Chromatic Aberration

Below are two clips (300x500 from 1200x1600 files) from succesive exposures. I know they are very poor pictures:

http://albums.photopoint.com/j/View?u=10256&a=12913443&p=48413605
http://albums.photopoint.com/j/View?u=10256&a=12913443&p=48413606

The first shows a severe case of what Phil Askey has called chromatic aberration. I have been uncomfortable with this name for the phenomenon, because I associate the term "chromatic aberration" with a common linear phenomenon arising from the variation of refractive index of a transparent medium as a function of wavelength.

It seems to me that these two images show a non-linear process, in part associated with the difference in brightness of adjacent areas.

I am unable at the moment to think of experiments to explain the phenomenon further. I am strongly suspicious that it is not primarily caused by shortcomings of the lens. --Gerry

http://home.columbus.rr.com/gfoley/

http://www.dpreview.com/learn/Glossary/Optical/Chromatic_Aberrations_01.htm

Gerry said:

Below are two clips (300x500 from 1200x1600 files) from succesive
exposures. I know they are very poor pictures:

http://albums.photopoint.com/j/View?u=10256&a=12913443&p=48413605
http://albums.photopoint.com/j/View?u=10256&a=12913443&p=48413606

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name for
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

It seems to me that these two images show a non-linear process, in
part associated with the difference in brightness of adjacent areas.

I am unable at the moment to think of experiments to explain the
phenomenon further. I am strongly suspicious that it is not
primarily caused by shortcomings of the lens.
--
Gerry

http://home.columbus.rr.com/gfoley/

Click to expand...

Hi Gerry,

the differences between the two exposures are simple to explain: the more a picture is out of focus, the more cromatic aberrations (if allready there by the lens) are visible. I noticed that many times, and try to wait till the camera has focused completely (difficult/not possible with moving objects).
tc

Gerry said:

Below are two clips (300x500 from 1200x1600 files) from succesive> exposures. I know they are very poor pictures:> > http://albums.photopoint.com/j/View?u=10256&a=12913443&p=48413605 > http://albums.photopoint.com/j/View?u=10256&a=12913443&p=48413606 > > The first shows a severe case of what Phil Askey has called> chromatic aberration. I have been uncomfortable with this name for> the phenomenon, because I associate the term "chromatic aberration"> with a common linear phenomenon arising from the variation of> refractive index of a transparent medium as a function of> wavelength.> > It seems to me that these two images show a non-linear process, in> part associated with the difference in brightness of adjacent areas.> > I am unable at the moment to think of experiments to explain the> phenomenon further. I am strongly suspicious that it is not> primarily caused by shortcomings of the lens.> --> Gerry> > http://home.columbus.rr.com/gfoley/

Click to expand...

crowning said:

Hi Gerry,
the differences between the two exposures are simple to explain:
the more a picture is out of focus, the more cromatic aberrations
(if allready there by the lens) are visible. I noticed that many
times, and try to wait till the camera has focused completely
(difficult/not possible with moving objects).

Click to expand...

I think a major difference between the two frames is in the proportion of the aircraft fuselage which is reflecting the sun directly into the camera. I cannot see any significant difference in sharpness which would be attributabke to autofocusing. To my eye, the images are equally fuzzy. -))

Gerry,

To my eye as well, the two pictures are equally out of focus. I believe the difference in Chromatic Aberrration that you see is due to the difference (however slight it may seem) in angle. Remember, the CCD in a digital camera is significantly smaller than a 35mm Film frame. This causes even very minor differences in angle to result in major differences in refraction off the glass of the lens. I believe that the rush to market cameras with 3MP CCD's has left the quality of the lenses mounted on the cameras a bit behind. While I can and do experience CA on my 2MP C2100, It is infrequent, and much more subtle in degree than some I have seen on images posted using 3MP cameras. This is a subjective rather than scientific conclusion on my part. I have read articles that point in this direction. I would love to be able to build a lab in the basement to to test my theories, however, the care and feeding of family is currently of higher priority.

Mike D.

Gerry said:

I think a major difference between the two frames is in the
proportion of the aircraft fuselage which is reflecting the sun
directly into the camera. I cannot see any significant difference
in sharpness which would be attributabke to autofocusing. To my
eye, the images are equally fuzzy. -))

Click to expand...

Gerry -

HOORAY FOR YOU, for finally raising this issue!

I have been thinking of doing the same myself,
for some time now--but I haven't taken the time.

Based on long photographic experience (together
with some training in physics), I believe the following:

The digicam image defects (rather carelessly) described as
"chromatic aberrations" in many posts and articles are,
in most cases, a MIXTURE of optical and imager defects.
In these "mixtures," true chromatic aberration is often
a minor contributor--and in many cases, not visible at all.

I believe the major contributors to the "CA" mix are
as follows (approximately in order of importance):

1. Imager "overflow" effects - These are referred to as
"blooming" by Phil. This problem is apparent in many
images, and is NOT CA, or in any way a lens defect.

2. Flare - Problems are often seen with light "leaking"
from light to dark areas in high-contrast subjects.
"Blooming" is probably the predominant cause of this
in digicams. However, the lens can also cause this;
and when it does, "flare" is usually responsible. Good
lens coatings and simple lens designs are the main
defenses on this one.

3. Chromatic aberration - This can and does occur--
but not nearly to the extent that one might think,
when every instance of blooming and/or flare is
described as CA. A key thing to look for here is COLOR.
True CA typically separates the light into MORE THAN
ONE color. Visible separation of the reds and greens
is very typical. Think of the output of a prism, which
can be thought of as INTENDED to produce CA.
Note also that true CA does NOT depend on high
contrast. If the colors are being separated by CA,
this will occur in low-contrast subjects as well.

I submit that the "real deal" (admittedly oversimplified)
goes something like this:

When you see a digicam image that shows one-color,
contrast-dependent "leakage" from light to dark areas,
you are very likely seeing a mixture of blooming and flare,
NOT true chromatic aberration. This description applies
to MOST of the so-called CA images that we see.

When you see a multi-color, prism-like effect--particularly
in low-contrast images, you are almost certainly seeing
true chromatic aberration. Images like this are much
rarer than blooming images. However, I could point out
some such images, on-line.

There is one current model digicam that has such obvious
and bad (true) CA, that it could be thought of as the CA
poster kid. (But I don't want to name it in this post, 'cause
I don't want a nice, technical discussion of optical effects
to be diluted by flames from fans of that camera.)

Gerry said:

Below are two clips (300x500 from 1200x1600 files) from succesive
exposures. I know they are very poor pictures:

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name fore
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

It seems to me that these two images show a non-linear process, in
part associated with the difference in brightness of adjacent areas.

I am unable at the moment to think of experiments to explain the
phenomenon further. I am strongly suspicious that it is not
primarily caused by shortcomings of the lens.
--
Gerry

http://home.columbus.rr.com/gfoley/

Click to expand...

Hello Guys,

Old Ed said:

Gerry -

HOORAY FOR YOU, for finally raising this issue!

Click to expand...

To add another theory to the mix, there was a thread in the Canon forum and therein it was argued that CA as generally defined was actually the result of IR and it was suggested that using a flash would alter the balance of lighting and reduce or eliminate the effect. I guess outdoors you could add on a hot mirror filter.

CA does seem to occur more frequently on cloudy days when the color temperture actually increases sightly, but I'm not totally convinced that this is the only source of the problem.

The only thing I am sure of is that the camera companies will NOT offer an explanation as they seem quite unable to deal with the effect. This has to be in part because of the lack of competition amongst the Japanese companies. And as long as we keep buying they have no reason to innovate.

ms

CA has got to have something to do with exposure levels. My Casio QV3000 is not bad for CA problems, BUT if I get a shot that comes out overexposed slightly then these are the ones that are much more likely to show CA up.

I've found a good quick fix for the usual tree branch CA problem in PSP7.

Select the general branch area having the problem, this is not too critical using the Freehand tool. Then go to Colors> Adjust> Hue/Saturation/Brightness and choose Magentas and turn the Saturation down to -100. Cured.

Old Ed said:

HOORAY FOR YOU, for finally raising this issue!

I have been thinking of doing the same myself,
for some time now--but I haven't taken the time.

Based on long photographic experience (together
with some training in physics), I believe the following:

The digicam image defects (rather carelessly) described as
"chromatic aberrations" in many posts and articles are,
in most cases, a MIXTURE of optical and imager defects.
In these "mixtures," true chromatic aberration is often
a minor contributor--and in many cases, not visible at all.

I believe the major contributors to the "CA" mix are
as follows (approximately in order of importance):

1. Imager "overflow" effects - These are referred to as
"blooming" by Phil. This problem is apparent in many
images, and is NOT CA, or in any way a lens defect.

2. Flare - Problems are often seen with light "leaking"
from light to dark areas in high-contrast subjects.
"Blooming" is probably the predominant cause of this
in digicams. However, the lens can also cause this;
and when it does, "flare" is usually responsible. Good
lens coatings and simple lens designs are the main
defenses on this one.

3. Chromatic aberration - This can and does occur--
but not nearly to the extent that one might think,
when every instance of blooming and/or flare is
described as CA. A key thing to look for here is COLOR.
True CA typically separates the light into MORE THAN
ONE color. Visible separation of the reds and greens
is very typical. Think of the output of a prism, which
can be thought of as INTENDED to produce CA.
Note also that true CA does NOT depend on high
contrast. If the colors are being separated by CA,
this will occur in low-contrast subjects as well.

I submit that the "real deal" (admittedly oversimplified)
goes something like this:

When you see a digicam image that shows one-color,
contrast-dependent "leakage" from light to dark areas,
you are very likely seeing a mixture of blooming and flare,
NOT true chromatic aberration. This description applies
to MOST of the so-called CA images that we see.

When you see a multi-color, prism-like effect--particularly
in low-contrast images, you are almost certainly seeing
true chromatic aberration. Images like this are much
rarer than blooming images. However, I could point out
some such images, on-line.

There is one current model digicam that has such obvious
and bad (true) CA, that it could be thought of as the CA
poster kid. (But I don't want to name it in this post, 'cause
I don't want a nice, technical discussion of optical effects
to be diluted by flames from fans of that camera.)

Gerry said:

Below are two clips (300x500 from 1200x1600 files) from succesive
exposures. I know they are very poor pictures:

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name fore
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

It seems to me that these two images show a non-linear process, in
part associated with the difference in brightness of adjacent areas.

I am unable at the moment to think of experiments to explain the
phenomenon further. I am strongly suspicious that it is not
primarily caused by shortcomings of the lens.
--
Gerry

http://home.columbus.rr.com/gfoley/

Click to expand...

Click to expand...

I agree with Old Ed. I have seen real CA in some shots and, as Ed says, it typically involves a blue fringe on one side of an dark object and a red fringe on the other. I would further add that there is possible amplification of the effects mentioned by Ed by the very agressive color interpolation, sharpening, and other processing in many cameras. Leon

Old Ed said:

HOORAY FOR YOU, for finally raising this issue!

I have been thinking of doing the same myself,
for some time now--but I haven't taken the time.

Based on long photographic experience (together
with some training in physics), I believe the following:

The digicam image defects (rather carelessly) described as
"chromatic aberrations" in many posts and articles are,
in most cases, a MIXTURE of optical and imager defects.
In these "mixtures," true chromatic aberration is often
a minor contributor--and in many cases, not visible at all.

I believe the major contributors to the "CA" mix are
as follows (approximately in order of importance):

1. Imager "overflow" effects - These are referred to as
"blooming" by Phil. This problem is apparent in many
images, and is NOT CA, or in any way a lens defect.

2. Flare - Problems are often seen with light "leaking"
from light to dark areas in high-contrast subjects.
"Blooming" is probably the predominant cause of this
in digicams. However, the lens can also cause this;
and when it does, "flare" is usually responsible. Good
lens coatings and simple lens designs are the main
defenses on this one.

3. Chromatic aberration - This can and does occur--
but not nearly to the extent that one might think,
when every instance of blooming and/or flare is
described as CA. A key thing to look for here is COLOR.
True CA typically separates the light into MORE THAN
ONE color. Visible separation of the reds and greens
is very typical. Think of the output of a prism, which
can be thought of as INTENDED to produce CA.
Note also that true CA does NOT depend on high
contrast. If the colors are being separated by CA,
this will occur in low-contrast subjects as well.

I submit that the "real deal" (admittedly oversimplified)
goes something like this:

When you see a digicam image that shows one-color,
contrast-dependent "leakage" from light to dark areas,
you are very likely seeing a mixture of blooming and flare,
NOT true chromatic aberration. This description applies
to MOST of the so-called CA images that we see.

When you see a multi-color, prism-like effect--particularly
in low-contrast images, you are almost certainly seeing
true chromatic aberration. Images like this are much
rarer than blooming images. However, I could point out
some such images, on-line.

There is one current model digicam that has such obvious
and bad (true) CA, that it could be thought of as the CA
poster kid. (But I don't want to name it in this post, 'cause
I don't want a nice, technical discussion of optical effects
to be diluted by flames from fans of that camera.)

Gerry said:

Below are two clips (300x500 from 1200x1600 files) from succesive
exposures. I know they are very poor pictures:

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name fore
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

It seems to me that these two images show a non-linear process, in
part associated with the difference in brightness of adjacent areas.

I am unable at the moment to think of experiments to explain the
phenomenon further. I am strongly suspicious that it is not
primarily caused by shortcomings of the lens.
--
Gerry

http://home.columbus.rr.com/gfoley/

Click to expand...

Click to expand...

Steve -

Thanks for your inputs! A few follow-ups below...

SteveB said:

CA has got to have something to do with exposure levels. My Casio
QV3000 is not bad for CA problems, BUT if I get a shot that comes
out overexposed slightly then these are the ones that are much more
likely to show CA up.

Click to expand...

Please note that TRUE CA has absolutely nothing to do with
exposure levels; but blooming and flare effects DO! So some
of us crotchety old purists would be much happier if you'd
avoid using the CA misnomer in these situations.

Member said:

I've found a good quick fix for the usual tree branch CA problem in
PSP7.
Select the general branch area having the problem, this is not too
critical using the Freehand tool. Then go to
Colors> Adjust> Hue/Saturation/Brightness and choose Magentas and
turn the Saturation down to -100. Cured.

Click to expand...

Thank you for this tip. I just bought (but haven't yet installed)
PSP7; so I'm on the lookout for useful PSP techniques.

The phrase 'chromatic aberration' although not completely accurate is widely accepted in the digital photography community and within the r&d departments of manufacturers (it certainly wasn't 'coined' by me).

In fact it's caused by the direct interaction between the main lens system and the microlenses which sit over every pixel of the CCD. It's to do with very bright rays of light refracting off the edges of these microlenses.

Gerry said:

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name for
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

Click to expand...

The top photo, which I assume was cropped from an area near the edge of the frame, definitely exhibits both chromatic aberration and blooming. The bottom photo, clipped from an area near the center, shows no chromatic aberration. Ths photo may also show some flare as well.

One manifestation of chromatic aberration and much less of the other two. The reduction in flare I attribute to the less intense reflection from the body of the plane. The lack of chromatic aberration I assume is because this picture is cropped from an area near the center of the frame.

In today's digicams, chromatic aberration most often manifests itself as slichtly different image sizes for different colors of light. This results in color fringes around objects near the edges of the frame, with one color bordering the edge of the object away from the center of the frame and its complement bordering the edge toward the center. In most digicam pictures, these fringes are usually yellow or yellow-green and its bluish complement. These fringes don't show at all in the middle of the frame but get progressively worse as the edges are approached. In the top picture, the tail of the plane shows these fringes.

Blooming is an entirely different effect, caused by excess charge spilling into neighboring pixels. Blooming, in contrast (no real pun intended) to chromatic aberration, can appear anywhere in the frame and always occurs around bright objects like the specular highlight on the body of the plane shown here. Blooming can result in swelling of the highlights, hence its name, and often color fringing. Unlike the color fringing from chromatic aberration, the fringe may be asymmetrical in shape and is usually of one color all the way around. In this case, as in many cases, magenta or purple.

Lens flare is another effect appearing in photos caused by strong specular highlights. Lens flare may appear as a general bright halo around the object or even as a number of bright circles extending through the center of the frame and even of the side of the frame opposite from the highlight causing the flare. The cause of this is internal reflections in the lens and is not confined to digicams. Zoom lenses are particularly prone to this problem because of the large number of glass surfaces in these lenses.

A fourth effect is halation. Like lens flare, this looks like a white halo around the offending highlight. Unlike blooming, these halos don't show any color fringes, except in cases where both halation and blooming are present. Like blooming, halation's origin is in the CCD (or film). Unlike blooming, halation is an optical effect in which light spreads from the offending highlight by internal reflections within the CCD.

Old Ed said:

HOORAY FOR YOU, for finally raising this issue!

I have been thinking of doing the same myself,
for some time now--but I haven't taken the time.

Based on long photographic experience (together
with some training in physics), I believe the following:

The digicam image defects (rather carelessly) described as
"chromatic aberrations" in many posts and articles are,
in most cases, a MIXTURE of optical and imager defects.
In these "mixtures," true chromatic aberration is often
a minor contributor--and in many cases, not visible at all.

I believe the major contributors to the "CA" mix are
as follows (approximately in order of importance):

1. Imager "overflow" effects - These are referred to as
"blooming" by Phil. This problem is apparent in many
images, and is NOT CA, or in any way a lens defect.

2. Flare - Problems are often seen with light "leaking"
from light to dark areas in high-contrast subjects.
"Blooming" is probably the predominant cause of this
in digicams. However, the lens can also cause this;
and when it does, "flare" is usually responsible. Good
lens coatings and simple lens designs are the main
defenses on this one.

3. Chromatic aberration - This can and does occur--
but not nearly to the extent that one might think,
when every instance of blooming and/or flare is
described as CA. A key thing to look for here is COLOR.
True CA typically separates the light into MORE THAN
ONE color. Visible separation of the reds and greens
is very typical. Think of the output of a prism, which
can be thought of as INTENDED to produce CA.
Note also that true CA does NOT depend on high
contrast. If the colors are being separated by CA,
this will occur in low-contrast subjects as well.

I submit that the "real deal" (admittedly oversimplified)
goes something like this:

When you see a digicam image that shows one-color,
contrast-dependent "leakage" from light to dark areas,
you are very likely seeing a mixture of blooming and flare,
NOT true chromatic aberration. This description applies
to MOST of the so-called CA images that we see.

When you see a multi-color, prism-like effect--particularly
in low-contrast images, you are almost certainly seeing
true chromatic aberration. Images like this are much
rarer than blooming images. However, I could point out
some such images, on-line.

There is one current model digicam that has such obvious
and bad (true) CA, that it could be thought of as the CA
poster kid. (But I don't want to name it in this post, 'cause
I don't want a nice, technical discussion of optical effects
to be diluted by flames from fans of that camera.)

Gerry said:

Below are two clips (300x500 from 1200x1600 files) from succesive
exposures. I know they are very poor pictures:

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name fore
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

It seems to me that these two images show a non-linear process, in
part associated with the difference in brightness of adjacent areas.

I am unable at the moment to think of experiments to explain the
phenomenon further. I am strongly suspicious that it is not
primarily caused by shortcomings of the lens.
--
Gerry

http://home.columbus.rr.com/gfoley/

Click to expand...

Click to expand...

Phil Askey said:

The phrase 'chromatic aberration' although not completely accurate
is widely accepted in the digital photography community and within
the r&d departments of manufacturers (it certainly wasn't 'coined'
by me).

In fact it's caused by the direct interaction between the main lens
system and the microlenses which sit over every pixel of the CCD.
It's to do with very bright rays of light refracting off the edges
of these microlenses.

Click to expand...

In defense of Gerry, your glossary entry on CA goes into detail on the classical lens CA, and only mentions blooming as a compounding factor, with no mention of the refracting (reflecting?) off the edges of microlenses. I think Gerry is referring to the fact that in many of the forums the topic comes up periodically, and is generally treated as a lens CA issue, as if that were the complete story.

It seems to me that Old Ed's explanation is getting closer to the real issue, although the I suspect the blooming issue needs more fleshing out.

It doesn't strike me as exactly correct that blooming is just reflection/refraction off the microlenses (if that was what your were saying), because then one would expect the effect to be dependent solely on differences in intensity, and independent of the overall exposure. On the other hand, it appears that the effect is position dependent: the purple fringe on the left side of an image is on the right side of the high contrast boundary, and on the right side it's on the left (and top=bottom, bottom=top). Thus, migration of photoelectons from full wells (blooming?) does not fully account for the effect either.

Like Gerry, old ed, and the others, I'd be interested in a good mechanistic explanation of the effect.

Member said:

Gerry said:

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name for
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

Click to expand...

Click to expand...

I have followed this thread with interest, and while I don't have the level of knowledge here displayed, I can recall enough of my physics to follow it.

Regarding the following, (edited)

...Unlike the color fringing from chromatic aberration, the

Member said:

fringe may be asymmetrical in shape and is usually of one color all
the way around. In this case, as in many cases, magenta or purple.

Click to expand...

I had been under the assumption that the common "purplish" or red color fringing was due to true chromatic aberration, i.e., slight difference in the focus point of different wavelengths (colors) of light due to dispersion. As red and violet are the two extremes (in terms or refractive index) I'd assumed that was why they were the most prevalent aberration.

Is it then a function of a camera's lens/microlens optical path that the extremes are brought to focus, and the median colors, (the yellow-green that you mentioned) are the colors where chromatic aberration is most visible?

As for blooming, this "leaking" charge, is it a purely an electronic phenomenon as opposed to an optical one? And why would the leaking charge be a certain color?

Regarding Mr. Askey's post, Since this is due to the lens/microlens interaction, that explains why film cameras generally exhibit less CA -- there is simply less lenses to correct for, right?

Thanks all, I mean no offense, I just like to understand as well as I'm able.

Be well. Take pictures. Enjoy life.

F.J.

Actually, achromatc lenses typically are designed to bring two colors (wavelengths of light) into focus. For optics designed for the human eye, these colors are typically green and yellow. This leaves the yellow-green to be out of focus in one diection while the reds and blues go out of focus the other way. From the color fringing in digital cameras, I would conclude that the fringing from the primary lens would be designed for focus at similar wavelengths.

The microlenses in front of the CCD are not achromats, so I would expect them to exhibit the chromatic aberrations typical of simple lenses, with the fringes colored red-orange and blue-violet. Form this I conclude that the chromatic aberration of the primary lens does indeed contribute to the color fringing.

What I called "blooming" in my previous post may be halation. The microlenses may indeed contribute some color fringing, but I find the purple color of these fringes difficult to explain by that mechanism. I would instead expect the fringes either to be blue-violet or red-orange but not purple.

The asymmetry I mentioned occurs because the image is shifted off the photosensor after the picture is taken. This may result in preferential blooming in the direction of the shift.

BTW, after seeing, the picture again, the highlight also shows definite signs of lens flare, giving the top edge of the plane's body highlight a scalloped appearance. You can also see evidence of "rays" extending from the cusp of each scallop toward the upper right of the photo.

F.J. said:

Regarding the following, (edited)

...Unlike the color fringing from chromatic aberration, the

F.J.41613 said:

fringe may be asymmetrical in shape and is usually of one color all
the way around. In this case, as in many cases, magenta or purple.

Click to expand...

I had been under the assumption that the common "purplish" or red
color fringing was due to true chromatic aberration, i.e., slight
difference in the focus point of different wavelengths (colors) of
light due to dispersion. As red and violet are the two extremes
(in terms or refractive index) I'd assumed that was why they were
the most prevalent aberration.

Is it then a function of a camera's lens/microlens optical path
that the extremes are brought to focus, and the median colors, (the
yellow-green that you mentioned) are the colors where chromatic
aberration is most visible?

As for blooming, this "leaking" charge, is it a purely an
electronic phenomenon as opposed to an optical one? And why would
the leaking charge be a certain color?

Regarding Mr. Askey's post, Since this is due to the lens/microlens
interaction, that explains why film cameras generally exhibit less
CA -- there is simply less lenses to correct for, right?

Thanks all, I mean no offense, I just like to understand as well as
I'm able.

Be well. Take pictures. Enjoy life.

F.J.

Click to expand...

Phil -

Thank you very much for your input... it's great that
you take the time to notice what's going on the the
Forums, and to directly participate.

There is an old saying that "A rose by any other name
would smell as sweet." This can be taken as an argument
that terminology isn't important. But it IS, if language is
to work as a communications tool.

If our "community" got together and decided to call roses--say--"backhoes," a lot of mischief would result. Not only
would those outside our community be VERY confused
when we talked about our beautiful bouquets of backhoes,
but what would we then call a bona-fide backhoe?

Since our digicams still use lenses, and some of those lenses
have (true) CA, I would like to see "chromatic aberration"
keep its well established optical definition; and we would
then coin new and appropriate terms to describe the defects
of solid-state imagers (e.g., "blooming").

Phil Askey said:

In fact it's caused by the direct interaction between the main lens
system and the microlenses which sit over every pixel of the CCD.
It's to do with very bright rays of light refracting off the edges
of these microlenses.

Gerry said:

The first shows a severe case of what Phil Askey has called
chromatic aberration. I have been uncomfortable with this name for
the phenomenon, because I associate the term "chromatic aberration"
with a common linear phenomenon arising from the variation of
refractive index of a transparent medium as a function of
wavelength.

Click to expand...

Click to expand...

Alan Ball said:

The top photo, which I assume was cropped from an area near the
edge of the frame, definitely exhibits both chromatic aberration
and blooming. The bottom photo, clipped from an area near the
center, shows no chromatic aberration. Ths photo may also show
some flare as well.

Click to expand...

Alan Ball is absolutely right. Calling the left and top of the original 1600x1200 images 1,1 and the right bottom 1600.1200, the image with the bad bad color had the plane centered around 1170,200 and the one with the less bad color had the plane nearly in the middle at 885,768.

I am going to look out for this relationship that Alan suggests between position and the effect. By the way, the actual focal length of the lens is said by the camera maker to be 70 mm. on a 2.1 Mp chip.

I am grateful to all those who have picked up this theme and contributed to it. You have all enlightened me, but unhappily not enough. I think we and the camera makers still have a way to go in understanding precisely what is going on here. And if anyone can come up with a better name, please do!

Thanks very much.

You don't have to understand exactly what is going on in this case to fix the problem in postprocessing. In this case it just means resizing the blue channel slightly so its image corresponds more closely to the other two channels. If your image edtior can't do this directly, you can do it with the Photoshop-compatible plug-in called PanoTools, the same one Vincent Bockaert suggests for barrel distortion correction. (See the Learn -> Image Techniques -> Barrel Distortion Correction entry on this site.) Changing the d parameter of the blue channel by a few tenths of a percent will do the trick.

No such easy fix for blooming or lens flare, unfortunately.

Gerry said:

I am going to look out for this relationship that Alan suggests
between position and the effect.

Click to expand...

Member said:

I am grateful to all those who have picked up this theme and
contributed to it. You have all enlightened me, but unhappily not
enough. I think we and the camera makers still have a way to go in
understanding precisely what is going on here. And if anyone can
come up with a better name, please do!

Thanks very much.

Click to expand...

Old Ed said:

Since our digicams still use lenses, and some of those lenses
have (true) CA, I would like to see "chromatic aberration"
keep its well established optical definition; and we would
then coin new and appropriate terms to describe the defects
of solid-state imagers (e.g., "blooming").

Click to expand...

I strongly agree with you for another reason. It is possible to remove the

effects of true CA with some software that will do a differential radial re-sizing
of the RGB images, but these filters do nothing really to handle blooming (I
have been calling it sensor "overload"). There are some other techniques
and filters that will reduce the blooming somewhat. It has become difficult
to help people who want to use these tools, when you can't tell which problem
they wish to work on.

Darrell

Alan Ball said:

Actually, achromatc lenses typically are designed to bring two
colors (wavelengths of light) into focus. For optics designed for
the human eye, these colors are typically green and yellow. This
leaves the yellow-green to be out of focus in one diection while
the reds and blues go out of focus the other way. From the color
fringing in digital cameras, I would conclude that the fringing
from the primary lens would be designed for focus at similar
wavelengths.

Click to expand...

Right, right -- the eye is most sensitive to green. Cameras would obviously be designed with this in mind, which goes right along with there being a greater number of "green-sensing" elements in the array.

Member said:

The microlenses in front of the CCD are not achromats, so I would
expect them to exhibit the chromatic aberrations typical of simple
lenses, with the fringes colored red-orange and blue-violet. Form
this I conclude that the chromatic aberration of the primary lens
does indeed contribute to the color fringing.

Click to expand...

Okay, makes sense, and I should've known that the microlenses are not achromatic -- too difficult/expensive to produce.

Member said:

What I called "blooming" in my previous post may be halation. The
microlenses may indeed contribute some color fringing, but I find
the purple color of these fringes difficult to explain by that
mechanism. I would instead expect the fringes either to be
blue-violet or red-orange but not purple.

The asymmetry I mentioned occurs because the image is shifted off
the photosensor after the picture is taken. This may result in
preferential blooming in the direction of the shift.

Click to expand...

Yes, I get that, too; whew, there is alot contributing to these effects isn't there?

Thank you for being kind enough to answer my questions!

Sincerely,

F.J.

Member said:

BTW, after seeing, the picture again, the highlight also shows
definite signs of lens flare, giving the top edge of the plane's
body highlight a scalloped appearance. You can also see evidence
of "rays" extending from the cusp of each scallop toward the upper
right of the photo.

Click to expand...