ANOTHER CODE PAGE?

ANOTHER CODE PAGE EMERGES?

Over the years there have been a number of  'Code Pages' either as 'new scans' or, as in this new case, a 'better quality' image. At the request of one of our followers, I have examined this particular document to get a better idea as to where it may have come from and what process it may have been put through. The image above is a copy of this 'new' code page.

The page turned up some time ago on a UK based bloggers site, in fact, I reviewed it at the time. It has since been reawakened and hence this post.

Origin of the page

This version of the code page came from the site of Marcel Varallo, a Melbourne based IT engineer. It is interesting to note that Marcel makes no claims about the version in his post, any claims that have been made are by others, Marcel has only 2 post on the subject, October 2019 and an earlier one in February 2017. To be fair, he had spent some time on his posts and my comments here are in no way a reflection on Marcel, my focus is the document itself which has been copied and commented on by others.

The original page
Some weeks ago I wrote a post 'What do we know about the code page', the page I displayed in that post was the one downloaded from the Adelaide University Wiki who in turn, had obtained it from the Adelaide Advertiser. Gerry Feltus also obtained his copy from the Adelaide Advertiser and both his and the University copies are identical. Both of these have the same file size of 631 kb.

The Metadata
In that previous post, I published the metadata related to the Wiki version, this data tells the story of the image as shown below:

As you can see it's quite an extensive set of information.

The Verallo Version
I have put Marcel's image through the same analysis process and below is the Metadata for that image:

I would draw your attention to some key facts:

1. The Wiki version shows a resolution of 400 DPI whilst Marcelo's version shows a resolution of 96 DPI.  (The DPI is Dots per inch which is the print resolution, not the resolution of the image)

2. The Wiki version shows that the image has 2.6 million megapixels, Marcelo's version shows 4.6 million pixels, an additional 2 million pixels.

3. The Wiki Version shows that the image has 8 bits per sample, Marcelo's version does not show the sample rate but, an inspection of that image's properties shows that it has 24 bits per sample.

3. The original date of the Wiki version shows that it was created on 14th of June 2002 using Photoshop, this was the date, I believe, that the Advertiser first created the image from a hard copy photograph that had been in their possession since 1949.

4. Marcelo's version does not show the date of creation but, it does show that it was created using Paint.net Version 3.5.10 which was released on October 9th, 2011. In other words, an old version of Paint.net was used to create Marcelo's image.

5. The file size of Marcelo's version is 2.7 mbs which is 4 + times the size of the Wiki version file being 631 kb.

This is not to say Marcelo used that package to create it. When the Properties of the image are examined the file name refers to 'TRISH' and the date 16th April 2012 as the original owner and probable date of creation given that the version of Paint.net was still a current version at that time. (To examine 'properties' simply right click on the image file and select properties.)

The version attributed to Trish, has the same size ratio (Squareness) as the size of the image found on the Alamy stock image site with a creation date of 2009, thus predating the TRISH version by around 3 years and suggesting that is where it originally came from. I published a post about this version in September 2019. Click here to view the Alamy post.


Conclusion
What does all this mean? We can conclude:

1. That Marcelo's version measuring 2407 X 1900 pixels and at 96 DPI would print out at a physical size of 63.69 cms X 50.27 cms.

2. The original image from the Wiki prints out at 11.44 cms X  9.14 cms. This latter is the size of the original copy of the Rubaiyat of Omar Khayyam from which the code page came.

3. Marcelo's version is approximately 5.5 times the size of that original book.

4. The original version obtained by Marcelo had a ' bit depth' of 8 bits, yet the version as presented has a 'bit depth' of 24 bits. This is achieved by altering the 'quality' of a grayscale image within an image editing package by increasing the 'quality' to 100%. It gives the grayscale image the same number of bits as a colour image, 24 bits. In doing so, it adds in this case, 2 million additional, but 'empty', pixels.

5. The original Wiki image had 2.6 million pixels, each one of those pixels contained colour information, they each had a correct and original information value.

• The additional 2 million pixels are effectively 'empty' of any additional information, they add no new information to the original Wiki image. 
• When an image is inflated with an additional load of pixels, they tend to take on the attributes of the nearest existing pixels. 
• What that means is that if there is an original pixel with a colour black then the introduced pixel alongside it also takes on the colour black. 
• Here's where it gets really complex, these new pixels are not necessarily added evenly across the existing image, they can be 'randomly' distributed.
• The effect of this distribution can be to destroy any fine detail within the existing image but on some occasions, it can apparently sharpen some.
• An examination of the Marcelo version confirms the loss of sharp/fine details found in the original Wiki version. 
• In a number of areas on the altered image, it was observed that 'artifacts' were introduced, the dark 'stain' at the top right of the code page is one example.

The bottom line is that this version of the code page from Marcelo is not suitable for document examination and therefore cannot be recommended for that purpose.

IMPORTANT

If you are carrying out an examination of a document, in this case, a good quality photograph, then you need to print it out. It should be printed at the same physical size as the original document. This is a forensic document examination practice when the actual original document is not available.

When you print an image out it should be at the best quality setting of your printer and that printer should be capable of the same print resolution as the image. Many home printers would fall short of the 400 DPI resolution requirement of the code page.

The printer itself should be a good quality inkjet printer, the heat process involved in laser printing can lead to distortion of fine details and they generally are not able to give the best colour match having only 4 basic colours. My current inkjet printer, for example, has 6 colours but I will shortly be acquiring one with 9 colours.

There are significant differences between an image that is going to look good and one that is going to be used to reproduce as near as possible, an original document which is to be printed out and examined. Document examination of this type cannot be done on screen, it must be a physical examination. I will post more on that aspect in the near future.

In closing, Marcel, if you are reading this, please understand that this post is not meant to be a criticism of you. Document examination is a science all by itself. In much the same way as I would not understand the vagaries of Java, Python, Selenium, and others, I respectfully suggest that the techniques, tools, and processes that I use may, at the moment at least, be foreign to you. However, I am sure you would catch up quickly.