[Doug Nelson]

I actually understand very little of this, but it's fascinating just the same
http://blog.wired.com/27bstroke6/200...chers-ana.html

[Swampy]

Very interesting indeed. I have become very sceptical about the aunthenticity of news photos since the editing that was done by Sports Illustrated many years ago at the time of the Olympics (perhaps Barcelona?). It's been a long time, but it was the women's relay race where one of the USA track "stars" fell in the latter portion of her relay lap. The photo, as originally shot, showed the baton projecting from between her legs like a huge penis. The version that ran in SI had the baton edited out. SI did run a retraction explaining that they felt the photo "was less offensive" as edited.

[Jafo]

I read the article and love the idea. The article links to the source code for the program that does the analasys:

http://blog.wired.com/27bstroke6/files/jpegquality.txt

But, sadly, I don't know how to compile it.
This is a general request either for the information of how to compile this, or for a version compiled for WinXP platform.

[Jafo]

Well, I got the program to compile, but all it really does it print out quantization tables for both luminance and chrominance. Then it averages the two error percentages and gives the user an overall estimation of the quality in percent.

Not sure how the user is supposed to use that information, really.

#File: d3f2f66e9a7c6da98602f782e425ae2d63e608a6.jpg

Quantization table
Precision=0; Table index=0 (luminance)
6 4 4 4 5 4 6 5
5 6 9 6 5 6 9 11
8 6 6 8 11 12 10 10
11 10 10 12 16 12 12 12
12 12 12 16 12 12 12 12
12 12 12 12 12 12 12 12
12 12 12 12 12 12 12 12
12 12 12 12 12 12 12 12
Estimated quality level = 89.79%
Precision=0; Table index=1 (chrominance)
7 7 7 13 12 13 24 16
16 24 20 14 14 14 20 20
14 14 14 14 20 17 12 12
12 12 12 17 17 12 12 12
12 12 12 17 12 12 12 12
12 12 12 12 12 12 12 12
12 12 12 12 12 12 12 12
12 12 12 12 12 12 12 12
Estimated quality level = 86.57%
Average quality: 90.80% (91%)