More VueScan on 8.0.1 color trouble

[Ellestad]

Is anyone else having color difficulty with 8.0.1? I have been uing VueScan
for at least 4 years now but I haven't had this much trouble in the past.
I'd like to stick with VueScan (I have other scanning software including
SilverFast) because I think it offers more control but this kind of time
setback will turn me away.

I have 8.0.1 now scanning reasonably with negative, both color and b&w. At
first I had the color problems with negative as well but persistence
eventually put negative scanning into resonably normal results. I can find
nothing that I did that changed the scanning results, however, other than
that I kept trying.

Scanning transparencies is still yielding useless results, however. When
scanning 35mm Ektachrome E6 slides I get normal looking images from about
RGB 30 and up. But below that everything is crushed into black or a uniform
R,G or B variant of black. The image is also heavily blocked up into crude
pixilation. Is anyone else having anything else like this? Does it suggest
anything to look for or correct? Since the scanner works beautifully with
SilverFast and Polacolor Insight I have to believe that this isn't a
hardware issue.

I'm afraid that VueScan 8.0.1 has a major bug or two, at least when it comes
to my SprintScan 120 on FireWire. Having these kind of problems and not
being able to get very expedient help kind of suggests that VueScan isn't
suitable for professional use.

Thanks for any help.

Tim Ellestad

 

[Erik Krause]

Is anyone else having color difficulty with 8.0.1?

Me not. Did you try to delete or rename vuescan.ini? Did you try more
recent versions?

 

[Ellestad]

I think that this is the latest version.

I did try a re-install. I did delete the .ini file several times.

This is a new development from older versions.

Tim Ellestad

 

[Erik Krause]

I think that this is the latest version.

Most recent is 8.0.10. If you got this perhaps it is a good idea to
switch back to 8.0.9 or a bit earlier. I use 8.0.2 with good success.

I did try a re-install. I did delete the .ini file several times.

It would be helpful if you write your settings. Some days ago Ed posted
he had "Improved white balance" in version 8.0.10 and some days before
that he "Improved color correction" for "Input|Media" = "Slide" in
version 8.0.8. Could be that causes your problem...

This is a new development from older versions.

Don't understand.

BTW (just read your other posting): You can rename vuescan and the
folder where it resides as you like. It still runs, as long as you dont
rename it's support file vuescan.dat.

 

[Ellestad]

Most recent is 8.0.10. If you got this perhaps it is a good idea to
switch back to 8.0.9 or a bit earlier. I use 8.0.2 with good success.


It would be helpful if you write your settings. Some days ago Ed posted
he had "Improved white balance" in version 8.0.10 and some days before
that he "Improved color correction" for "Input|Media" = "Slide" in
version 8.0.8. Could be that causes your problem...

Don't understand.

I wasn't having this color trouble with older versions, although older
versions had a few problems of their own making it seemingly desirable to
have the latest version.

BTW (just read your other posting): You can rename vuescan and the
folder where it resides as you like. It still runs, as long as you dont
rename it's support file vuescan.dat.

After seeing this I went searching through the disk and cleaned out some
forgotten VueScan stuff from the past. No effect on my problem, however. As
I indicated before - color neg now scans usably and transparencies scan
about normal on values that end up above about RGB 40 but below that they
crudely block up into black colored near-black in coarse pixelation.

Erik - Thanks for your input on this. I'm open to any suggestions. I've
tried just about all settings and combinations that I can possibly come up
with to no avail. At this stage I have to suspect that this is a significant
bug with VueScan and my scanner. More as the plot thickens.

Tim

 

[Erik Krause]

Erik - Thanks for your input on this. I'm open to any suggestions.

Ok. Did you try advanced workflow? Did you use generic color negative?

Assuming your RGB exposure settings are ok scan to raw and check
whether the blacks (brightest for negative) are missing from the raw
scan, too (you can set gamma to 2.2 and then invert manually f.e. in
photoshop). If not, something is wrong with subsequent processing (most
likely film base color values too low or black point setting too high).
If yes, most likely RGB exposure is to high.

In the first case lock film base color and set values (on color tab)
manually to 1.0. Use color balance none for testing. If this helps you
definitely should use advanced workflow.

In the second case try locking the exposure and set it manually (lower
than automatic). If this helps, you should notify Ed of this behaviour.
Most likely there is a bug in the calculation of the RGB exposure.

If possible, compare the RGB exposure and film base color values with
older versions. If they differ much there's something wrong. If you
can't get a solution, use an older version until Ed fixes the bugs...

 

[Ellestad]

Ok. Did you try advanced workflow? Did you use generic color negative?

Yes, I've always used advanced workflow. The problem I'm describing is with
transparencies - in this case 35mm Ektachrome.

Assuming your RGB exposure settings are ok scan to raw and check
whether the blacks (brightest for negative) are missing from the raw
scan, too (you can set gamma to 2.2 and then invert manually f.e. in
photoshop). If not, something is wrong with subsequent processing (most
likely film base color values too low or black point setting too high).
If yes, most likely RGB exposure is to high.

Yeah, I tried everything here.

In the first case lock film base color and set values (on color tab)
manually to 1.0. Use color balance none for testing. If this helps you
definitely should use advanced workflow.

In the second case try locking the exposure and set it manually (lower
than automatic). If this helps, you should notify Ed of this behaviour.
Most likely there is a bug in the calculation of the RGB exposure.

Yes, the image manipulated normally above the values that are errant. The
areas of the image that show the problem are of such density that they
should come up at about RGB 30 to 40 but instead are being delivered at near
zero black or a one color variant of black in heavy pixelation (looks
"posterized"). These areas change very little even with gross changes of the
controls that you suggested above. In video production we would have to
adjust (and color balance) the pedestal setting to correct a picture
situation like this.

If possible, compare the RGB exposure and film base color values with
older versions. If they differ much there's something wrong. If you
can't get a solution, use an older version until Ed fixes the bugs...

Thanks once again for your interest. I appreciate the feedback from someone
so experienced with image control and VueScan.

Sincerely,

Tim Ellestad

 

[Erik Krause]

Yeah, I tried everything here.

You did not write anything about RAW files. Please check whether the
information you miss is present in the raw files. Please give more
details. If you write 'tried everything' I know nothing about your
results...

 

[Ellestad]

You did not write anything about RAW files. Please check whether the
information you miss is present in the raw files. Please give more
details. If you write 'tried everything' I know nothing about your
results...

Yes, I did try RAW files with color negative scanning and the same problems
persisted. I can get scans from color negs that I can work with by finding
combinations of RGB and gamma settings as you have suggested. I have been
unable to fix the problem in scans of Ektachrome with any combination of
color, level and gamma settings. The entire range of combinations of these
settings yields only a greater or lesser display of the problem as I
previously described. I have tried various selections of original media as
well with the same results. Changing white balance, color spaces, etc.
yielded notable changes in the color rendering but did not eliminate the
problem.

Thanks again for your continued interest. I appreciate the help.

Sincerely,

Tim Ellestad

 

[Erik Krause]

Yes, I did try RAW files with color negative scanning and the same problems
persisted.

Did you open the RAW files in your image processor and look whether
there is usable data where there is pure black in the processed files?

If you open RAW files you must set gamma in levels dialog to 2.2 (or
higher) as a first step after opening or use a Gamma 1.0 color space
like suggested on http://www.aim-dtp.net/aim/techniques/index.htm

 

[Ellestad]

Did you open the RAW files in your image processor and look whether
there is usable data where there is pure black in the processed files?

Yes, I explored the RAW files in Photoshop but there was nothing there in
the overly black areas to bring back.

If you open RAW files you must set gamma in levels dialog to 2.2 (or
higher) as a first step after opening or use a Gamma 1.0 color space

Yes, I changed Gamma as part my attempts to solve the problem. My screen is
calibrated via Optical and Spyder with Gamma set to 1.8. My screen display
and performance is consistent with my service bureau.

like suggested on http://www.aim-dtp.net/aim/techniques/index.htm

Thanks again for your continued effort. I guess I'll take this problem to Ed
Hamrick as per his specified submission format.