Posted By:
DaveI tried using the border size to determine the card size. I used Bender, pitching, trees in background as my reference card. I was also using scans done at 2x, 200 dpi, which is about 500 kB per picture (very high resolution). I did this on my computer screen with an engraved steel engineering ruler, so the measurements are good but lucky to be within .02 inches.
Using this I came up with nominal size of the border as 1.261 by 2.401 inches. I then used this as a yardstick against Tris Speaker, and came up with a card size of 1.422 x 2.615. The actual size of this card is 1.437 x 2.630. So, I was off by about 0.015 in each dimension, low in each case.
I actually think this is good given that there are 4 measurments involved in each computation, and each could have been off this much.
I wouldn't use the border values here as the gospel; to do this right needs optical magnificaion on real cards and a more accurate ruler, like the calipers. And, multiple cards to validate unformity. But, I think shows it is possible to use.
Trying this on low res scans really would be useless.
