On Mon, Mar 17, 2008 at 1:39 PM, norman wrote:

I have been copying some old colour transparencies using my digital camera and most of the images produced suffer from chromatic aberration somewhere within them. I have tried to find some procedure to remove these blemishes but, so far, have not found anything I can get to work. All suggestions and how to get round the difficulty will be gratefully received.

Can you post a small sample image somewhere?

Chris

On Mon, Mar 17, 2008 at 03:09:47PM -0500, Chris Mohler wrote:

On Mon, Mar 17, 2008 at 1:39 PM, norman wrote:

I have been copying some old colour transparencies using my digital camera and most of the images produced suffer from chromatic aberration somewhere within them. I have tried to find some procedure to remove these blemishes but, so far, have not found anything I can get to work. All suggestions and how to get round the difficulty will be gratefully received.

Not having worked on this problem in particular, I don't really know. My first approximate guess would be to decompose the image into color channels and apply deblurring filters on the seperate channels. You should be able to make reasonable progress with this as as a first order approximation; "all" chromatic aberration is, is the different colors blurring different amounts.

Jeff

Given a sample image I can be more specific than the following:

1 Decompose the image into LAB channels. 2 Despeckle the AB channels (oilify with low exponent is also an option) 3 Recompose

On Tue, Mar 18, 2008 at 6:44 AM, Jeffrey Brent McBeth wrote:

On Mon, Mar 17, 2008 at 03:09:47PM -0500, Chris Mohler wrote: > On Mon, Mar 17, 2008 at 1:39 PM, norman wrote: > > I have been copying some old colour transparencies using my digital > > camera and most of the images produced suffer from chromatic aberration > > somewhere within them. I have tried to find some procedure to remove > > these blemishes but, so far, have not found anything I can get to work. > > All suggestions and how to get round the difficulty will be gratefully > > received.

Not having worked on this problem in particular, I don't really know. My first approximate guess would be to decompose the image into color channels and apply deblurring filters on the seperate channels. You should be able to make reasonable progress with this as as a first order approximation; "all" chromatic aberration is, is the different colors blurring different amounts.

Jeff

--
---------------------------------------------------------------------------- "The man who does not read good books has no advantage over the man who cannot read them."
-- Mark Twain
----------------------------------------------------------------------------

I think there used to be a plugin in the registry for doing this, but I can't remember the name of it. Dunno if it would work in 2.4 or not.

--- norman wrote:

I have been copying some old colour transparencies using my digital camera and most of the images produced suffer from chromatic aberration somewhere within them. I have tried to find some procedure to remove these blemishes but, so far, have not found anything I can get to work. All suggestions and how to get round the difficulty will be gratefully received.

Norman

* norman [03-18-08 09:46]:

Therefore, I believe what I should do is concentrate on photographic techniques with a view to minimising the chromatic aberration as much as possible. Any advice in this direction would be very welcome.

Unfortunately, ca is controlled in lense manufacture and design and has two solutions, software or better glass. Minimizing ca via technique would severly limit your scope, imo.

----- Original Message ----
From: norman
To: gimp-user@lists.XCF.Berkeley.EDU Sent: Tuesday, March 18, 2008 8:52:22 AM Subject: Re: [Gimp-user] chromatic aberration

< snip >

Unfortunately, ca is controlled in lense manufacture and design and has two solutions, software or better glass. Minimizing ca via technique would severly limit your scope, imo.

I would expect most lenses these days to be made such that they do not cause CA. From what I have read, there is another factor to be taken into account, the chip responsible for recording the image. I gathered that small chips are quite prone to CA and the larger the chip the lower the CA and that it virtually disappears in the 1:1 (35 mm) format. The camera I am using is an upper end, point and press so perhaps I need a better camera with a larger chip.

-------------------------------------------------------

CA is indeed a function of the lens quality. You're also right that a smaller sensor makes CA more visible, that's just simple geometry. If the lens produces an abberation of any given size, then if the sensor is half the size, the apparent effect of the abberation is doubled.

Unfortunately, only the best lenses have this effect almost entirely eliminated. You'll find some that are called "Apochromatic" or just "Apo". They tend to be much more expensive than "normal" lenses (typically called "achromatic"). I have a perfectly respectable, but low-end, Nikon zoom lens designed originally for film use that generates what to me is an entirely unacceptable amount of CA at the long end of its zoom on my DX-format D-SLR.

Software can certainly help with this, and "that other product" has this built in. Then again, you can buy a couple of really nice lenses for the price you'll pay for that product ;>

Meanwhile, you're more likely to have trouble because of poor focus, camera shake, and other more mundane issues, than you are from CA in general. I'd say just forget about it, and focus (sorry ;) on your artistic abilities. Let's face it, the lenses that most of the "greats" used were total junk compared to the most basic point and shoot now. See Ken Rockwell's comments on "it's not the camera" at http://www.kenrockwell.com/tech/notcamera.htm

Cheers, Simon

"You can tell whether a man is clever by his answers. You can tell whether a man is wise by his questions." — Naguib Mahfouz

_____________________________________

On Tue 18-Mar-2008 at 08:13 -0700, Simon Roberts wrote:

Software can certainly help with this, and "that other product" has this built in. Then again, you can buy a couple of really nice lenses for the price you'll pay for that product ;>

Not an immediate solution, but 'over at the hugin project' we have a demonstrated technique for automatic correction of transverse chromatic aberration. Sponsorship for turning this into an everyday tool is available under the Google Summer of Code:

http://wiki.panotools.org/SoC_2008_ideas#tCA_Correction

----- Original Message ----
From: norman
...

CA is indeed a function of the lens quality. You're also right that a smaller sensor makes CA more visible, that's just simple geometry. If the lens produces an abberation of any given size, then if the sensor is half the size, the apparent effect of the abberation is doubled.

Unfortunately, only the best lenses have this effect almost entirely eliminated. You'll find some that are called "Apochromatic" or just "Apo". They tend to be much more expensive than "normal" lenses (typically called "achromatic"). I have a perfectly respectable, but low-end, Nikon zoom lens designed originally for film use that generates what to me is an entirely unacceptable amount of CA at the long end of its zoom on my DX-format D-SLR.

I feel sure that you must be correct. I have never seen any noticeable fringing or CA effects with my ordinary photography it is only with this project I set myself of copying a lot of old colour transparencies. In the old days I used to often feel frustrated at not being able to do a great deal with colour slides such as I did in my darkroom with black and white film. Thus, I saw this as chance to catch up on history and at the same time, maybe, produce some interesting images digitally. It now looks as though I shall be frustrated yet again.

Norman

-----------------------------------------------------------------------------------------------------------

Are you "copying" the slides, or are you scanning them?

If you're using a slide copying attachment and effectively rephotographing them onto your digital camera, then the CA of the copying equipment will be a factor. On the other hand, if you're scanning them, then CA isn't usually an issue, because scanning is a different mechanism entirely. If you're scanning and then seeing CA, I believe the CA must be in the original slide. Perhaps you just didn't notice before? We do tend to view digital images at much higher magnifications than we used to view silver halide images. (To be fair, that might not be true of slides though!)

Cheers, Simon

_____________________________________

?On Tue, 2008-03-18 at 16:23 +0000, norman wrote:

On Tue, 2008-03-18 at 15:46 +0000, Bruno Postle wrote:

On Tue 18-Mar-2008 at 08:13 -0700, Simon Roberts wrote:

Software can certainly help with this, and "that other product" has this built in. Then again, you can buy a couple of really nice lenses for the price you'll pay for that product ;>

Not an immediate solution, but 'over at the hugin project' we have a demonstrated technique for automatic correction of transverse chromatic aberration. Sponsorship for turning this into an everyday tool is available under the Google Summer of Code:

http://wiki.panotools.org/SoC_2008_ideas#tCA_Correction

This seems to assume that the optics are the cause of the CA whereas I understand that CA is also caused by the chip in e digital camera. Will this process take care of that?

The chip doesn't cause CA, it's only the lens. A smaller sensor is more sensitive for CA, because the pixels are smaller and enlarge every lens problem. A CA that stays on one pixel in a Nikon D3 (and is invisible) would cover a lot of pixels on a 1/1.8" chip and would become visible.

Rolf ?
http://meetthegimp.org - weekly videopodcast about GIMP and digital photography

?On Tue, 2008-03-18 at 16:23 +0000, norman wrote:

On Tue, 2008-03-18 at 15:46 +0000, Bruno Postle wrote:

On Tue 18-Mar-2008 at 08:13 -0700, Simon Roberts wrote:

Software can certainly help with this, and "that other product" has this built in. Then again, you can buy a couple of really nice lenses for the price you'll pay for that product ;>

Not an immediate solution, but 'over at the hugin project' we have a demonstrated technique for automatic correction of transverse chromatic aberration. Sponsorship for turning this into an everyday tool is available under the Google Summer of Code:

http://wiki.panotools.org/SoC_2008_ideas#tCA_Correction

This seems to assume that the optics are the cause of the CA whereas I understand that CA is also caused by the chip in e digital camera. Will this process take care of that?

The chip doesn't cause CA, it's only the lens. A smaller sensor is more sensitive for CA, because the pixels are smaller and enlarge every lens problem. A CA that stays on one pixel in a Nikon D3 (and is invisible) would cover a lot of pixels on a 1/1.8" chip and would become visible.

Rolf ?
http://meetthegimp.org - weekly videopodcast about GIMP and digital photography

?On Tue, 2008-03-18 at 15:33 +0000, norman wrote:

I feel sure that you must be correct. I have never seen any noticeable fringing or CA effects with my ordinary photography it is only with

this

project I set myself of copying a lot of old colour transparencies. In the old days I used to often feel frustrated at not being able to do a great deal with colour slides such as I did in my darkroom with black and white film. Thus, I saw this as chance to catch up on history and

at

the same time, maybe, produce some interesting images digitally. It

now

looks as though I shall be frustrated yet again.

Norman

I think your camera is OK, the problem is your "close up filter" that you use for enlarging the slides. These things are prone to CA. You can get away with a lens made out of two glasses (called achromatic and being expensive), but most of these things are made out of a single glass and mess up the colours. I never have heard about an apochromatic lens to put in front of a point and shoot.

You can try to use your camera without this lens in macro mode and crop the image later in GIMP. With luck there is enough resolution left.

Rolf

http://meetthegimp.org - weekly videopodcast about GIMP and digital photography

On Tue 18-Mar-2008 at 16:23 +0000, norman wrote:

http://wiki.panotools.org/SoC_2008_ideas#tCA_Correction

This seems to assume that the optics are the cause of the CA whereas I understand that CA is also caused by the chip in e digital camera. Will this process take care of that?

Nope.

* norman [03-18-08 10:53]:

< snip >

Unfortunately, ca is controlled in lense manufacture and design and has two solutions, software or better glass. Minimizing ca via technique would severly limit your scope, imo.

I would expect most lenses these days to be made such that they do not cause CA. From what I have read, there is another factor to be taken into account, the chip responsible for recording the image. I gathered that small chips are quite prone to CA and the larger the chip the lower the CA and that it virtually disappears in the 1:1 (35 mm) format. The camera I am using is an upper end, point and press so perhaps I need a better camera with a larger chip.

While this may be correct to some extent, *glass* is a *major* factor. I can show you ca from my Nikon D3 (a full frame sensor) with a 70-300G Nikon lense, a *cheap* lense.

Simon Roberts wrote:

See Ken Rockwell's comments on "it's not the camera" at http://www.kenrockwell.com/tech/notcamera.htm

An excellent discussion, although Rockwell fails to mention Ansel Adams's darkroom artistry (see wikipedia on Ansel Adams and the associated reference 18).

I have found that a small low cost program called Ptlens does a very fair job of reducing CA, and various other lens distortions.. It only runs on Win2K, XP or Vista or Unix with a windows simulator, either free standing or as a PS plugin. Profiles are available for many cameras. Details available on:
http://epaperpress.com/ptlens/

It costs $15 for a full license

Peter Taylor

norman wrote:

I have been copying some old colour transparencies using my digital camera and most of the images produced suffer from chromatic aberration somewhere within them. I have tried to find some procedure to remove these blemishes but, so far, have not found anything I can get to work. All suggestions and how to get round the difficulty will be gratefully received.

Should be easy(-ish) :

Split image into red, green, blue channels. Apply lens correction to red and blue channels to align with green. Recombine channels.

Hi,

On Wed, 2008-03-19 at 14:16 +1100, David Hodson wrote:

Should be easy(-ish) :

Split image into red, green, blue channels. Apply lens correction to red and blue channels to align with green. Recombine channels.

That would work well if the light was combined out of exactly three well-defined frequencies, one red, one green, one blue. But that is usually not the case. It could still be an improvement though...

Sven

norman wrote:

On Wed, 2008-03-19 at 08:32 +0000, norman wrote:

I have found that a small low cost program called Ptlens does a very fair job of reducing CA, and various other lens distortions.. It only runs on Win2K, XP or Vista or Unix with a windows simulator, either free standing or as a PS plugin. Profiles are available for many cameras. Details available on:
http://epaperpress.com/ptlens/

It costs $15 for a full license

Sounds interesting, would you know if it will run using Wine?

Ill answer this myself - It appears to work in Wine on Ubuntu 7.10. Ten tries are allowed free of charge so I will try it on a few images and report back.

Norman

Seems to work on Slack 12, but only seems to open jpegs.

Andrew