Re: re[2]: filmscanners: OT (slightly): Epson 640U

["Richard N. Moyer" <dickmoyer@mail.earthlink.net>]

If you use Ektachrome E6, you will obtain saturated colors far 
exceeding the gamut of sRGB. For this media, you should use a wider 
color space such as Ekta Space which encompasses (just barely) the 
gamut of E6. In addition, you should not save these files using the 
sRGB profile, or they will be clipped, meaning that space compression 
attributes such as greyscale distortions, and lost tonality - meaning 
loss of saturation - will take place irreversibly. Unless you rescan. 
If you are concerned that you haven't captured saturated colors in 
the first place, you can drop your scanned slide image file (making 
sure it is not scanned into sRGB onto (drag and drop) Chromix's 
ColorThink gamut analysis software (Mac only). This will then give 
you (in any coordinate system you want L.a.b., xyz, luv, etc.) the 
gamut latitude of your photograph image and compare that gamut with 
gamuts of sRGB, EktaSpace, AdobeRGB, or KodakProRGB.

The reason these (gamut conversions) changes are irreversible is 
because the compression/expansion algorithms and transforms are not 
perfect and because the repsective gamuts are highly irregular three 
dimensional "globs". Stuffing one glob into another - or expanding 
one in another causes non linear distortions.

sRGB matches the gamut of most "good" monitors, and was chosen as a 
default by Microsoft Windows, and other PC software. But it should be 
emphasized that this is a restricting space and it will (potentially) 
clip your work. Once your image is in sRGB, it cannot be "widened" to 
a larger space without again introducing gamut expansion artifacts, 
such as posterazation, and more greyscale (crossover) distortions. 
Think carefully about color spaces, and plan your workflow 
accordingly!




>In a message dated 3/3/2001 1:31:55 PM EST, arwbackup@worldnet.att.net writes:
>
>>  This is after vuescan gets it and does a conversion from the sRGB provided
>>  from the scanner.  Unless Ed has changed it and not noted this it the
>release
>>  notes he only brings back the values from the scanner under sRGB.
>
>Yes, VueScan transfers the raw samples as linear samples
>(i.e. not gamma corrected) using the same color primaries as
>sRGB.  If you do a scan of a Q60 calibration slide using this scanner,
>you'll see that only the saturated yellows are slightly out of gamut
>and some of the saturated cyans are slightly out of gamut.  Since
>these colors occur rarely (if ever) in real photographs, this isn't
>really a problem.
>
>Regards,
>Ed Hamrick