Vignetting on Microfilm

[Jim Cobban]

In many cases it seems to me that the technicians who were hired to
microfilm public records were poorly trained in the operation of their
equipment in that they chose to use an excessive aperture. The result is
that the microfilm images suffer from vignetting; the outer portions of the
image are darker (in a positive image) than the center. This creates
problems any time such an image is digitized or printed.

My local FHC and the Library and Archives of Canada have installed
digitizing scanners to replace their obsolete (and usually broken down)
analog microfilm printers. However these systems do not seem to have any
features for electronically inserting a digital vignetting filter into the
image scanning, so it remains difficult to obtain a scanned image where the
whole page is usable.

Does anyone have any experience with getting around this in the amateur
market?

[singhals]

Jim Cobban wrote:

In many cases it seems to me that the technicians who were hired to
microfilm public records were poorly trained in the operation of their
equipment in that they chose to use an excessive aperture. The result is
that the microfilm images suffer from vignetting; the outer portions of the
image are darker (in a positive image) than the center. This creates
problems any time such an image is digitized or printed.

My local FHC and the Library and Archives of Canada have installed
digitizing scanners to replace their obsolete (and usually broken down)
analog microfilm printers. However these systems do not seem to have any
features for electronically inserting a digital vignetting filter into the
image scanning, so it remains difficult to obtain a scanned image where the
whole page is usable.

Does anyone have any experience with getting around this in the amateur
market?

I'm told many full-price photoware programs will fix this with one
click. Not having invested in a full-price anything for many years, I
can't speak from experience.

Still, you might want to take it to a DIY photo booth and see what you
can do.

Also -- the FHC's scanner, you can sometimes compensate by choosing the
light carefully -- turn the large knob on the front of the viewer.

Cheryl

[Gjest]

Hello Cheryl , Only Adobe Photoshop can fix it !! , not Paintshop Pro
, or any other thing calling its self a full value graphic program , , I
doubt that adobe can remove it all , I am willing to try if I had a file !!
, PLEASE Do not send one without ASKING First !!!!!!!!! , Phil

----- Original Message -----
From: "singhals" <[email protected]>
To: <[email protected]>
Sent: Sunday, July 17, 2005 9:10 AM
Subject: Re: Vignetting on Microfilm

Jim Cobban wrote:

In many cases it seems to me that the technicians who were hired to
microfilm public records were poorly trained in the operation of their
equipment in that they chose to use an excessive aperture. The result
is
that the microfilm images suffer from vignetting; the outer portions of
the
image are darker (in a positive image) than the center. This creates
problems any time such an image is digitized or printed.

My local FHC and the Library and Archives of Canada have installed
digitizing scanners to replace their obsolete (and usually broken down)
analog microfilm printers. However these systems do not seem to have
any
features for electronically inserting a digital vignetting filter into
the
image scanning, so it remains difficult to obtain a scanned image where
the
whole page is usable.

Does anyone have any experience with getting around this in the amateur
market?




I'm told many full-price photoware programs will fix this with one
click. Not having invested in a full-price anything for many years, I
can't speak from experience.

Still, you might want to take it to a DIY photo booth and see what you
can do.

Also -- the FHC's scanner, you can sometimes compensate by choosing the
light carefully -- turn the large knob on the front of the viewer.

Cheryl



--
No virus found in this incoming message.
Checked by AVG Anti-Virus.
Version: 7.0.323 / Virus Database: 267.8.16/50 - Release Date: 7/15/05

[Robert M. Riches Jr.]

(Top-posting left as-is. Any complaints to the first
top-poster.)

On 2005-07-17, [email protected] <[email protected]> wrote:

Hello Cheryl , Only Adobe Photoshop can fix it !! , not Paintshop Pro
, or any other thing calling its self a full value graphic program , , I
doubt that adobe can remove it all , I am willing to try if I had a file !!
, PLEASE Do not send one without ASKING First !!!!!!!!! , Phil

----- Original Message -----
From: "singhals" <[email protected]
To: <[email protected]
Sent: Sunday, July 17, 2005 9:10 AM
Subject: Re: Vignetting on Microfilm


Jim Cobban wrote:

In many cases it seems to me that the technicians who were hired to
microfilm public records were poorly trained in the operation of their
equipment in that they chose to use an excessive aperture. The result
is
that the microfilm images suffer from vignetting; the outer portions of
the
image are darker (in a positive image) than the center. This creates
problems any time such an image is digitized or printed.

...

Have you tried the Gimp? I'm not a Gimp expert, but I have
seen articles about other things the Gimp can do that would
lead me to believe it ought to be able to do that.

Back in the 1981-1983 time frame, this type of thing was
discussed in a signal processing class. To do this
correction in the frequency domain, just take a 2D FFT of
the image, zero out the low frequency components, and take
the reverse 2D FFT. That would get rid of the variations
that are low in frequency, meaning the ones that cover a
large area. It would leave intact the high frequency
variations, the ones covering small areas of the image. Are
there any commonly available signal processing (Matlab,
S-Plus, R, Octave, etc.) that do 2D FFTs?

In the spatial domain, at each pixel, you'd want to cancel
out the difference between the whole-picture mean value and
the mean value of a blob of image sized around maybe 1% to
10% of the area of the image around that pixel. Given
library routines to read the image from a file into memory
and write the image back out to a file, that should be
fairly easy to program.

Robert Riches
[email protected]
(Yes, that is one of my email addresses.)

[Dennis Lee Bieber]

On Mon, 18 Jul 2005 04:43:12 GMT, "Robert M. Riches Jr."
<[email protected]> declaimed the following in
soc.genealogy.computing:

In the spatial domain, at each pixel, you'd want to cancel
out the difference between the whole-picture mean value and
the mean value of a blob of image sized around maybe 1% to
10% of the area of the image around that pixel. Given
library routines to read the image from a file into memory
and write the image back out to a file, that should be
fairly easy to program.

Assuming a smooth/predictable fade-off (circular, I suspect),

creating a circular gradation "selection" (or alpha channel, or whatever
Photoshop calls is), then applying various gamma, brightness, contrast,
etc. operations to bring up the corners -- and reversing the selection
to bring down the center.

--

==============================================================
[email protected] | Wulfraed Dennis Lee Bieber KD6MOG
[email protected] | Bestiaria Support Staff
==============================================================
Home Page: <http://www.dm.net/~wulfraed/
Overflow Page: <http://wlfraed.home.netcom.com/

[Dave Hinz]

On Mon, 18 Jul 2005 16:54:38 +0000 (UTC), [email protected] <[email protected]> wrote:

What ???? , top post this !!!!!!! ,

Top-posting is what you're doing - screwing up the order of the message
by putting the answer before what it's answering, making it impossible
to know what you're talking about without jumping around in the message.

what is gimp ?

gnu image manipulation program - gimp.org - has many photoshop-like
features, free, and very powerful.

NEVER heard of it ,

google has.

I don`t understand a word of what you wrote , Phil

Could you please put answers after what they're there to answer? It's
hard to know what you're saying also.

[Gjest]

What ???? , top post this !!!!!!! , what is gimp ? NEVER heard of it ,
I don`t understand a word of what you wrote , Phil
----- Original Message -----
From: "Robert M. Riches Jr." <[email protected]>
To: <[email protected]>
Sent: Sunday, July 17, 2005 9:43 PM
Subject: Re: Vignetting on Microfilm

(Top-posting left as-is. Any complaints to the first
top-poster.)

On 2005-07-17, [email protected] <[email protected]> wrote:
Hello Cheryl , Only Adobe Photoshop can fix it !! , not Paintshop
Pro
, or any other thing calling its self a full value graphic program , ,
I
doubt that adobe can remove it all , I am willing to try if I had a file
!!
, PLEASE Do not send one without ASKING First !!!!!!!!! , Phil

----- Original Message -----
From: "singhals" <[email protected]
To: <[email protected]
Sent: Sunday, July 17, 2005 9:10 AM
Subject: Re: Vignetting on Microfilm


Jim Cobban wrote:

In many cases it seems to me that the technicians who were hired to
microfilm public records were poorly trained in the operation of
their
equipment in that they chose to use an excessive aperture. The
result
is
that the microfilm images suffer from vignetting; the outer portions
of
the
image are darker (in a positive image) than the center. This creates
problems any time such an image is digitized or printed.

...

Have you tried the Gimp? I'm not a Gimp expert, but I have
seen articles about other things the Gimp can do that would
lead me to believe it ought to be able to do that.

Back in the 1981-1983 time frame, this type of thing was
discussed in a signal processing class. To do this
correction in the frequency domain, just take a 2D FFT of
the image, zero out the low frequency components, and take
the reverse 2D FFT. That would get rid of the variations
that are low in frequency, meaning the ones that cover a
large area. It would leave intact the high frequency
variations, the ones covering small areas of the image. Are
there any commonly available signal processing (Matlab,
S-Plus, R, Octave, etc.) that do 2D FFTs?

In the spatial domain, at each pixel, you'd want to cancel
out the difference between the whole-picture mean value and
the mean value of a blob of image sized around maybe 1% to
10% of the area of the image around that pixel. Given
library routines to read the image from a file into memory
and write the image back out to a file, that should be
fairly easy to program.

Robert Riches
[email protected]
(Yes, that is one of my email addresses.)



--
No virus found in this incoming message.
Checked by AVG Anti-Virus.
Version: 7.0.323 / Virus Database: 267.8.16/50 - Release Date: 7/15/05

[Robert M. Riches Jr.]

On 2005-07-18, Dave Hinz <[email protected]> wrote:

On Mon, 18 Jul 2005 16:54:38 +0000 (UTC), [email protected] <[email protected]> wrote:
What ???? , top post this !!!!!!! ,

Top-posting is what you're doing - screwing up the order of the message
by putting the answer before what it's answering, making it impossible
to know what you're talking about without jumping around in the message.

what is gimp ?

gnu image manipulation program - gimp.org - has many photoshop-like
features, free, and very powerful.

NEVER heard of it ,

google has.

I don`t understand a word of what you wrote , Phil

Could you please put answers after what they're there to answer? It's
hard to know what you're saying also.

Thanks, Dave.

After reading joe2phil's followup/reply that he sent by
email, I had been wondering how I could answer his questions
without seeming to be undiplomatic. You answered them all
very well.

Robert Riches
[email protected]
(Yes, that is one of my email addresses.)

[Dave Hinz]

On Mon, 18 Jul 2005 18:05:03 GMT, Robert M. Riches Jr. <[email protected]> wrote:

Thanks, Dave.
After reading joe2phil's followup/reply that he sent by
email, I had been wondering how I could answer his questions
without seeming to be undiplomatic. You answered them all
very well.

Well, that was my third edit...