Vuescan version 8.3.32: "Improved profiling of IT8 calibration targets"

[Mendel Leisk]

Vuescan version 8.3.32 is indicated as having: "Improved profiling of
IT8 calibration targets"

I've tried the new version, doing scan-from-disk with the same 16bit
linear scan of an IT8 target. My old and new histograms were
*identical*, with all variables/settings unchanged, down to the 1/100
of unit for "Mean" "Std Deviation" and "Mean", Photoshop CS2.

That was actually a little suprising, considering I reposition the IT8
crop matrix by hand (that is one more variable I suppose I could have
eliminated, if I had noted the numbers in the crop tab). I hid the old
profile prior to the calibration process, so I'm sure I've not mixed up
the old and new profiles.

At any rate, I can't see *any* difference. Anyone else tried, and/or
have any idea what else "improved" may stand for?

 

[Ralf R. Radermacher]

Vuescan version 8.3.32 is indicated as having: "Improved profiling of
IT8 calibration targets"

I've tried the new version, doing scan-from-disk with the same 16bit
linear scan of an IT8 target. My old and new histograms were
*identical*, with all variables/settings unchanged, down to the 1/100
of unit for "Mean" "Std Deviation" and "Mean", Photoshop CS2.

Must be the same algorithm as the one used for dramatically improving
the IR cleaning function with Minolta scanners. ;-)

SCNR,
Ralf

 

[Don]

Vuescan version 8.3.32 is indicated as having: "Improved profiling of
IT8 calibration targets"

I've tried the new version, doing scan-from-disk with the same 16bit
linear scan of an IT8 target. My old and new histograms were
*identical*, with all variables/settings unchanged, down to the 1/100
of unit for "Mean" "Std Deviation" and "Mean", Photoshop CS2.

It's been clear for a long time that various VueScan's claims are just
hot air.

Starting with the 2 years of false claims that VueScan works with the
5400 to recent IR cleaning "improvements" etc.

Apparently with each new release (in order to try and obfuscate all
the bugs) the author randomly chooses a feature and declares it
"improved". :-/

Don.

 

[DenverDad]

I haven't tried the latest version to compare profiles myself. But I
seem to recall that people had observed some bugs with the
implementation of the profiling. Something about the profiles not
"taking" the first time through, or maybe it depended on whether or not
one did a preview scan first? I don't remember for sure because I had
not experienced these issues myself. But, perhaps the "improvements"
were more about overcoming some bugs as opposed to improved
quality/accuracy of the profiles. Or alternatively, maybe one of
those bugs is still there and no new profile files(s) were actually
written - so you're really working with your old profiles.

Just speculating.

Jeff

 

[Roger S.]

I've tried the latest version and profiling works fine (it does output
a new .icc file). I can't see any improvement, but "works fine" is
good enough for me.
Now, let's see if 8.3.33 fixes the annoying but where the cropping
changes between preview and the scan (gotta love waiting for a 4000dpi
scan and then having a side lopped off).

I'm also curious to see if the fixed IR exposure does anything. Before
I had to manually bump up IR cleaning from 1 or 2 to 3 to work right
for negs (slides work perfectly).

 

[Mendel Leisk]

Thanks, all!

I did verify by checking the date that a new icc profile was created. I
gather whatever the change may be, it is not reflected in the resulting
color balance

 

[Roger S.]

Spoke too soon, now I have to test 8.3.34 as well:

What's new in version 8.3.34

Improved cropping when using infrared cleaning