Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 




      :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: filmscanners: image samples of digital artifacts


> The old JPEG (not JPEG 2000) does code three channels
> Y, Cr, Cb. The channels Cr and Cb are downsampled.
> Then each channel is divided in blocks of 8x8. For
> each such block you do a Discret Cosinus Transform
> (DCT), devide each of the 64 resulting values by one
> of 64 numbers defined by the quantization table
> (higher frequency values are divided by higher numbers
> then low frequency values), and then Huffman
> (arithmetic coding is also possible but is less
> common) entropy encoded. This is true for lossy
> compression. Now if you do a high compression you
> divide the values after the DCT by higher factors so
> you get more 0s. Because of that the transition of one
> 8x8 block becomes less smooth and you see 8x8 block in
> the final image.

I guessed as much, but I had thought it was Guffman, not Huffman, and I
think I forgot to carry a 1.



Copyright © Lexa Software, 1996-2009.