switch between a JPG and FPX.

Using exiftool we can take a look at one of the images from the camera:

exiftool P0004795.FPX
ExifTool Version Number         : 12.73
File Name                       : P0004795.FPX
Directory                       : GitHub/digicam_corpus/Kodak/DC260/DC260_01
File Size                       : 251 kB
File Modification Date/Time     : 2024:01:06 12:54:20-07:00
File Access Date/Time           : 2024:01:06 13:20:46-07:00
File Inode Change Date/Time     : 2024:01:06 13:04:34-07:00
File Permissions                : -rwxrwxrwx
File Type                       : FPX
File Type Extension             : fpx
MIME Type                       : image/vnd.fpx
Code Page                       : Unicode UTF-16, little endian
Data Object ID                  : 13BC5A58-6B90-1B6B-12C9-0800201177F8
Data Object Status              : Exists, Not Purgeable
Creating Transform              : Source Image
Using Transforms                : 
Cached Image Height             : 1024
Cached Image Width              : 1536
Comp Obj User Type Len          : 16
Comp Obj User Type              : FlashPix_Object
Visible Outputs                 : 1
Maximum Image Index             : 1
Maximum Transform Index         : 0
Maximum Operation Index         : 0
Thumbnail Clip                  : (Binary data 18480 bytes, use -b option to extract)
Revision Number                 : 1
Create Date                     : 2024:01:06 12:53:29
Modify Date                     : 2024:01:06 12:53:29
Software                        : KODAK DIGITAL SCIENCE DC260
Image Width                     : 1536
Image Height                    : 1024
Subimage Width                  : 1536
Subimage Height                 : 1024
Subimage Color                  : RGB
Subimage Numerical Format       : 8-bit, Unsigned
Decimation Method               : None (Full-sized Image)
JPEG Tables                     : (Binary data 558 bytes, use -b option to extract)
Number Of Resolutions           : 1
Max JPEG Table Index            : 1
Scene Type                      : Original Scene
Software Release                : KODAK DIGITAL SCIENCE DC260
Make                            : Eastman Kodak Company
Camera Model Name               : KODAK DIGITAL SCIENCE DC260
Serial Number                   : 7577
Exposure Time                   : 1/180
F Number                        : 4.7
Exposure Program                : Program AE
Exposure Compensation           : 0
Subject Distance                : 0.520 m
Metering Mode                   : Center-weighted average
Light Source                    : Unknown
Focal Length                    : 24.0 mm
Max Aperture Value              : 4.6
Flash                           : No Flash
Exposure Index                  : 90
Sharpness Approximation         : 0
File Source                     : Digital Camera
Sensing Method                  : One-chip color area
Extension Create Date           : 2024:01:06 12:53:29
Extension Modify Date           : 2024:01:06 12:53:29
Creating Application            : Picoss
Extension Name                  : ijuhsimasa
Extension Persistence           : Always Valid
Extension Description           : Data Object Store 000001
Storage-Stream Pathname         : /Data Object Store 000001
Extension Class ID              : 56616000-C154-11CE-8553-00AA00A1F95B
Used Extension Numbers          : 1
Screen Nail                     : (Binary data 4304 bytes, use -b option to extract)
Subimage Tile Count             : 384
Subimage Tile Width             : 64
Subimage Tile Height            : 64
Num Channels                    : 3
Audio Stream                    : (Binary data 30780 bytes, use -b option to extract)
Aperture                        : 4.7
Image Size                      : 1536x1024
Megapixels                      : 1.6
Shutter Speed                   : 1/180
Preview Image                   : (Binary data 4164 bytes, use -b option to extract)
Focal Length                    : 24.0 mm

The file also does identify in PRONOM:

sf P0004795.FPX 
---
siegfried   : 1.11.0
scandate    : 2024-01-17T23:13:59-07:00
signature   : default.sig
created     : 2023-12-17T15:54:41+01:00
identifiers : 
  - name    : 'pronom'
    details : 'DROID_SignatureFile_V116.xml; container-signature-20231127.xml'
---
filename : 'P0004795.FPX'
filesize : 250880
modified : 2024-01-06T12:54:20-07:00
errors   : 
matches  :
  - ns      : 'pronom'
    id      : 'x-fmt/56'
    format  : 'Kodak FlashPix Image'
    version : 
    mime    : 'image/vnd.fpx'
    class   : 'Image (Raster)'
    basis   : 'extension match fpx; container name CompObj with byte match at 53, 36 (signature 2/2)'
    warning :

If you notice, PRONOM has two signatures for the FlashPix format, this image was identified with signature #2. The first signature looks for the string “FlashPix Object”, but the second looks for the CLSID which is unique to each compound object format. FlashPix has the CLSID: {56616700-c154-11ce-8553-00aa00a1f95b}. Looking at many of the other samples I have there is much variation on the use of the string and CLSID.

FlashPix samples:
FlashPix Object({56616000-C154-11CE-8553-00AA00A1F95B}
FlashPix Object({56616800-C154-11CE-8553-00AA00A1F95B}
Picture It! FlashPix'{56616700-C154-11CE-8553-00AA00A1F95B}
LPI FlashPix'{56616700-c154-11ce-8553-00aa00a1f95b}
FlashPix_Object'{56616700-C154-11CE-8553-00AA00A1F95B}
'{56616700-C154-11CE-8553-00AA00A1F95B}
Picture It!'{56616700-c154-11ce-8553-00aa00a1f95b}
Flashpix Toolkit Application'{56616700-c154-11ce-0000-000000000000}

The images from the Kodak Camera use “FlashPix_Object” string so with the underscore it doesn’t match the first signature, but others I made using Picture It! software used a couple variations. Many don’t use the string at all. Others use a sightly different CLSID in both uppercase and lowercase. We will have to suggest adjustments to the current signature to identify them all.

Looking at the contents of the OLE container we can see some interesting things.

Path = P0004795.FPX
Type = Compound
Physical Size = 250880
Extension = compound
Cluster Size = 512
Sector Size = 64

Size         Compressed     Name
------------ ------------  ------------------------
188          192           [5]Data Object 000001
272          320           [1]CompObj
388          448           [5]Extension List
144          192           [5]Global Info
                           Data Object Store 000001
18704        18944         [5]SummaryInformation
816          832           Data Object Store 000001/[5]Image Contents
272          320           Data Object Store 000001/[1]CompObj
988          1024          Data Object Store 000001/[5]Extension List
1624         1664          Data Object Store 000001/[5]Image Info
4332         4608          Data Object Store 000001/[5]Screen Nail_bd0100609719a180
                           Data Object Store 000001/Resolution 0005
                           Data Object Store 000001/Audio_bd0100609719a180
1112         1152          Data Object Store 000001/[5]KDC_bd0100609719a180
72           128           Data Object Store 000001/[5]SummaryInformation
108          128           Data Object Store 000001/Audio_bd0100609719a180/[5]Audio Info
30808        31232         Data Object Store 000001/Audio_bd0100609719a180/Audio Stream 000000
6208         6656          Data Object Store 000001/Resolution 0005/Subimage 0000 Header
176378       176640        Data Object Store 000001/Resolution 0005/Subimage 0000 Data
------------ ------------  ------------------------
242414       244480        16 files, 3 folders

The main CompObj is where we find the identification information, but the Data Object Store 000001 directory is where all the image data is stored. In a multiple resolution image we might see additional Resolution directories. You may also notice a mention of an Audio directory. Yes, this image was captured and then audio was recorded with it. Not a video, but an audio clip associated with the image. FlashPix can contain audio streams. This isn’t the first time we have seen this, HP camera’s also have this function which as it turns out is stored in a FlashPix exif extension within a JPEG.

The FlashPix native format may have disappeared, but the format lives on as an extension to Exif data, allowing you to embed audio and other media within a JPEG file. The code for FlashPix was given to ImageMagick and is maintained by them.

PNG Plus

Posted on by Thor Leave a comment

Usually in the software world file formats are fairly efficient, the structure is meant to provide a way to store the data of the software being used. There isn’t much need to add additional unnecessary additions. This isn’t always true, but in the early days, disk space was expensive so compression and efficiency ruled. There also wasn’t much need to hide anything or complicate things. That is unless it is intended. This makes me think of two things, Polyglots and Steganography.

Steganography is the art of embedding data within an image. With digital images you can hide another image within the main image by using the most and least significant bits. Fun use of technology, but not something you normally would find in your regular desktop software.

Ange is the master at polyglots. If you haven’t watched his presentation on funky file formats, you are missing out.

Imagine my surprise when I was researching the Picture It! software and the MIX file format only to discover Microsoft decided to make their own polyglot of sorts for their PNG Plus format which replaced the MIX format, then both obsolete when Digital Image was discontinued in 2007. The PNG Plus format was the native format for the Microsoft Picture It! and Digital Image software often found with the Microsoft Works or Digital Imaging suite of software.

Save Menu from Digital Image Pro

According to the help within Digital Image:

The PNG Plus format uses the standard PNG extension but provides saving of layers and pages within the PNG format. Since the PNG format cannot do this natively, how did Microsoft accomplish this? Well, by throwing an OLE container into the middle of the file of course!

PNG Plus files are your regular PNG format and will identify as such. But they are just a low resolution thumbnail of the full image. Let’s take a look:

exiftool PictureIt7-s02.png 
ExifTool Version Number         : 12.70
File Name                       : PictureIt7-s02.png
File Size                       : 26 kB
File Modification Date/Time     : 2023:12:26 22:01:58-07:00
File Access Date/Time           : 2024:01:01 12:31:07-07:00
File Inode Change Date/Time     : 2023:12:26 22:01:58-07:00
File Permissions                : -rwx------
File Type                       : PNG
File Type Extension             : png
MIME Type                       : image/png
Image Width                     : 500
Image Height                    : 333
Bit Depth                       : 8
Color Type                      : RGB with Alpha
Compression                     : Deflate/Inflate
Filter                          : Adaptive
Interlace                       : Noninterlaced
SRGB Rendering                  : Perceptual
Gamma                           : 2.2
White Point X                   : 0.3127
White Point Y                   : 0.329
Red X                           : 0.64
Red Y                           : 0.33
Green X                         : 0.3
Green Y                         : 0.6
Blue X                          : 0.15
Blue Y                          : 0.06
Warning                  : [minor] Text/EXIF chunk(s) found after PNG IDAT (may be ignored by some readers)
Title                           : PictureIt7-s02
Image Size                      : 500x333
Megapixels                      : 0.167

Looks like there is some additional data after the IDAT chunk.

hexdump -C PictureIt7-s02.png | head
00000000  89 50 4e 47 0d 0a 1a 0a  00 00 00 0d 49 48 44 52  |.PNG........IHDR|
00000010  00 00 01 f4 00 00 01 4d  08 06 00 00 00 f6 13 9d  |.......M........|
00000020  37 00 00 00 01 73 52 47  42 00 ae ce 1c e9 00 00  |7....sRGB.......|
00000030  00 04 67 41 4d 41 00 00  b1 8f 0b fc 61 05 00 00  |..gAMA......a...|
00000040  00 20 63 48 52 4d 00 00  7a 26 00 00 80 84 00 00  |. cHRM..z&......|
00000050  fa 00 00 00 80 e8 00 00  75 30 00 00 ea 60 00 00  |........u0...`..|
00000060  3a 98 00 00 17 70 9c ba  51 3c 00 00 24 f4 49 44  |:....p..Q<..$.ID|
00000070  41 54 78 5e ed dd 4d a8  15 57 be 28 f0 1e 08 1e  |ATx^..M..W.(....|
00000080  e3 47 8e 49 ab c7 d8 81  03 09 41 9c 28 38 e8 80  |.G.I......A.(8..|
00000090  d0 9c 0e 08 0e 1a 11 c2  15 07 5e 5a 07 4d c7 2b  |..........^Z.M.+|

The header looks the same as any PNG file, so lets look a little further:

00002560  ff 1f fa 5f 90 66 c9 e6  ad 88 00 00 00 00 63 6d  |..._.f........cm|
00002570  4f 44 4e 88 09 c1 00 00  40 00 63 70 49 70 d0 cf  |ODN.....@.cpIp..|
00002580  11 e0 a1 b1 1a e1 00 00  00 00 00 00 00 00 00 00  |................|
00002590  00 00 00 00 00 00 3e 00  03 00 fe ff 09 00 06 00  |......>.........|
000025a0  00 00 00 00 00 00 00 00  00 00 01 00 00 00 01 00  |................|
000025b0  00 00 00 00 00 00 00 10  00 00 02 00 00 00 01 00  |................|
*
00002970  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff 52 00  |..............R.|
00002980  6f 00 6f 00 74 00 20 00  45 00 6e 00 74 00 72 00  |o.o.t. .E.n.t.r.|
00002990  79 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |y...............|
000029a0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
000029b0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 16 00  |................|
000029c0  05 00 ff ff ff ff ff ff  ff ff 01 00 00 00 7e 7f  |..............~.|
000029d0  3f b5 a5 f6 86 43 a1 a1  a3 02 24 d2 88 ef 00 00  |?....C....$.....|
000029e0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
000029f0  00 00 03 00 00 00 40 12  00 00 00 00 00 00 44 00  |......@.......D.|
00002a00  61 00 74 00 61 00 53 00  74 00 6f 00 72 00 65 00  |a.t.a.S.t.o.r.e.|
00002a10  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00003930  00 00 00 00 00 00 00 00  00 00 00 00 00 00 43 48  |..............CH|
00003940  4e 4b 49 4e 4b 20 04 00  07 00 0c 00 00 03 00 02  |NKINK ..........|
00003950  00 00 00 0a 00 00 f8 01  0c 00 ff ff ff ff 18 00  |................|
00003960  54 45 58 54 00 00 01 00  00 00 54 45 58 54 00 02  |TEXT......TEXT..|
00003970  00 00 22 00 00 00 18 00  46 44 50 50 00 00 43 00  |..".....FDPP..C.|
00003980  4f 00 4e 00 54 00 45 00  4e 00 54 00 53 00 00 00  |O.N.T.E.N.T.S...|
00003990  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
000039f0  00 00 1f 00 00 00 00 0a  00 00 00 00 00 00 01 00  |................|
00003a00  43 00 6f 00 6d 00 70 00  4f 00 62 00 6a 00 00 00  |C.o.m.p.O.b.j...|
00003a10  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00004530  00 00 00 00 00 00 00 00  00 00 00 00 00 00 01 00  |................|
00004540  fe ff 03 0a 00 00 ff ff  ff ff 00 00 00 00 00 00  |................|
00004550  00 00 00 00 00 00 00 00  00 00 1a 00 00 00 51 75  |..............Qu|
00004560  69 6c 6c 39 36 20 53 74  6f 72 79 20 47 72 6f 75  |ill96 Story Grou|
00004570  70 20 43 6c 61 73 73 00  ff ff ff ff 01 00 00 00  |p Class.........|
00004580  00 00 00 00 f4 39 b2 71  00 00 00 00 00 00 00 00  |.....9.q........|
00004590  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00006570  00 00 00 00 00 00 00 00  00 00 00 00 00 00 ba 84  |................|
00006580  43 51 00 00 00 18 69 54  58 74 54 69 74 6c 65 00  |CQ....iTXtTitle.|
00006590  00 00 00 00 50 69 63 74  75 72 65 49 74 37 2d 73  |....PictureIt7-s|
000065a0  30 32 3a 70 9c 00 00 00  00 14 74 45 58 74 54 69  |02:p......tEXtTi|
000065b0  74 6c 65 00 50 69 63 74  75 72 65 49 74 37 2d 73  |tle.PictureIt7-s|
000065c0  30 32 f2 8f d5 89 00 00  00 00 49 45 4e 44 ae 42  |02........IEND.B|
000065d0  60 82                                             |`.|

What what do we have here? Near the end of the file before the IEND chunk is an OLE file with the very recognizable hex values of “D0CF11E0“. Let’s strip out the OLE file and take a look.

Path = PictureIt7-s02-ole
Type = Compound
WARNINGS:
There are data after the end of archive
Physical Size = 8704
Tail Size = 7764
Extension = compound
Cluster Size = 512
Sector Size = 64

   Date      Time    Attr         Size   Compressed  Name
------------------- ----- ------------ ------------  ------------------------
2023-12-26 22:01:58 D....                            DataStore
2023-12-26 22:01:58 D....                            Text
                    .....         2560         2560  Text/CONTENTS
                    .....           86          128  Text/[1]CompObj
                    .....           96          128  DataStore/3
                    .....            4           64  DataStore/1
                    .....          121          128  DataStore/0
                    .....           57           64  DataStore/2
                    .....           98          128  DataStore/5
                    .....            4           64  DataStore/4
                    .....         1254         1280  DataStore/7
                    .....            4           64  DataStore/6
                    .....            4           64  DataStore/8
------------------- ----- ------------ ------------  ------------------------
2023-12-26 22:01:58               4288         4672  11 files, 2 folders

Interesting, I don’t think I have come across a standard format with a container embedded within. I have come across many OLE and ZIP containers which contain other common formats within, but this format is definitely unique. Others have added features in the IDAT chunk, such as a web shell. I am sure there are others out there. The CompObj file found within the Text directory is very similar to the Microsoft Works and Publisher format. Although trying to open the file in Publisher doesn’t work!

hexdump -C PictureIt7-s02-ole/Text/\[1\]CompObj | head
00000000  01 00 fe ff 03 0a 00 00  ff ff ff ff 00 00 00 00  |................|
00000010  00 00 00 00 00 00 00 00  00 00 00 00 1a 00 00 00  |................|
00000020  51 75 69 6c 6c 39 36 20  53 74 6f 72 79 20 47 72  |Quill96 Story Gr|
00000030  6f 75 70 20 43 6c 61 73  73 00 ff ff ff ff 01 00  |oup Class.......|
00000040  00 00 00 00 00 00 f4 39  b2 71 00 00 00 00 00 00  |.......9.q......|
00000050  00 00 00 00 00 00                                 |......|

PRONOM uses binary and container signatures to identify file formats. Even though this file format contains a valid OLE container, because it is within a regular binary file format, I don’t believe a container signature would work. The difficulty will be to clearly identify this new format without falsely identifying a regular PNG instead. The OLE file format header is not in a consistent location to use a specific offset. Making the string a variable location can causes some undo processing, so lets look to see if there is anything else we can use to make a positive ID.

The PNG file format is based on chunks, you have to have IHDR, then an IDAT and the IEND chunk. If we take a look at a regular PNG file using a libpng tool pngcheck, we see this:

pngcheck -cvt rgb-8.png 
File: rgb-8.png (759 bytes)
  chunk IHDR at offset 0x0000c, length 13
    256 x 256 image, 24-bit RGB, non-interlaced
  chunk tEXt at offset 0x00025, length 44, keyword: Copyright
    ? 2013,2015 John Cunningham Bowler
  chunk iTXt at offset 0x0005d, length 116, keyword: Licensing
    compressed, language tag = en
    no translated keyword, 101 bytes of UTF-8 text
  chunk IDAT at offset 0x000dd, length 518
    zlib: deflated, 32K window, maximum compression
  chunk IEND at offset 0x002ef, length 0
No errors detected in rgb-8.png (5 chunks, 99.6% compression).

The required chunk are there, but a couple extra, the tEXt and iTXt, which are textual metadata you can add. Now lets look at a PNG Plus file:

pngcheck -cvt PictureIt7-s02.png         
File: PictureIt7-s02.png (26066 bytes)
  chunk IHDR at offset 0x0000c, length 13
    500 x 333 image, 32-bit RGB+alpha, non-interlaced
  chunk sRGB at offset 0x00025, length 1
    rendering intent = perceptual
  chunk gAMA at offset 0x00032, length 4: 0.45455
  chunk cHRM at offset 0x00042, length 32
    White x = 0.3127 y = 0.329,  Red x = 0.64 y = 0.33
    Green x = 0.3 y = 0.6,  Blue x = 0.15 y = 0.06
  chunk IDAT at offset 0x0006e, length 9460
    zlib: deflated, 32K window, fast compression
  chunk cmOD at offset 0x0256e, length 0
    Microsoft Picture It private, ancillary, unsafe-to-copy chunk
  chunk cpIp at offset 0x0257a, length 16384
    Microsoft Picture It private, ancillary, safe-to-copy chunk
  chunk iTXt at offset 0x06586, length 24, keyword: Title
    uncompressed, no language tag
    no translated keyword, 15 bytes of UTF-8 text
  chunk tEXt at offset 0x065aa, length 20, keyword: Title
    PictureIt7-s02
  chunk IEND at offset 0x065ca, length 0
No errors detected in PictureIt7-s02.png (10 chunks, 96.1% compression).

It looks like we have the required chunks and some textual chunks but also a couple chunks which pngcheck describes as private and identify’s them as Microsoft Picture It chunks. The cpIp chunk is the one which contains the OLE container. This is the chunk we need to identify in a signature. The problem is the offset for the cpIp chunk is not the same each time. Here is one from Digital Image 10 Pro.

  chunk cpIp at offset 0x737a7, length 245760
    Microsoft Picture It private, ancillary, safe-to-copy chunk

Significantly further in the file that the other example. These samples currently identify as PNG 1.2 files. PRONOM fmt/13 so we can use the signature and add to it, but it currently doesn’t look for IDAT only the iTXt chunk, which is probably not optimal. For PNG Plus, lets get the header which includes IHDR, IDAT, then the cpIp chunk then an end of file sequence for IEND. Take a look at my signature and samples, I am curious how many PNG Plus files are out there hidden to the world.

Turns out there is another PNG flavor which has been enhanced to allow for layers and pages. Adobe Fireworks uses a PNG format as their native format. They also use private chunks, but not within an OLE container. They use additional chunks, but before the IDAT chunk:

  chunk prVW at offset 0x00092, length 1700
    Macromedia Fireworks preview chunk (private, ancillary, unsafe to copy)
  chunk mkBF at offset 0x00742, length 72
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkTS at offset 0x00796, length 36716
    Macromedia Fireworks(?) private, ancillary, unsafe-to-copy chunk
  chunk mkBS at offset 0x0970e, length 190
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x097d8, length 1251
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x09cc7, length 1358
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0a221, length 1145
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0a6a6, length 339
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0a805, length 695
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0aac8, length 3799
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0b9ab, length 7733
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0d7ec, length 2741
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0e2ad, length 5153
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk
  chunk mkBT at offset 0x0f6da, length 10775
    Macromedia Fireworks private, ancillary, unsafe-to-copy chunk

It’s hard to know which each of the chunks are for and if they are all required for the Fireworks PNG format. From the book on PNG.

In addition to supporting PNG as an output format, Fireworks actually uses PNG as its native file format for day-to-day intermediate saves. This is possible thanks to PNG’s extensible “chunk-based” design, which allows programs to incorporate application-specific data in a well-defined way. Macromedia has embraced this capability, defining at least four custom chunk types that hold various things pertinent to the editor. Unfortunately, one of them (pRVW) violates the PNG naming rules by claiming to be an officially registered, public chunk type, but this was an oversight and should be fixed in version 2.0.

Picture It!

Posted on by Thor 1 Comment

Most everyone has heard of Microsoft Office, the suite of applications used by millions everyday. Less people know about Microsoft Works, which was a lower cost alternative, but was quite popular as a home office suite of applications. One tool which often came with the Works suite was a digital image tool called Picture It!

Picture It! was a photo editing tool first released by Microsoft in 1996 geared to making photo editing easy and affordable.

Picture It! used a wizard type interface which walked you through acquiring an image and adding to it. One of the key features of the software was the ability to “stack” objects like layers. Because of this feature a new file format was used to save this information to disk. Meet the Microsoft Image (Picture) Extension format, commonly known as the MIX file format. It is very similar to the FlashPix image format, which was supposed to be an image file format to solve many delivery issues, but didn’t seem to gain hold despite being created by Kodak, HP, and others. In fact many of the MIX files I found on Microsoft disks are actually FlashPix files.

The MIX extension was also used by another Microsoft program, PhotoDraw, which causes confusion as they were similar, but PhotoDraw has some added features which may not be compatible with Picture It!. Both formats are based on the Microsoft Compound Object (OLE) container, and have a similar structure. Let’s take a look at a MIX file from Picture It! version 1. 

7z l PictureIt1-s02.mix                 

--
Path = PictureIt1-s02.mix
Type = Compound
Physical Size = 48128
Extension = compound
Cluster Size = 512
Sector Size = 64

   Date      Time    Attr         Size   Compressed  Name
------------------- ----- ------------ ------------  ------------------------
                    .....          328          384  [5]Data Object 000001
                    .....          396          448  [5]Transform 000004
                    .....          872          896  [5]Operation 000001
                    .....          320          320  [1]CompObj
                    .....          292          320  [5]Global Info
                    .....          872          896  [5]Operation 000002
                    .....          144          192  [5]Operation 000003
                    .....          684          704  [5]Transform 000008
                    .....         1028         1088  [5]Transform 000009
                    .....          328          384  [5]Data Object 000009
                    .....          324          384  [5]Data Object 000005
2023-12-27 11:04:39 D....                            Data Object Store 000001
                    .....          328          384  [5]Data Object 000010
                    .....        20932        20992  [5]SummaryInformation
                    .....          200          256  [5]Microsoft Embedding Info
2023-12-27 11:04:39 D....                            Data Object Store 000001/Resolution 0001
                    .....         1400         1408  Data Object Store 000001/[5]Image Contents
                    .....          230          256  Data Object Store 000001/[1]CompObj
2023-12-27 11:04:39 D....                            Data Object Store 000001/Resolution 0000
                    .....           28           64  Data Object Store 000001/Resolution 0000/Subimage 0000 Data
                    .....           80          128  Data Object Store 000001/Resolution 0000/Subimage 0000 Header
2023-12-27 11:04:39 D....                            Data Object Store 000001/Resolution 0003
2023-12-27 11:04:39 D....                            Data Object Store 000001/Resolution 0002
                    .....           28           64  Data Object Store 000001/Resolution 0002/Subimage 0000 Data
                    .....          208          256  Data Object Store 000001/Resolution 0002/Subimage 0000 Header
2023-12-27 11:04:39 D....                            Data Object Store 000001/Resolution 0005
2023-12-27 11:04:39 D....                            Data Object Store 000001/Resolution 0004
                    .....           28           64  Data Object Store 000001/Resolution 0004/Subimage 0000 Data
                    .....         1792         1792  Data Object Store 000001/Resolution 0004/Subimage 0000 Header
                    .....          124          128  Data Object Store 000001/[5]SummaryInformation
                    .....           28           64  Data Object Store 000001/Resolution 0005/Subimage 0000 Data
                    .....         6976         7168  Data Object Store 000001/Resolution 0005/Subimage 0000 Header
                    .....           28           64  Data Object Store 000001/Resolution 0003/Subimage 0000 Data
                    .....          544          576  Data Object Store 000001/Resolution 0003/Subimage 0000 Header
                    .....           28           64  Data Object Store 000001/Resolution 0001/Subimage 0000 Data
                    .....          128          128  Data Object Store 000001/Resolution 0001/Subimage 0000 Header
------------------- ----- ------------ ------------  ------------------------
2023-12-27 11:04:39              38698        39872  29 files, 7 folders

This is a simple MIX file with one line of text, but contains a lot of content inside the OLE container. If I try and use the PRONOM registry to identify the file, I get:

sf PictureIt1-s02.mix 
---
siegfried   : 1.11.0
scandate    : 2023-12-27T11:06:32-07:00
signature   : default.sig
created     : 2023-12-17T15:54:41+01:00
identifiers : 
  - name    : 'pronom'
    details : 'DROID_SignatureFile_V116.xml; container-signature-20231127.xml'
---
filename : 'PictureIt1-s02.mix'
filesize : 48128
modified : 2023-12-27T11:04:40-07:00
errors   : 
matches  :
  - ns      : 'pronom'
    id      : 'fmt/111'
    format  : 'OLE2 Compound Document Format'
    version : 
    mime    : 
    class   : 'Text (Structured)'
    basis   : 'byte match at 0, 30'
    warning :

Hmm, we know it is an OLE compound document, but it should identify as a Picture It! file as PRONOM has defined a PUID for the format. fmt/936 has been defined as “Microsoft Picture It! Image File 1”. So I am not sure why this file from version 1 is not identifying correctly. Let’s take a look. The PRONOM container signature for fmt/936 is looking for this:

    <ContainerSignature Id="17015" ContainerType="OLE2">
      <Description>Microsoft Picture It! Image File</Description>
      <Files>
        <File>
          <Path>CompObj</Path>
          <BinarySignatures>
            <InternalSignatureCollection>
              <InternalSignature ID="17015">
                <ByteSequence Reference="BOFoffset">
                  <SubSequence Position="1" SubSeqMinOffset="32"
                               SubSeqMaxOffset="32">
                    <Sequence>'Microsoft Picture It! version 1 Picture'</Sequence>
                  </SubSequence>
                </ByteSequence>
              </InternalSignature>
            </InternalSignatureCollection>
          </BinarySignatures>
        </File>
      </Files>
    </ContainerSignature>

The container signature is looking into the OLE container for the “CompObj” file (which seems to be required), then looks for the string “Microsoft Picture It! version 1 Picture” starting at the 32nd byte. That is pretty specific. The sample file I am using as an example has the following string of bytes.

hexdump -C PictureIt1-s02/\[1\]CompObj 
00000000  01 00 fe ff 03 0a 00 00  ff ff ff ff 00 68 61 56  |.............haV|
00000010  54 c1 ce 11 85 53 00 aa  00 a1 f9 5b 1e 00 00 00  |T....S.....[....|
00000020  4d 69 63 72 6f 73 6f 66  74 20 50 69 63 74 75 72  |Microsoft Pictur|
00000030  65 20 49 74 21 20 50 69  63 74 75 72 65 00 27 00  |e It! Picture.'.|
00000040  00 00 7b 35 36 36 31 36  38 30 30 2d 43 31 35 34  |..{56616800-C154|
00000050  2d 31 31 43 45 2d 38 35  35 33 2d 30 30 41 41 30  |-11CE-8553-00AA0|
00000060  30 41 31 46 39 35 42 7d  00 13 00 00 00 50 69 63  |0A1F95B}.....Pic|
00000070  74 75 72 65 49 74 21 2e  50 69 63 74 75 72 65 00  |tureIt!.Picture.|

Ok, so this sample has a similar string but is missing the “version 1” text. It seems the samples used to created the PRONOM signature was working off samples which included the version 1 in the header of CompObj. Maybe when Microsoft learned they would be making a version 2, they decided a version number should be included going forward. Let’s take a look a file from version 2 to compare:

hexdump -C PictureIt2-s01/\[1\]CompObj 
00000000  01 00 fe ff 03 0a 00 00  ff ff ff ff 50 28 72 2d  |............P(r-|
00000010  4b 8c d0 11 a9 6f 00 a0  c9 05 41 0d 28 00 00 00  |K....o....A.(...|
00000020  4d 69 63 72 6f 73 6f 66  74 20 50 69 63 74 75 72  |Microsoft Pictur|
00000030  65 20 49 74 21 20 76 65  72 73 69 6f 6e 20 32 20  |e It! version 2 |
00000040  50 69 63 74 75 72 65 00  27 00 00 00 7b 32 44 37  |Picture.'...{2D7|
00000050  32 32 38 35 30 2d 38 43  34 42 2d 31 31 44 30 2d  |22850-8C4B-11D0-|
00000060  41 39 36 46 2d 30 30 41  30 43 39 30 35 34 31 30  |A96F-00A0C905410|
00000070  44 7d 00 f4 39 b2 71 50  00 00 00 4d 00 69 00 63  |D}..9.qP...M.i.c|

Ok, so it looks like they did update the version string for version 2. This file also does not identify correctly. A quick look at the wikipedia page for Microsoft Picture It! tells us they continued to release the software until version 10. Is there a different string for each version?

Diving into this and gathering many samples has brought a lot of variants to surface. Let’s see if we can list all the CompObj header variants.

Version 1 samples:
Picture It! Picture'{56616800-C154-11CE-8553-00AA00A1F95B}
Microsoft Picture It! Picture'{56616800-C154-11CE-8553-00AA00A1F95B}
Microsoft Picture It! version 1 Picture'{56616800-C154-11CE-8553-00AA00A1F95B}
Picture It! Collage'{56616800-C154-11CE-8553-00AA00A1F95B}

Version 2 samples:
Microsoft Picture It! version 2 Picture'{2D722850-8C4B-11D0-A96F-00A0C905410D}

Version 3 samples:
Microsoft Picture It! version 3 Picture'{18B8D020-B4FD-11D0-A97E-00A0C905410D}

Version 4 samples:
Microsoft Picture It! version 4 Picture'{18B8D020-B4FD-11D0-A97E-00A0C905410D}

PhotoDraw version 1 samples:
Microsoft PhotoDraw version 1 Picture'{18B8D020-B4FD-11D0-A97E-00A0C905410D}

PhotoDraw version 2 samples:
Microsoft PhotoDraw version 2 Picture'{18B8D021-B4FD-11D0-A97E-00A0C905410D}

FlashPix samples:
FlashPix Object({56616000-C154-11CE-8553-00AA00A1F95B}
FlashPix Object({56616800-C154-11CE-8553-00AA00A1F95B}
Picture It! FlashPix'{56616700-C154-11CE-8553-00AA00A1F95B}
LPI FlashPix'{56616700-c154-11ce-8553-00aa00a1f95b}
FlashPix_Object'{56616700-C154-11CE-8553-00AA00A1F95B}
'{56616700-C154-11CE-8553-00AA00A1F95B}
Picture It!'{56616700-c154-11ce-8553-00aa00a1f95b}
Flashpix Toolkit Application'{56616700-c154-11ce-0000-000000000000}

Ok, there is a lot to discuss here. First of all, it seems MIX was only used in Picture It! until version 5 (2001), then the Picture It! software used a new format, PNG Plus to store the layered stacks. More on that in a future post! Although some later versions seems to be able to open the older MIX format. Version 4 of the MIX format seems to be the last as the 2001 software had only version 4 files on it. Probably safe to say only the 4 versions are needed for identification.

You may notice the additional unique identifier I included in each format. This is called a Class ID for the OLE format, which A LOT of formats use. Each “format” has a unique ID associated with it to help distinguish it from other formats. This Unique ID could possibly be a better solution for identification. It does cross over with the PhotoDraw format, but the FlashPix format seems to have a unique ID. With all the variations in the version 1 strings, the ID remains the same. For version 3 and 4 the ID is the same, which could mean they are interchangeable. It is also the same as PhotoDraw version 1. Not to complicate things.

So it seems in order to get proper identification of these similar formats we need to:

  • Clean up version 1 identification for fmt/936
  • Add a signature for 2, 3, and 4
  • Add a version 2 signature for the PhotoDraw format
  • Add some additional signature variations for the FlashPix format.

The Class ID’s could be used to distinguish different versions and formats, but many of the ID’s are identical, this could mean they are the same format. But for now we can just add the additional variation strings and it should identify everything for now. The FlashPix format needs more research as there is so many different variations and it’s so close to the MIX format. Take a look at my GitHub submission, maybe you have some additional variations to add?

Digital Negatives

Posted on by Thor Leave a comment

One of the important parts about Digital Preservation is to gather significant properties of the digital files we hope to preserve. This can allow us to base our risk assessments off of more data than just an extension. For example, a TIFF file is a mighty good preservation format. Well documented and adopted by the preservation community, and with hundreds if not thousands of software tools to render and make use of the format. But if a TIFF file uses compression like LZW, or if it happens to have multiple pages, those are good things to know about. Most formats might have a stable set of properties, but sometimes can have properties which adds more risk to the format becoming difficult to render or migrate.

A DNG or Digital Negative developed by Adobe was supposed to solve the issues with proprietary RAW digital camera formats. Rendering a PhaseOne IIQ file often times requires the full CaptureOne software which can be expensive. Adobe spends quite a bit of resources in adding support to its Camera RAW toolkit and adding the ability to take majority of these RAW formats and move them into a DNG. There is also more and more camera manufacturers who image directly to a DNG as their native RAW format. This is the case for Apple’s ProRAW format which uses the DNG specification.

Another manufacturer is the Insta360 camera’s. Their 360 camera’s can use two lenses to capture 180 degrees from each and then stitch into a 360 photo or video. They can capture compressed images and videos, but also in RAW. Because of the two lenses and sensors, their DNG’s can get quite large. For this reason I recently asked PRONOM to adjust their signatures to allow for a bigger offset of DNG information in the larger RAW images. 

exiftool IMG_20230913_141939_00_039.dng 
ExifTool Version Number         : 12.70
File Name                       : IMG_20230913_141939_00_039.dng
File Size                       : 143 MB
File Type                       : DNG
File Type Extension             : dng
MIME Type                       : image/x-adobe-dng
Exif Byte Order                 : Little-endian (Intel, II)
Subfile Type                    : Full-resolution image
Image Width                     : 5984
Image Height                    : 11968
Bits Per Sample                 : 16
Compression                     : Uncompressed
Photometric Interpretation      : Color Filter Array
Make                            : Arashi Vision
Camera Model Name               : Insta360 X3

DNG files are actually based on the TIFF format, TIFF/EP to be precise, which means there is some good history behind the format and understanding of its structure. DNG does add many new tags and new features, so there is much more going on. Here is a TIFFInfo view of a DNG. Lots of new tags…..

tiffinfo IMG_20230913_141939_00_039.dng 
TIFFReadDirectory: Warning, Unknown field with tag 33421 (0x828d) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 33422 (0x828e) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 50937 (0xc6f9) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 50938 (0xc6fa) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 50940 (0xc6fc) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 51009 (0xc741) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 51107 (0xc7a3) encountered.
=== TIFF directory 0 ===
TIFF Directory at offset 0x889946c (143234156)
  Subfile Type: (0 = 0x0)
  Image Width: 5984 Image Length: 11968
  Bits/Sample: 16
  Sample Format: unsigned integer
  Compression Scheme: None
  Photometric Interpretation: 32803 (0x8023)
  Orientation: row 0 top, col 0 lhs
  Samples/Pixel: 1
  Rows/Strip: 11968
  Planar Configuration: single image plane
  Make: Arashi Vision
  Model: Insta360 X3
  Software: v1.0.69_build1
  DateTime: 2023:09:13 14:19:40
  Tag 33421: 2,2
  Tag 33422: 1,2,0,1
  EXIFIFDOffset: 0x8
  GPSIFDOffset: 0x3e6
  DNGVersion: 1,3,0,0
  DNGBackwardVersion: 1,3,0,0
  UniqueCameraModel: Insta360 X3

An IFD (Image File Directory) is the building block of a TIFF file. A TIFF file can have multiple IFD’s within a single file. But an IFD can also be a thumbnail, metadata or GPS info. For a DNG, they use the IFD structure as well, but often, the first IFD is a lower resolution of the full image. 

 <File:FileType>DNG</File:FileType>
 <File:FileTypeExtension>dng</File:FileTypeExtension>
 <File:MIMEType>image/x-adobe-dng</File:MIMEType>
 <File:ExifByteOrder>Little-endian (Intel, II)</File:ExifByteOrder>
 <IFD0:SubfileType>Reduced-resolution image</IFD0:SubfileType>
 <IFD0:ImageWidth>256</IFD0:ImageWidth>
 <IFD0:ImageHeight>171</IFD0:ImageHeight>
 <IFD0:BitsPerSample>8 8 8</IFD0:BitsPerSample>
 <IFD0:Compression>Uncompressed</IFD0:Compression>
 <IFD0:PhotometricInterpretation>RGB</IFD0:PhotometricInterpretation>
 <IFD0:Make>Canon</IFD0:Make>
 <IFD0:Model>Canon EOS RP</IFD0:Model>
...
 <SubIFD:SubfileType>Full-resolution image</SubIFD:SubfileType>
 <SubIFD:ImageWidth>6384</SubIFD:ImageWidth>
 <SubIFD:ImageHeight>4224</SubIFD:ImageHeight>
 <SubIFD:BitsPerSample>16</SubIFD:BitsPerSample>
 <SubIFD:Compression>JPEG</SubIFD:Compression>

But not always the same way.

 <IFD0:SubfileType>Full-resolution image</IFD0:SubfileType>
 <IFD0:ImageWidth>5984</IFD0:ImageWidth>
 <IFD0:ImageHeight>11968</IFD0:ImageHeight>
 <IFD0:BitsPerSample>16</IFD0:BitsPerSample>
 <IFD0:Compression>Uncompressed</IFD0:Compression>
 <IFD0:PhotometricInterpretation>Color Filter Array</IFD0:PhotometricInterpretation>
 <IFD0:Make>Arashi Vision</IFD0:Make>
 <IFD0:Model>Insta360 X3</IFD0:Model>

 <IFD0:SubfileType>Reduced-resolution image</IFD0:SubfileType>
 <IFD0:ImageWidth>4032</IFD0:ImageWidth>
 <IFD0:ImageHeight>3024</IFD0:ImageHeight>
 <IFD0:BitsPerSample>8 8 8</IFD0:BitsPerSample>
 <IFD0:Compression>JPEG</IFD0:Compression>
 <IFD0:PhotometricInterpretation>YCbCr</IFD0:PhotometricInterpretation>
 <IFD0:Make>Apple</IFD0:Make>
 <IFD0:Model>iPhone 13 Pro</IFD0:Model>
...
 <SubIFD:SubfileType>Full-resolution image</SubIFD:SubfileType>
 <SubIFD:ImageWidth>4032</SubIFD:ImageWidth>
 <SubIFD:ImageHeight>3024</SubIFD:ImageHeight>
 <SubIFD:BitsPerSample>12 12 12</SubIFD:BitsPerSample>
 <SubIFD:Compression>JPEG</SubIFD:Compression>

It can get confusing, especially for tools we use to extract metadata and significant properties from a DNG for preservation. Within Rosetta, the preservation system I use at work, there is no dedicated DNG extractor, so we use JHOVE, as it is the tool we use for our TIFF images. This presents a problem as the process only extracts properties for the first IFD assuming it is the main IFD, but in many cases it reports back the image is much smaller in pixel dimensions than it actually is. More work is needed to improve extracting correct significant properties for DNG and other RAW image formats.

Adobe released a new version of DNG this year. In June, DNG version 1.7.0.0 was finalized. The new version brought a few new features, two of which are including JPEG XL compression and a new HDR colorimetric value. In order to add JPEG XL compression DNG version 1.7 is required. Here is how one looks in exiftool, created with Adobe DNG Converter 16.1.

exiftool _MG_9375_1.dng 
ExifTool Version Number         : 12.70
File Name                       : _MG_9375_1.dng
File Size                       : 5.4 MB
File Type                       : DNG
File Type Extension             : dng
MIME Type                       : image/x-adobe-dng
Exif Byte Order                 : Little-endian (Intel, II)
Make                            : Canon
Camera Model Name               : Canon EOS DIGITAL REBEL XT
Preview Image Start             : 91884
Orientation                     : Rotate 270 CW
Rows Per Strip                  : 171
Preview Image Length            : 10305
Software                        : Adobe DNG Converter 16.1 (Macintosh)
Modify Date                     : 2023:12:18 11:45:06
Artist                          : unknown
Image Width                     : 3516
Image Height                    : 2328
Bits Per Sample                 : 16
Compression                     : JPEG XL
DNG Version                     : 1.7.1.0
DNG Backward Version            : 1.7.1.0

I had recently submitted a new signature for DNG 1.7 to PRONOM, but I found this new DNG version falls outside the signature I created. I had made the assumption all DNG’s report their version based on the last two values of 0.0, so I created the signature to look for 1.7.0.0. This is wrong now that I can see an example of version 1.7.1.0.

In order to fix the issue, I would need to change all the DNG signatures to remove the last two bytes so:

12C601000400000001070000 would change to 12C60100040000000107

This would allow for identification if some DNG files have a point version.

The pace at which manufacturers are producing camera’s with new features is much faster than the Digital Preservation community can keep up with. As new technologies get released, we play catch up trying to identify new formats and variations to existing ones. I guess that is job security?

Adobe Acrobat Capture

Posted on by Thor 1 Comment

During the recent PRONOM Research Week, I noticed a file format with no description and no signature.

x-fmt/217Adobe ACD

All I had to go on was it was an Adobe format and the acronym “ACD”. One of the first results that came up in a google search was a post in the Adobe forums with someone asking what to do with some old ACD and ACI files they found on a disc, circa 2000, labeled “Adobe Capture”. The only thing I remember about Adobe Capture was some scanning tools related to Adobe Acrobat, but I didn’t remember coming across any ACD files related to Acrobat.

Initially it wasn’t easy to find more information on this format. Eventually I was able to narrow it down to stand-alone software adobe released called “Adobe Acrobat Capture”. Originally released in 1995 it was eventually discontinued in 2010. The software was marketed under the ePaper name and connected to Acrobat through the creation of a PDF from scanned images. The software was compatible with many scanner models and would process the scanned images, run Optical Character recognition, and export to a searchable PDF. These tools are built into Adobe Acrobat today.

One of the reasons the software was being so elusive is the fact it was sold with a high price tag and required the use of a hardware key, or dongle, in order to process scans. The hardware key also managed the type of license you purchased which may limit the number of pages you are allowed to scan within a certain period of time. So the software is very difficult to run today, if you do happen to find a copy out there in Internet land.

In order to document these file formats for preservation purposes I needed to find some samples. I was excited to find a demonstration CD on the Internet Archive, but unfortunately it contained no examples of the ACD file format.

A little sleuthing on the Wayback Machine helped me find a few user guides and brochures. I was also able to find there was three versions of Adobe Acrobat Capture. In a Product Brochure, you can see a screenshot of the software with a document open with the ACD extension.

If you are OCD like me you might have noticed the window in this screenshot is typical of the older Windows 3.1 or Windows NT system. So this was indeed an older product released by Adobe.

The Adobe Acrobat Capture 3.0 Demonstration CD-ROM from the Internet Archive luckily has a UserGuide PDF on the disc and was able to help me understand the ACD format a little more.

Looks like the ACD format is an intermediate format used by the software to manage the process between scanning and export to PDF. ACD was also defined as an “Acrobat Capture Document” which makes sense. They were also mentioned as being “multipage files in Acrobat Capture Document (ACD)”. The UserGuide also mentioned an ACP format which it referenced as “one-page files are in Acrobat Capture Page (ACP) format.” So more research is needed.

Lets start with Adobe Acrobat Capture 2.0 as I managed to get a few samples from an installer I found. Here is a hexdump of an ACD file and its corresponding ACI file.

hexdump -C CONTRACT.ACD | head
00000000  02 04 47 47 c9 00 86 b5  01 00 b6 27 02 00 01 00  |..GG.......'....|
00000010  f5 00 5e 00 3b 96 02 00  01 6e 63 6a 00 00 88 68  |..^.;....ncj...h|
00000020  00 00 26 00 44 3a 5c 43  4f 44 45 5c 47 47 5c 50  |..&.D:\CODE\GG\P|
00000030  52 4f 44 55 43 54 2e 33  32 53 5c 49 4e 5c 63 6f  |RODUCT.32S\IN\co|
00000040  6e 74 72 61 63 74 2e 61  63 69 00 00 00 00 00 00  |ntract.aci......|
00000050  7c 33 c0 27 00 40 ff ff  ff 00 03 00 03 00 00 00  ||3.'.@..........|
00000060  00 00 00 00 00 00 40 00  00 00 00 00 00 03 00 00  |......@.........|
00000070  00 00 00 00 00 00 00 40  00 00 00 00 09 00 0a ab  |.......@........|
00000080  04 0b 14 b5 04 39 19 00  40 00 00 00 00 0c 14 b0  |.....9..@.......|
00000090  04 38 19 b0 04 08 00 0a  7f 06 d3 11 89 06 39 17  |.8............9.|

hexdump -C CONTRACT.ACI | head
00000000  49 49 2a 00 b3 0c 02 00  35 80 78 a0 80 35 c0 78  |II*.....5.x..5.x|
00000010  a4 80 35 40 3c 54 40 01  e2 b2 01 e2 b2 01 e2 b2  |..5@<T@.........|
00000020  01 e2 b2 01 e2 b2 01 e2  b2 01 e2 b2 01 e2 b2 01  |................|
00000030  e2 b2 01 e2 b2 01 e2 b2  01 e2 b2 01 e2 b2 01 e2  |................|
00000040  b2 01 e2 b2 01 e2 b2 01  e2 b2 01 e2 b2 01 e2 b2  |................|
00000050  01 e2 b2 01 e2 b2 01 e2  b2 01 e2 b2 01 e2 b2 01  |................|
00000060  e2 b2 01 e2 b2 01 e2 b2  01 e2 b2 01 e2 b2 01 e2  |................|
00000070  b2 01 e2 b2 01 e2 b2 01  e2 b2 01 e2 b2 01 e2 b2  |................|
00000080  01 e2 b2 01 e0 b0 01 e0  b0 01 e0 b0 01 e0 b0 01  |................|
00000090  e0 b0 01 e0 b0 01 e0 b0  01 e0 b0 01 e0 b0 01 e0  |................|

The ACD file is unique, PRONOM and even TrID was unaware of the format. But to the keen observer, the ACI format is very recognizable. You may have seen this header before:

Lets take a closer look at an ACI file to see if they are a true TIFF image or if there is any customization to the format.

tiffinfo CONTRACT.ACI 
=== TIFF directory 0 ===
TIFF Directory at offset 0x20cb3 (134323)
  Subfile Type: (0 = 0x0)
  Image Width: 2544 Image Length: 3295
  Resolution: 300, 300
  Bits/Sample: 1
  Compression Scheme: CCITT RLE
  Photometric Interpretation: min-is-white
  Samples/Pixel: 1
  Rows/Strip: 32
  Planar Configuration: single image plane
  Software: HALO Desktop Imager

exiftool -D CONTRACT.ACI 
    - ExifTool Version Number         : 12.60
    - File Name                       : CONTRACT.ACI
    - Directory                       : TUTORIAL/SAMPOUT
    - File Size                       : 134 kB
    - File Modification Date/Time     : 1995:07:10 16:02:08-06:00
    - File Access Date/Time           : 2023:11:14 15:41:02-07:00
    - File Inode Change Date/Time     : 2023:11:08 08:34:18-07:00
    - File Permissions                : -rwxrwxrwx
    - File Type                       : TIFF
    - File Type Extension             : tif
    - MIME Type                       : image/tiff
    - Exif Byte Order                 : Little-endian (Intel, II)
  254 Subfile Type                    : Full-resolution image
  256 Image Width                     : 2544
  257 Image Height                    : 3295
  258 Bits Per Sample                 : 1
  259 Compression                     : CCITT 1D
  262 Photometric Interpretation      : WhiteIsZero
  273 Strip Offsets                   : (Binary data 625 bytes, use -b option to extract)
  277 Samples Per Pixel               : 1
  278 Rows Per Strip                  : 32
  279 Strip Byte Counts               : (Binary data 448 bytes, use -b option to extract)
  282 X Resolution                    : 300
  283 Y Resolution                    : 300
  305 Software                        : HALO Desktop Imager
    - Image Size                      : 2544x3295
    - Megapixels                      : 8.4

Looks like a true TIFF image with no special tags or unique properties. They are 1-bit TIFF’s compressed with CCITT RLE. Not sure there would be any need to create a special signature for these ACI files.

Looking closer at the ACD file format, we can see they reference ACI files, so probably safe to assume the ACD file doesn’t contain the full raster data for each image:

hexdump -C Report.acd
00000000  02 04 47 47 c9 00 9a 8b  00 00 d4 ce 00 00 03 00  |..GG............|
00000010  f5 02 5f 00 00 61 01 00  01 6e 63 6a 01 00 30 5f  |.._..a...ncj..0_|
00000020  00 00 27 00 63 3a 5c 63  61 70 74 75 72 65 32 5c  |..'.c:\capture2\|
00000030  73 61 6d 70 6c 65 73 5c  6f 75 74 5c 52 65 70 6f  |samples\out\Repo|
00000040  72 74 5f 30 30 30 31 2e  61 63 69 00 00 01 00 00  |rt_0001.aci.....|
00000050  00 00 00 00 00 00 00 00  00 00 e8 03 00 00 01 00  |................|
00000060  01 00 00 00 00 00 00 00  00 00 08 00 52 65 70 6f  |............Repo|
00000070  72 74 30 31 00 00 00 00  70 33 d8 27 00 40 ff ff  |rt01....p3.'.@..|
*
00005f40  07 00 40 6f 00 09 00 40  01 6e 63 6a 02 00 52 2c  |..@o...@.ncj..R,|
00005f50  00 00 27 00 63 3a 5c 63  61 70 74 75 72 65 32 5c  |..'.c:\capture2\|
00005f60  73 61 6d 70 6c 65 73 5c  6f 75 74 5c 52 65 70 6f  |samples\out\Repo|
00005f70  72 74 5f 30 30 30 32 2e  61 63 69 00 00 00 00 00  |rt_0002.aci.....|
00005f80  00 00 00 00 4e 0c fe ff  ff ff e8 03 00 00 01 00  |....N...........|
00005f90  01 00 00 00 00 00 00 00  00 00 08 00 52 65 70 6f  |............Repo|
00005fa0  72 74 30 32 00 00 00 00  4c 31 f0 27 00 40 ff ff  |rt02....L1.'.@..|

From the limited sample set I have access, all the ACD files begin with the same Hex values, “02044747C900”. Along with the common header we can assume there should be at least one ACI file referenced in the first part of the file. Because it is referenced as a filepath, the ACI string would be variable in its offset.

Adobe Acrobat Capture 3.0 turns out to be a different format. But looks familiar………

hexdump -C Contract.acd | head
00000000  50 4b 03 04 14 00 00 00  08 00 3b ba 6e 57 23 9d  |PK........;.nW#.|
00000010  8e b8 3d 00 00 00 3e 00  00 00 09 00 40 00 46 49  |..=...>.....@.FI|
00000020  4c 45 53 2e 4c 53 54 0a  00 20 00 00 00 00 00 00  |LES.LST.. ......|
00000030  00 00 00 80 e6 e9 ca 50  17 da 01 80 e6 e9 ca 50  |.......P.......P|
00000040  17 da 01 80 e6 e9 ca 50  17 da 01 4e 55 18 00 4e  |.......P...NU..N|
00000050  55 43 58 09 00 46 00 49  00 4c 00 45 00 53 00 2e  |UCX..F.I.L.E.S..|
00000060  00 4c 00 53 00 54 00 8b  76 74 76 31 8c e5 e5 f2  |.L.S.T..vtv1....|
00000070  0c 76 f6 f7 0d f0 0f f6  0c 71 b5 0d 09 0a 75 e5  |.v.......q....u.|
00000080  e5 f2 0b f5 75 f3 f4 71  0d b6 35 e4 e5 02 31 fc  |....u..q..5...1.|
00000090  1c 7d 5d 0d 6d 9d f3 f3  4a 8a 12 93 4b f4 12 93  |.}].m...J...K...|

sf Contract.acd 
---
siegfried   : 1.10.1
scandate    : 2023-11-15T09:10:01-07:00
signature   : default.sig
created     : 2023-10-11T15:10:17-06:00
identifiers : 
  - name    : 'pronom'
    details : 'DROID_SignatureFile_V114.xml; container-signature-20230822.xml'
---
filename : 'Contract.acd'
filesize : 79002
modified : 2023-11-14T23:17:53-07:00
errors   : 
matches  :
  - ns      : 'pronom'
    id      : 'x-fmt/263'
    format  : 'ZIP Format'
    version : 
    mime    : 'application/zip'
    basis   : 'byte match at [[0 4] [78886 3] [78980 4]]'
    warning : 'extension mismatch'

Yep, its a zip container file. lets take a peek inside to see what it is composed of. 

7z l Contract.acd 
--
Path = Contract.acd
Type = zip
Physical Size = 79002

   Date      Time    Attr         Size   Compressed  Name
------------------- ----- ------------ ------------  ------------------------
2023-11-14 23:17:54 ....A           62           61  FILES.LST
2023-11-14 23:17:54 ....A          410          226  Contract.acd
2023-11-14 23:17:52 ....A       150213        78093  Contract.acp
------------------- ----- ------------ ------------  ------------------------
2023-11-14 23:17:54             150685        78380  3 files

The the Contract ACD file is like a nesting doll, an ACD within an ACD. Lets see what the ACD and ACP is made of.

hexdump -C Contract.acd | head
00000000  00 01 00 00 00 02 04 47  47 2d 01 9a 01 00 00 02  |.......GG-......|
00000010  00 00 00 02 00 01 01 00  00 00 01 00 00 00 04 04  |................|
00000020  00 00 00 09 00 57 69 6e  67 64 69 6e 67 73 05 00  |.....Wingdings..|
00000030  41 72 69 61 6c 0b 00 43  6f 75 72 69 65 72 20 4e  |Arial..Courier N|
00000040  65 77 0f 00 54 69 6d 65  73 20 4e 65 77 20 52 6f  |ew..Times New Ro|
00000050  6d 61 6e 05 01 00 00 00  02 00 00 00 78 01 00 00  |man.........x...|
00000060  0f 00 54 69 6d 65 73 20  4e 65 77 20 52 6f 6d 61  |..Times New Roma|
00000070  6e 00 00 00 20 0b 00 00  c0 0a 00 00 00 00 00 00  |n... ...........|
00000080  00 06 00 00 00 0f 00 54  69 6d 65 73 20 4e 65 77  |.......Times New|
00000090  20 52 6f 6d 61 6e 00 00  00 20 0c 00 00 00 0c 00  | Roman... ......|

hexdump -C Contract.acp | head
00000000  25 50 44 46 2d 31 2e 33  0d 25 e2 e3 cf d3 0d 0a  |%PDF-1.3.%......|
00000010  31 20 30 20 6f 62 6a 0d  3c 3c 20 0d 2f 54 79 70  |1 0 obj.<< ./Typ|
00000020  65 20 2f 43 61 74 61 6c  6f 67 20 0d 2f 50 61 67  |e /Catalog ./Pag|
00000030  65 73 20 32 20 30 20 52  20 0d 2f 53 74 72 75 63  |es 2 0 R ./Struc|
00000040  74 54 72 65 65 52 6f 6f  74 20 34 20 30 20 52 20  |tTreeRoot 4 0 R |
00000050  0d 2f 43 41 50 54 5f 49  6e 66 6f 20 3c 3c 20 2f  |./CAPT_Info << /|
00000060  56 20 33 30 31 20 2f 46  53 20 5b 20 28 57 69 6e  |V 301 /FS [ (Win|
00000070  67 64 69 6e 67 73 29 28  41 72 69 61 6c 29 28 43  |gdings)(Arial)(C|
00000080  6f 75 72 69 65 72 20 4e  65 77 29 28 54 69 6d 65  |ourier New)(Time|
00000090  73 20 4e 65 77 20 52 6f  6d 61 6e 29 5d 20 2f 4c  |s New Roman)] /L|

The ACD has some of the same hex values as the previous version, but with some extra bytes at the beginning and it looks like the ACP is a straight up PDF. But may have some interesting tags, like “CAPT_info”.

The problem we will face when trying to write a signature for this version of ACD is the container signature needs a static file name to reference, and it appears the name of the container is also the name of the ACD file within the container. So every file will be different. I wish there was a way in the PRONOM signature syntax to reference an extension and ignore the filename, but currently there no method to do this. The only thing inside the container which seems to be consistent is the file “FILES.LST”. So lets take a peek inside if it. 

hexdump -C FILES.LST | head
00000000  5b 41 43 44 31 5d 0d 0a  49 53 43 4f 4d 50 4f 53  |[ACD1]..ISCOMPOS|
00000010  49 54 45 3d 54 52 55 45  0d 0a 4e 55 4d 46 49 4c  |ITE=TRUE..NUMFIL|
00000020  45 53 3d 31 0d 0a 46 49  4c 45 4e 41 4d 45 31 3d  |ES=1..FILENAME1=|
00000030  43 6f 6e 74 72 61 63 74  2e 61 63 70 0d 0a        |Contract.acp..|

Ok, there seems to be some static information that is unique to the ACD format. I bet the string “[ACD1]” would be sufficient enough to make a solid signature.

This is a good format example of a limited amount of information on the file format used by a well known company which has become obsolete and disappeared. Take a look at my signatures, maybe you have some old ACD files you were unaware of!

TIFF

Posted on by Thor Leave a comment

Lets talk TIFF, or Tagged Image File Format. It is well documented and accepted by the community. The format has been around since 1986, first developed by Aldus as a image format for scanners. The TIFF format is now used worldwide as a preferred format for scanning and preservation of cultural heritage objects.

As amazing as the format is, there are a few features of the format which can be a preservation risk. I want to focus on three of those risks.

The Tagged Image File Format has a well known header:

A TIFF file begins with an 8-byte image file header, containing the following
information:
Bytes 0-1: The byte order used within the file. Legal values are:
“II” (4949.H) LSB (IBM)
“MM” (4D4D.H) MSB (Mac)
Bytes 2-3 An arbitrary but carefully chosen number (42).
Bytes 4-7 The offset (in bytes) of the first IFD.

Putting this poster of the TIFF structure in your office will impress your co-workers, guaranteed. Thanks Ange!

The three risks I have been pondering lately are:

  • Multiple IFD’s
  • Metadata
  • DNG format

TIFF version 6.0 was released in 1992 and is the most recent version. Although some vendors are free to add their own private tags. In 1995 Adobe added an addendum which added some additions for use with PageMaker.

One of the main features of the TIFF format is its ability to hold multiple pages. In Adobe’s words:

TIFF has always supported what amounts to a singly linked list of IFD’s in a single TIFF file, via the “next IFD pointer,” though most applications currently ignore any IFD beyond the first one. Probably the best use for a linked list of IFD’s is when you want to store multiple different but related images in the same file—a ‘burst’ of images from a camera, for example.

Adobe PageMaker® 6.0 TIFF Technical Notes

Take note of the highlighted text, software like Adobe Photoshop will ignore any IFD beyond the first one. Even worse, Photoshop won’t even mention there are additional IFD’s. I have used many document scanners which default to multipage TIFF capture and have lost pages because of this. Because of this I have always built my workflows around single page TIFF’s for all scanning and we check against this as a rule.

What also makes this hard is how some capture software uses additional IFD’s. CaptureOne is a popular imaging software used by photographers and cultural heritage institutions. We have used it to connect to our PhaseOne cameras for capture of books and other flat objects. By default the software exports the final TIFF image with a thumbnail.

With the “No Thumbnail” unchecked we get this TIFF structure:

identify _MG_0193.tif 
_MG_0193.tif[0] TIFF 3456x5184 3456x5184+0+0 8-bit sRGB 51.3136MiB 0.030u 0:00.026
_MG_0193.tif[1] TIFF 107x160 107x160+0+0 8-bit sRGB 0.000u 0:00.007

 <IFD0:ImageWidth>3456</IFD0:ImageWidth>
 <IFD0:ImageHeight>5184</IFD0:ImageHeight>
 <IFD1:SubfileType>Reduced-resolution image</IFD1:SubfileType>
 <IFD1:ImageWidth>107</IFD1:ImageWidth>
 <IFD1:ImageHeight>160</IFD1:ImageHeight>
 <IFD1:BitsPerSample>8 8 8</IFD1:BitsPerSample>

So Imagemagick identifies two pages 0 and 1 with the second a much smaller resolution than the first. Exiftool reports back IFD0 and IFD1 with IFD1 having a SubfileType of a Reduced-resolution image. Makes sense, it is a thumbnail. In looking at the specifications for TIFF 6.0, I can find no mention of the word “thumbnail”, but the specification does make mention of “reduced resolution” images:

If multiple subfiles are written, the first one must be the full-resolution image. Subsequent images, such as reduced-resolution images, may be in any order in the TIFF file.

The specification also gives us this warning:

TIFF readers must be prepared for multiple images (subfiles) per TIFF file, although they are not required to do anything with images after the first one.

Scary to think about how a reader is not required to do anything, not even warn against multiple IFD’s (Subfiles).

The EXIF specifications seem to expand on this through attributes:

Attribute information can be recorded in 2 IFDs (0th IFD, 1st IFD) following the TIFF structure, including the File Header. The 0th IFD records compressed image attributes (the image itself). The 1st IFD may be used for thumbnail images. 

Page 97 of EXIF Specification

Take a look at the information and Figure 6 on page 21-22 in the EXIF specification.

Adobe early on decided to use their own tags for thumbnail data. Since Photoshop 5, Adobe has stored the thumbnail in Tag 1036. 

 1036 Photoshop Thumbnail             : (Binary data 4625 bytes, use -b option to extract)

There is another TIFF structure sometimes used in older FAX compressed multipage TIFFs and now used in the DNG Specification. The SubIFD tag was writable using the libtiff “thumbnail” tool, but is now depreciated. Originally described in the TIFF/EP specification, DNG files use SubIFD trees.

DNG files are often talked about in the same way TIFF files are, and many tools handle both seamlessly. One of the major differences is that DNG files switch their IFD use. IFD0 is often the reduced-resolution thumbnail and SubIFD the full-resolution image.

<IFD0:SubfileType>Reduced-resolution image</IFD0:SubfileType>
<IFD0:ImageWidth>256</IFD0:ImageWidth>
<IFD0:ImageHeight>171</IFD0:ImageHeight> 

<SubIFD:SubfileType>Full-resolution image</SubIFD:SubfileType>
<SubIFD:ImageWidth>3516</SubIFD:ImageWidth>
<SubIFD:ImageHeight>2328</SubIFD:ImageHeight>

This can cause issues when trying to extract technical metadata from images, knowing which IFD to get the main image details requires a bit of work. I’ll save DNG for another blog post.

TIFF Metadata is a vital part of preservation. The metadata can provide technical properties of the file along with some descriptive information. It amazes me how much the embedded metadata can vary from a scanner or camera capture device. The digitization lab I worked in for years had scanners from Epson, Fujitu, Canon and others. Along with cameras made by Canon, PhaseOne, and Copibooks. Each one with a vastly different set of metadata using different standards. Even when each workflow produced final uncompressed TIFF images, they all varied in metadata.

The TIFF images with the leasT amount of metadata was from the Epson scanners. When using the free Epson Scan software, not a single metadata field was embedded, no dates, scanner model or manufacturer. More was embedded when you used the Silverfast professional software included with each Epson, but even then if you didn’t add any IPTC fields, the metadata was limited.

The most metadata came from the camera systems, especially the PhaseOne/CaptureOne systems. Even though it produced the most and had valuable properties, there were some issues. I already discussed the thumbnail issue, but PhaseOne decided they wanted to change how some of the tags were used.

CaptureOne has quite the list of white balance options. Which is great for the photographer, but not so great for adhering to the TIFF standard.

According to the EXIF TIFF Specification, there are only two values allowed for White Balance, Auto or Manual. A CaptureOne produced TIFF will have this value if Auto or Manual are not chosen:

41987 White Balance                   : Unknown (5)
37384 Light Source                    : Other

The different lighting situations should be identified using the “Light Source” 37384 tag, but alas they chose to add to white balance instead. When I asked about this, they responded that they requested this update to the TIFF spec, but they weren’t willing so they took matters into their own hands. You can read the conversation on the JHOVE issues page.

The TIFF format is very accepted in the Cultural Heritage community as a preferred preservation format. The specification is well understood and documented. I just hope we can continue to openly discuss issues that arise in preservation which add risk to our collections. Some issues are minor compared to others. Sometimes it’s the tools we use to validate formats like TIFF which are wrong and need to be corrected. The talk more about these issues and how to manage them.

Hemera Photo-Object

Posted on by Thor Leave a comment

Many years ago I dabbled in a little Graphic Design. Working for a commercial printer in the Pre-Press area, I was very familiar with all things graphics, but never had a great talent for design, especially drawing. I often needed the random clip art for a design I was working on, so I purchased the Hemera, The Big Box of Art, probably from my local CompUSA if that dates me.

Hemera Big Box of Art

The cool thing about clip art from Hemera is it was not your usual JPG or TIFF format, it was in a special Photo-Object format. This format included the raster image, but also included a mask or alpha channel for the main object. They marketed this format as an alternative to the sometimes larger formats of the day. GIF files didn’t have the color depth and PNG was new enough, Hemera was probably hoping this format would be the next greatest thing to happen to clip art.

A Hemera Photo-Object has the extension HPI. Lets take a closer look at a file and see what is under the hood. I pulled this file from Disc 1 on Archive.org

The HPI file has a unique header which should make identification really easy. But what do we see starting at offset 32? A JFIF! Just after a 32 byte header the file has a standard JPG file hidden inside. Now a standard JPG file does not have the ability to support an alpha channel so there must be something else they have within to mask this file. Lets look for the EOF file marker for the JPG format.

Well, well, well. It appears the JPG file is then followed by a standard PNG! Sneaky. The entire HPI file is a 32 byte HPI header, a JPG, followed by a PNG. One could easily carve out each of the formats and save as separate files if needed. There is a script you can use to do this for you, written by Ed Halley. The original Hemera software won’t run on modern systems.

Hemera had a good run for about 10 years before selling off their assets in 2004 to another stock image company. At one point Hemera even purchased the rights to all of Corel’s Premium photo library which I covered in my article about the Kodak PhotoCD format.

Image PAC Files

Posted on by Thor 3 Comments

I wouldn’t be surprised if you have never heard of an Image PAC file. You may know it by the more common name Kodak Photo CD Image. Kodak’s PhotoCD format actually refers to the system and Disc format used to store images for compatibility with other hardware. The Kodak PhotoCD format was pretty advanced for its time, it original purpose was to store scanned 35mm film to a disc which was playable on computers and other hardware. In fact, because it was meant to store 35mm rolls as they were scanned it was the first use of the linked Multi-session CD format made standard by the orange book specification. The format was widely adopted at first, but eventually lost favor and was abandoned by 2004.

The Kodak PhotoCD format was also used on many commercial CD-ROM products. One example was the Corel Professional CD series. Below is a photo of a case of 200 CD’s I recently acquired. Each has around a hundred PCD images and viewing software on disc. Most discs can be viewed here. Or you can view their “Sampler” CD-ROM.

The actual PCD image file format was referred to as an Image PAC File. The format was unique in the fact it has multiple resolutions built into a single file. It also stored the raster data in a format called Photo YCC color encoding metric, developed by Kodak. This requires conversion to RGB for many uses. Adobe Photoshop for many years had an import filter for the format built in which included ICC profiles for properly converting the source to a destination colorspace, but support was dropped in CS3 of their products.

Photoshop Kodak PCD import

The Image PAC PCD format was a proprietary format which Kodak protected aggressively, even to the point of threatening legal action to those who attempted to reverse engineer the format. This frustrated developers and was probably part of the reason the format was abandoned. Of course this didn’t deter some curious developers and was partially reversed engineered and is available in the NetPBM library formally knows as PBMPlus. The tool hpcdtoppm was developed to convert PCD to PBM.

The trick in preserving older obsolete formats is to find a way to first identify them, gather significant properties, then migrate to a modern format if appropriate with minimal loss of data. Luckily most PCD files have the ascii string “PCD_IPI” starting around offset 2048. This is basically how the PRONOM registry identifies the format and has assigned it fmt/211. Exiftool also supports the format in identifying some of the significant properties. 

ExifTool Version Number         : 12.62
File Name                       : 136009.PCD
Directory                       : /Users/thorsted/Desktop/blog/Kodak/PCD
File Size                       : 3.6 MB
File Modification Date/Time     : 2023:06:23 10:48:55-06:00
File Access Date/Time           : 2023:06:26 23:43:50-06:00
File Inode Change Date/Time     : 2023:06:27 11:18:38-06:00
File Permissions                : -rwx------
File Type                       : PCD
File Type Extension             : pcd
MIME Type                       : image/x-photo-cd
Specification Version           : 0.6
Authoring Software Release      : 3.0
Image Magnification Descriptor  : 1.0
Create Date                     : 1993:09:20 07:35:34-06:00
Image Medium                    : Color reversal
Product Type                    : 116/01 SPD 0064  #00
Scanner Vendor ID               : KODAK
Scanner Product ID              : FilmScanner 2000
Scanner Firmware Version        : 2.21
Scanner Firmware Date           : 
Scanner Serial Number           : 0296
Scanner Pixel Size              : 0b.30 micrometers
Image Workstation Make          : Eastman Kodak
Character Set                   : 95 characters ISO 646
Photo Finisher Name             : HADWEN GRAPHICS
Scene Balance Algorithm Revision: 3.1
Scene Balance Algorithm Command : Neutral SBA On, Color SBA On
Scene Balance Algorithm Film ID : Unknown (131)
Copyright Status                : Restrictions apply
Copyright File Name             : RIGHTS.USE
Orientation                     : Horizontal (normal)
Image Width                     : 3072
Image Height                    : 2048
Compression Class               : Class 1 - 35mm film; Pictoral hard copy
Image Size                      : 3072x2048
Megapixels                      : 6.3

Exiftool is able to gather much of the important properties including an original creation date and the pixel dimensions. It would be nice if was able to mention each of the resolution options as some later Pro versions of PCD had a 64 base for resolutions of 4096 x 6144.

Migration to a more modern open format is a common preservation strategy. The National Archives and Records Administration has the format NF00224 listed as needing to migrate to JPG, while others prefer migration to TIFF. Others have learned valuable lessons attempting to find the right method for migration. There is a right way and a wrong way as the Center for Digital Archaeology learned. The easiest method is to use the popular ImageMagick command-line tool. 

thorsted$ identify 136009.PCD 
136009.PCD PCD 768x512 768x512+0+0 8-bit YCC 3.44727MiB 0.020u 0:00.006
thorsted$ convert 136009.PCD[5] -colorspace sRGB +compress 136009.tif
thorsted$ identify 136009.tif
136009.tif TIFF 3072x2048 3072x2048+0+0 8-bit sRGB 18.0004MiB 0.000u 0:00.000

ImageMagick along with most other tools like IrfranView and XnView only see the base resolution of 768 x 512, but with an extra little addition to the command by adding “[5]” after the filename if forces the conversion to use the “Fifth” 16 Base resolution which is the highest resolution on most PCD files, the Pro versions may have higher. The other issue is the colorspace conversion. It is known there could be a loss of highlights. This webpage illustrates different tools and the issues with highlights. You can see the difference if I use -colorspace RGB instead of sRGB.

ImageMagick conversion using RGB vs sRGB colorspace setting.

Other tools such as the open source pcdtojpeg and paid pcdMagic both work well, but the only tool I have tested so far which keeps the original metadata is pcdMagic. 

ExifTool Version Number         : 12.62
File Name                       : 136009_1.tif
Directory                       : .
File Size                       : 38 MB
File Modification Date/Time     : 2023:06:27 12:06:26-06:00
File Access Date/Time           : 2023:06:27 12:06:29-06:00
File Inode Change Date/Time     : 2023:06:27 12:06:27-06:00
File Permissions                : -rw-r--r--
File Type                       : TIFF
File Type Extension             : tif
MIME Type                       : image/tiff
Exif Byte Order                 : Little-endian (Intel, II)
Subfile Type                    : Full-resolution image
Image Width                     : 3072
Image Height                    : 2048
Bits Per Sample                 : 16 16 16
Compression                     : Uncompressed
Photometric Interpretation      : RGB
Image Description               : color reversal: Unknown film. SBA settings neutral SBA on, color SBA on
Make                            : KODAK
Camera Model Name               : FilmScanner 2000
Strip Offsets                   : 1622
Samples Per Pixel               : 3
Rows Per Strip                  : 2048
Strip Byte Counts               : 37748736
Planar Configuration            : Chunky
Software                        : pcdMagic V1.4.19
Modify Date                     : 2023:06:27 12:06:26
Copyright                       : Copyright restrictions apply - see copyright file on original CD-ROM for details
Exif Version                    : 0231
Date/Time Original              : 1993:09:20 07:35:34
Create Date                     : 1993:09:20 07:35:34
Offset Time                     : -06:00
User Comment                    : color reversal: Unknown film. SBA settings neutral SBA on, color SBA on
Color Space                     : Uncalibrated
File Source                     : Film Scanner
Profile CMM Type                : Unknown (KCMS)
Profile Version                 : 2.1.0
Profile Class                   : Display Device Profile
Color Space Data                : RGB
Profile Connection Space        : XYZ
Profile Date Time               : 1998:12:01 18:58:21
Profile File Signature          : acsp
Primary Platform                : Microsoft Corporation
CMM Flags                       : Not Embedded, Independent
Device Manufacturer             : Kodak
Device Model                    : ROMM
Device Attributes               : Reflective, Glossy, Positive, Color
Rendering Intent                : Perceptual
Connection Space Illuminant     : 0.9642 1 0.82487
Profile Creator                 : Kodak
Profile ID                      : 0
Profile Copyright               : Copyright (c) Eastman Kodak Company, 1999, all rights reserved.
Profile Description             : ProPhoto RGB
Media White Point               : 0.9642 1 0.82489
Red Tone Reproduction Curve     : (Binary data 14 bytes, use -b option to extract)
Green Tone Reproduction Curve   : (Binary data 14 bytes, use -b option to extract)
Blue Tone Reproduction Curve    : (Binary data 14 bytes, use -b option to extract)
Red Matrix Column               : 0.79767 0.28804 0
Green Matrix Column             : 0.13519 0.71188 0
Blue Matrix Column              : 0.03134 9e-05 0.82491
Device Mfg Desc                 : KODAK
Device Model Desc               : Reference Output Medium Metric(ROMM)
Make And Model                  : (Binary data 40 bytes, use -b option to extract)
Image Size                      : 3072x2048
Megapixels                      : 6.3
Modify Date                     : 2023:06:27 12:06:26-06:00

There is a way to convert the PCD to TIF using ImageMagick, then using Exiftool to map some of the metadata over to the new TIFF file. It would look something like this:

exiftool -addtagsfromfile 136009.PCD '-EXIF:DateTimeOriginal<PhotoCD:CreateDate' '-EXIF:CreateDate<PhotoCD:CreateDate' '-ExifIFD:SerialNumber<PhotoCD:ScannerSerialNumber' '-ExifIFD:ExifImageWidth<PhotoCD:ImageWidth' '-ExifIFD:ExifImageHeight<PhotoCD:ImageHeight' '-IFD0:Make<PhotoCD:ScannerVendorID' '-IFD0:Model<PhotoCD:ScannerProductID' '-IFD0:Orientation<PhotoCD:Orientation' '-IFD0:Copyright<PhotoCD:CopyrightStatus' 136009.tif