Replies

I can think of no reasonable way for the "R" to be converted to Color with larger pixels, and that the rest of the characters would all be binary with smaller pixels, EXCEPT for the explanation I have offered.

The other explanation is that the 'R' on the original was fainter than the other letters, so it wasn't black/dark enough for the OCR or optimization routine to "recognize" it as a black letter. It would therefore leave it as part of the background rather than extract it like the other letters. I've read the graphics experts' arguments on both sides of whether that's possible, and the ones that say it is are at least as persuasive as the ones that say it isn't. Your explanation requires me to believe that it was so hard to find the letters 'B', 'A', 'R', 'C', and 'K' in other documents that they had to use multiple sources in different formats. I'm sure those letters would have been available on other birth certificates all digitized the same way at the same time--why would they not just use those? They found an all-black, small-pixel 'R' for the father's name--why not just use that one?

Your theory provokes an "oh come on, nobody would do it that way" response in me, while the computer processing theory get a "hmm, maybe." So far #2 wins.

Your explanation requires me to believe that it was so hard to find the letters 'B', 'A', 'R', 'C', and 'K' in other documents that they had to use multiple sources in different formats. I'm sure those letters would have been available on other birth certificates all digitized the same way at the same time--why would they not just use those? They found an all-black, small-pixel 'R' for the father's name--why not just use that one?

What did Sherlock Holmes say? "Once you've eliminated the impossible, whatever left, no matter how improbable, must be the truth."

I don't think you understand the significance of both the different sized pixels (impossible on a single original document) and the difference between binary bitmap and color bitmap. In computer graphics, whatever is being rendered to MUST ALL HAVE THE SAME PIXEL SIZE AND FORMAT. Meaning the Destination file is a specific defined format and pixel size. The variations we see can ONLY be caused by taking a binary image file, and converting it to the destination format while still retaining it's tell tale characteristics as a binary format.

Let me try to make this clear. All modern software will render in something like a 1024X768 (or better) 24bit or 32bit Color format. We don't do low resolution or black and white anymore. The destination file on a modern computer does not use binary images. It uses the above mentioned format, but it can easily represent binary images using this format. It simply creates Black by turning all the RGB components of the pixel word to 0. In a 24 bit word the bits would look like this: R=00000000 G=00000000 B=00000000 To produce white, the bits would look like this: R=11111111 G=11111111 B=11111111. Thus an original black and white (binary) image file can be represented by a high resolution color display system. (Other colors are created by varying the binary value of the color component intensity.)

Depending on the resolution, each black pixel from the original can be represented by some number of finer pixels on modern graphic displays, say 4 fine resolution pixels equal one original pixel. Or it can be 1 to 1. (but that would probably make it too small to read nowadays.)

What makes it obvious to anyone that is familiar with how graphics displays work is that in the letter "R" the pixels are FOUR TIMES LARGER than the pixels in the other characters. Sure, they can be reproduced by the graphics system, but to do so it required 4 times as many pixels. This makes it dead certain that the letter "R" was stored with a different original resolution than the rest of the letters. Take a look at this image to see what I mean. (Note the difference in size circled in red.)

Your theory provokes an "oh come on, nobody would do it that way" response in me, while the computer processing theory get a "hmm, maybe." So far #2 wins.

I think that's only because you don't understand theory # 1 that well. I have done a LOT of Graphics programing. That particular image of the word "Barack" is sourced from different original formats combined onto a standard destination format. That is the ONLY reason of which I can conceive for how it could possibly have different pixel size and color data information.

If you want to learn more, download Microsoft's DirectX Software Developers kit and read the information in the tutorial or help file.