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: Size of scan files

More information in the higher ISO film. LZW is a near lossless 
compression. With the lower ISO film, you had less information. And, 
not necessarily grain information in the 800 film, you could well 
have greater gamut/saturation, greater light latitude in the 800 
film. The Fuji 800 films are now pretty darned good on grain size. 
That's why you used the higher speed film, right? To get that 

Compression squeezes out redundant information. If you have valid 
"different" bit data in adjacent pixel locations, a compression 
scheme can't device a method of recreating that unique difference on 
expansion. Unless you tell it to ignore a certain level of 
difference, such as in JPEG. LZW won't let you do that.

This is not a case of "bad" compression. Bad compression is where the 
algorithm assumes that - say - five pixels are the same, when they 
really aren't. Then, on expansion, these five pixels show up as the 
same. That is how you get smaller file sizes. You loose information.

>I was just going through the files on the computer to see what I could
>easily archive to CDR and noticed a huge difference in file sizes.  I have
>some full frame scans from Fuji 800 print film that occupy 30MB as 8bit LZW
>TIFF files, yet I am going through Provia 100F scans at the moment which
>start out at 36MB 16bit from Vuescan and end up at about 15MB in 8bit.
>Logically you might expect this from halving the number of bits, but bear in
>mind that my LS30 only produces 10 bits per channel. :)  It looks like the
>excessive grain of the fast film compresses very poorly, while the almost
>non-existent grain of Provia 100F compresses very well.  This also makes
>sense, but I just hadn't noticed before now how huge the difference could
>PS Yes I know about things like run length compression and why it doesn't
>work very well when there is a lot of detail or noise. :)


Copyright © Lexa Software, 1996-2009.