VueScan raw file is not a true raw file!

[Ed Hamrick]

The process of building the IR channel is not as obvious as it seems.
In fact, IBM has patented this process.

Yes, a guy named Al Eggar did this patent when he worked for IBM.
He works for ASF (and now Kodak).

So what Ed has done is skirt the patent

Yes, the patent doesn't actually work nearly as well as the
algorithm I use (which is completely, totally different from
what's described in the patent).

So: what the heck is the "defect" channel?

It's just marketing-speak for the infrared channel. It's as
simple as that.

Presumably not to Vuescan's 64-bit RAW file, since it's
already done some processing on the data

No, the 64-bit RAW file is just the raw data as it comes
over the SCSI/USB/Firewire cable.

Regards,
Ed Hamrick

 

[Henk de Jong]

Hi Ed,

This thread I started is wandering a different direction then my original
posting... ;-)

My original question was a kind of "how about applying a colour profile
before saving the raw file?" which looks like in contradiction with the
information in the help-files.

If I am right, are you going to do something about this?
I think it is very important that we are sure that the raw files we save for
archiving purposes are with no processing at all, because VueScan is still
improving over time...
The option "Raw output with: Save" is not very clear about any
pre-processing like grain reduction and colour profiles...

With kind regards,
Henk de Jong

 

[Wilfred van der Vegte]

(...)

Yes, I've tried this technique. It works, although it's not as good as
ICE or Ed Hamrick's software.

As I said, there are more sophisticated ways of processing the
selection. I expect that, for instance, performing the blurring in
another layer with the blending options set to lighten (for slides) or
darken (for negs) would give better results. And adding some noise may
also help. Anyway, Ed's implementation still seems to have a bug (?)
related to dealing with the edges of the frame - which I hope he is
working on now.

 

[Ronald Bruck]

Yes, a guy named Al Eggar did this patent when he worked for IBM.
He works for ASF (and now Kodak).


Yes, the patent doesn't actually work nearly as well as the
algorithm I use (which is completely, totally different from
what's described in the patent).


It's just marketing-speak for the infrared channel. It's as
simple as that.


No, the 64-bit RAW file is just the raw data as it comes
over the SCSI/USB/Firewire cable.

Ed, great! Straight from the horse's mouth, so to speak

One last query: since the Minolta Pro is, IIRC, a 4 * 14-bit scanner,
that means the data has to be scaled to maintain dynamic range.
(Complicated by multiple passes.) Is that scaling done before writing
the RAW file, or in the postprocessing?

(I think it must be done in the postprocessing. That would explain why
RAW files look so dark when viewed in graphics apps?)

--Ron Bruck

 

[Bart van der Wolf]

SNIP

One last query: since the Minolta Pro is, IIRC, a 4 * 14-bit scanner,
that means the data has to be scaled to maintain dynamic range.
(Complicated by multiple passes.) Is that scaling done before writing
the RAW file, or in the postprocessing?

(I think it must be done in the postprocessing. That would explain why
RAW files look so dark when viewed in graphics apps?)

Several things need to be done to make a Raw file display well. The dark
appearance is mainly due to a linear datafile gamma and your display
subsequently applying another gamma conversion. The Raw data needs to
receive an inverse gamma correction which will be neutralized by the display
gamma.
Scaling mainly changes contrast, but colorbalancing is needed as well.

Bart

 

[Ed Hamrick]

My original question was a kind of "how about applying a colour profile
before saving the raw file?" which looks like in contradiction with the
information in the help-files.

The raw data is raw - there's no color profile applied.

I think it is very important that we are sure that the raw files we save for
archiving purposes are with no processing at all, because VueScan is still
improving over time...

Yes, that's how VueScan works.

The option "Raw output with: Save" is not very clear about any
pre-processing like grain reduction and colour profiles...

If you use this option, the grain reduction and infrared cleaning
is applied. Color profiles aren't applied.

Regards,
Ed Hamrick

 

[Ed Hamrick]

One last query: since the Minolta Pro is, IIRC, a 4 * 14-bit scanner,
that means the data has to be scaled to maintain dynamic range.
(Complicated by multiple passes.) Is that scaling done before writing
the RAW file, or in the postprocessing?

The 14-bit data is padded to 16 bits with two zero bits.

When doing multiple passes, the average pixel value is written.

(I think it must be done in the postprocessing. That would explain why
RAW files look so dark when viewed in graphics apps?)

No, raw files look dark because they have gamma 1.0.

Regards,
Ed Hamrick

 

[Henk de Jong]

Ed and Erik,

The raw data is raw - there's no color profile applied.

Okay, I believe you, just forget everything I wrote about colour profiles in
combination with raw files ;-)

Tonight I did a test with my ScanWit 2720S (the one without infra red) and a
slide.
I had two Raw files created: one with the option "Raw output with: Scan" and
one with the option "Raw output with: Save" All other settings in VueScan
are the same. Both scans are made with "Scanner color space" set to "ICC
Profile" which I have made from an IT8 target slide from Wolf Faust.
I opened the two resulting Raw files in Paint Shop Pro, applied a gamma
correction of 2.2, resized them much smaller, saved the files as JPEG and
uploaded them to my homepage.

The results can be seen at:
http://www.hsdejong.nl/vuescan/test.htm
It is clear that the first two images have a very different colour.
The one "Raw output with: Scan" has an acceptable colour. The other one is
much too green.

When I use the "Built-in Scanner color space" option the results, either
with "Raw output with: Scan" or "Raw output with: Save" are identical and
with acceptable colour.
The results of these raw files are the two bottom images on the before
mentioned web page. Both images have the same colour as the "Raw output
with: Scan" image and "Scanner color space" set to "ICC Profile".

My conclusion is:
1)
With ICC Built-in profiles the raw files with "Raw output with: Scan" and
"Raw output with: Save" are identical and with good colour.
2)
With a self made ICC profile (from a Wolf Faust slide) the "Raw output with:
Save" has a wrong colour and the one with the "Raw output with: Scan" is
okay.

Something is wrong, but what?

With kind regards after much testing,
Henk de Jong

 

[Henk de Jong]

Hi Ed,

What do you think about the conclusions regarding my experiments with "Raw
output with:"?


With kind regards,

Henk de Jong

 

[gour]

Hello Henk!

Excuse me for jumping in after 5 yrs

but I'm new Epson V700 user preparing to do batch scanning of ~2500 slides.

My conclusion is:
1)
With ICC Built-in profiles the raw files with "Raw output with: Scan" and
"Raw output with: Save" are identical and with good colour.
2)
With a self made ICC profile (from a Wolf Faust slide) the "Raw output with:
Save" has a wrong colour and the one with the "Raw output with: Scan" is
okay.

Have you got any reply on the above?

I plan to scan 35mm slides in 4800 dpi and archive them on HD/tapes for usage in video 'slide-show' projects later.

In several places on the Internet I read about the possible problem when using 'Raw output with: Save' and that's why plan to use 'Raw output with: Scan'.

However, I've W. Faust's targets, but would like to skip calibrating-monitor and profiling-scanner steps and just do 'raw batch scan' and then re-scan from files in Vuescan (after I'd calibrate my monitor and profile the scanner) to do post-processing phase and applying scans to ICC profiles.

Few questions:

• is the above workflow correct?

• is it confirmed that using built-in ICC profile provides that which I need for archiving of my slides?

• is it better to not crop in this phase when doing archiving-scans?

Sincerely,
Gour