Mike Chaney's Tech Corner

I acquired Profile Prism a while back as a free download with my last QIU sub.
The CANON 8400F scanner I have is underexposing with the histrogram clipped on the rh side as others with this scanner have reported.
Using VueScan Pro and following instructions.
I have a Color Munki from which I get good results, but have been busting to try PP.
Used a dark but not black A4 size backing. Could that be a reason for the underexposed scans?
Has anyone been successful with the CANON 8400F??
Are the newer LiDE scanners as capable as the older ones for this task??
When the image loads I answer No because I'm profiling the scanner. That is correct??
See the screenshot.

Thanks
John

Hi John,
I have produced a profile for my 8400F scanner in the past also for my newer and cheap Lide 110.
I've looked back at the images and they are dark with a similar histogram to yours.
PP gave error messages but the profiles, when used, seemed to produce good results in both cases.

Yes.

I assume you are producing a 48bit "raw" image from Vuescan.

Terry

Yes 48 bit Terry.
Does the 110 do a better job??
I've located one secondhand.

John

As it's an LED scanner, it has some advantages wrt colour Metamerism I believe.
Terry

It seems to me that up to the point where I'm at, the lighting variance across the chart and the poor shadow detail are the biggest hurdles to overcome with the 8400F.
Is that a failing of the scanner or my methods, or both?
In the past another 8400F user posted a histogram and report near identical to mine.


John

Terry, I purchased the LiDe 110 this morning.
I have a black sheet that occupies the entire scanner window.
First thing first. Profiling the scanner.
When I bring the 48bit scanned with 5x passes IT8 target into PP, it seems obvious to answer No. See the attached.
The same histogram jammed on the dark side. The familiar light variance report as with the 8400F.
Reading the instructions for the Canon scanner in Help (as copied below) it says to say Yes. See the second attachment.
Are my settings correct in both attachments??
The Yes histogram is encouraging, but the report isn't.
Can you pick where I'm going wrong?

Regards,

John

Canon Scanners

Included Canon software usually not sufficient to get good scans without clipping
Specialized scanning software such as VueScan recommended
Recommended VueScan settings:

First, click "Scanner", "Calibrate" with scanner lid closed, white pad in place
On the "Input" tab, select "48 bit RGB" for "Bits per pixel"
On the "Input" tab, select "300 dpi" for "Scan resolution"
On the "Output" tab, select "Output raw file"
UNcheck Output TIFF, JPEG, and index file so only the raw file is saved
UNcheck "Raw compression" on the same tab
Set "Raw file type" to "48 bit RGB" on the same tab
Scan
Open scan in Profile Prism and reply "Yes" to the linear 48 bit scan prompt
Proceed with profiling steps

This one for the Answering No attachment.
Don't know why it didn't go with the previous post.

John

Hi John,
First, just to clarify re.
You do not need to profile the scanner for making print profiles as a "raw" file is used from Vuescan.
Your results look similar to mine. I did check some other documentation I have which seems to confirm to answer "No".
Maybe we can get Mike to answer on this one.
The new Canon Driver for Lide scanners is now much improved apparently and it is possible to use the Twain feature via Qimage Ultimate to import and save a scan. There are settings in the Advanced mode & preferences to turn off colour adjustment and set 48 bit. However QU is a 16 bit app so I cannot remember what happens there. I have used this method for printer profiles and got reasonable results.
Terry

I can tell by the white punchouts that your bottom crop marks are not in the correct location.  In PP, the bottom corner markers go below the grayscale so that the entire target is cropped: don't use the white marks on the IT8.  This is outlined in the help:

http://www.ddisoftware.com/prism/help/steps.htm#crop

Regards,
Mike

Also, with any Canon LiDE scanner and most scanners built in the last ~5 years, I would recommend using the manufacturer's Twain software instead of VueScan.  While it is possible to get better results with VueScan, that software has gotten so complex in recent years that it is nearly unusable for getting scans with optimal exposure.  There are dozens of parameters that affect exposure and more than a few are linked to one another, making it almost like trying to adjust a carburetor that has 20 mixture screws, all of which are inter-related.  For most end users, it's far easier to get better scans using the manufacturer supplied Twain software than spending hours trying to decipher the hundreds of cryptic settings in VueScan.

That's my .02

Mike

Hi John,
re Mike's comment
I used twain from within Qimage Ultimate and made a scan and got good results. See first screen shot. The gaps at each end of the histogram are ok., I checked with Mike.
The 2 other screen shots are the Canon Scangear interface settings. NB.The color settings are "None". 48/16 bit is set but QU will save as 24bit, that's not a problem.
There's a little bug in QU at present: when the scan has finished, Image Examiner opens but the image is not visible. However, still click File-Save As to save it. Mike will fix this little problem soon.
Hope all that helps and gets you on the way again.
Terry

Thanks Mike & Terry, that's been a big help, although painful having bought VueScan specifically for this purpose.
Another duhhh moment for me with the marking of the corners Mike.
Copying your method Terry (only difference is the page setting is greyed out) I seem to be getting random results re the number of X'd patches.
Only by setting the brightness -20 in the Canon software was I able to get rid of the Warning and change it to Note, but there's no consistency to it.
Would you look at the end of the file name for the brightness setting and the report in the image.
Or at -20 have I hit a sweet spot with the brightness?? Going either way causes the Warning number of Xs.
Is that possibly how it works??
I notice my histogram is very similar to Terrys, but not as deep.
Is the first attachment a file I could carry on in the process with?? Or are there still problems?

Thanks,

John

-15 in brightness showing bottom right of the chart.

At 1am this morning I realized where I was going wrong. I'm going to keep my stupidity to myself.
The Quick Start instructions saved the day for me.
I think the volume of instructions had me rattled.
I went back to VueScan also.
The attached is giving good print & screen match. Exposure is 247.
And the Epson Print Preview is a perfect colour match. Color Munki profile print previews are no where near what ends up on the print. I thought the preview wasn't colour managed.
I've got a good feeling about PP and me getting along very well from now on.


Regards,

John

Hi John,
I thought you were profiling the scanner from previous posts!

I would not rely on the Epson print preview at all, it's definitely not a colour managed view.
The finished print is the real judge.
Anyway, I'm please to hear you are getting used to PP now.
Terry