CA reducing action - looks like the A2 is gonna need this!

William R Stirling said:

Tony SD said:

Err... hang on... Unless my 30+ years as a photomicographer have
been wasted:

Red and BLUE are at the opposite ends of the visible spectrum, and
achromatic lenses correct for these two extremes. Poorly corrected
lenses will show blue fringing one side of an image, and red the
other. Good achros correct well for both these wavelengths, but
leave a small out of correction element for green in the middle
(550nm) Apos then correct for this middle frequency.

I didn't think it was anything to do with violet - but hey! it's
years since I've done any of this so I may well be wrong!
--
TonySD

Click to expand...

Glasses of normal dispersion, which have an almost linear decrease
in refractive index with increasing wavelength, are used to produce
achromat objectives. Only two wavelengths can have the same focus,
and the remaining secondary spectrum produces greenish or purple
(violet) fringes on images of sharp edges. The higher quality
apochromat objectives use glasses having a partial dispersion where
the refractive index changes with wavelength more rapidly in either
the blue or red region. As a result, apochromats have a high degree
of chromatic correction in which up to four wavelengths can have
the same image location.

--
cheers!
Rick Stirling
http://clanstirling.net/gallery/

Click to expand...

--

Rick, Tony D and Dave, I have a question for you. Recently I have noticed a fair amount of purple fringing, purple on one side of an object and green on the other with my A1 at full wide angle. I suspect this is exascerbated by the bright skies at the high altitude I live at (7000 feet MSL in Colorado) and my habit of taking pictures of trees. The effect diminishes with higher f-stop and with the lens zoomed above 50 mm equiv, but is quite noticeable even at f5.6 and 40 mm equivalent. Here is a sample for f 5.6 and 28 mm equiv. Note the purple on the right of the pine trunk and the green on the left:



I recently completed several tests, varying the exposure (moving the brightest peak in the histogram more towards the center), and I still see purple fringing in almost every wide angle shot with bright skies and trees. Of course this often goes away when the picture is shot so that the sun or bright sky light source is behind me.

Am I seeing normal behavior for this lens?

Incidentally, I have been posting for a month and reading forum contributions since November and have appreciated the contributions each of you have made to this site.

Regards,

Pat

I have been looking at the recent lineup of 8MP cameras recently announced, AND all the sample images (if there are any). I owned a Minolta 7Hi for awhile, and recently I had the Sony 828 for 3 days, and returned it due to excessive purple fringing (CA, whatever you want to call it..........it's PURPLE). Who cares what it's called, it's not supposed to be there at least to that degree. Those A2 samples from Singapore show lots of purple....IF you look at them in original size. Try that, and you won't ask where it is. K.

William R Stirling said:

David Martin said:

See:
http://www.shaystephens.com/CA_Fixes.atn
Just duplicate the layer, apply the action, then erase back where
it is not needed
--
Regards,
DaveMart
Please see profile for equipment

Click to expand...

Sure wish someone would point me to an A2 image with CA. I haven't
been able to find one yet.

--
cheers!
Rick Stirling
http://clanstirling.net/gallery/

Click to expand...

--
Galleries: http://www.koo22photos.com/-/koo22photos/default.asp

fotog said:

It was an inflamatory troll-like post. We have barely seen any
samples come out of the camera, and he's sized it up. It's
especially troll-like since it was put in the subject line. Yes, he
seems well-versed in photogaphy, and yes, I have seen helpful posts.

And then we have a post pointing A2 owners to an action to reduce
PF. If it walks, talks, like a troll, guess what?

Click to expand...

Maybe I am another troll but, man, if you don't like what he writes, just stop reading it or at the very least, stop bashing. Ad hominem arguments are really not useful in a discussion.

Beside, PF reducing softwares may become very useful with cameras using Sony's 8 MP ccd and I thank DaveMart for his suggestion.

And FYI, I own a Minolta A1 and plan to change it for the A2.

jpmar

elk13 said:

William R Stirling said:

Tony SD said:

Err... hang on... Unless my 30+ years as a photomicographer have
been wasted:

Red and BLUE are at the opposite ends of the visible spectrum, and
achromatic lenses correct for these two extremes. Poorly corrected
lenses will show blue fringing one side of an image, and red the
other. Good achros correct well for both these wavelengths, but
leave a small out of correction element for green in the middle
(550nm) Apos then correct for this middle frequency.

I didn't think it was anything to do with violet - but hey! it's
years since I've done any of this so I may well be wrong!
--
TonySD

Click to expand...

Glasses of normal dispersion, which have an almost linear decrease
in refractive index with increasing wavelength, are used to produce
achromat objectives. Only two wavelengths can have the same focus,
and the remaining secondary spectrum produces greenish or purple
(violet) fringes on images of sharp edges. The higher quality
apochromat objectives use glasses having a partial dispersion where
the refractive index changes with wavelength more rapidly in either
the blue or red region. As a result, apochromats have a high degree
of chromatic correction in which up to four wavelengths can have
the same image location.

--
cheers!
Rick Stirling
http://clanstirling.net/gallery/

Click to expand...

--

Rick, Tony D and Dave, I have a question for you. Recently I have
noticed a fair amount of purple fringing, purple on one side of an
object and green on the other with my A1 at full wide angle. I
suspect this is exascerbated by the bright skies at the high
altitude I live at (7000 feet MSL in Colorado) and my habit of
taking pictures of trees. The effect diminishes with higher f-stop
and with the lens zoomed above 50 mm equiv, but is quite noticeable
even at f5.6 and 40 mm equivalent. Here is a sample for f 5.6 and
28 mm equiv. Note the purple on the right of the pine trunk and
the green on the left:

I recently completed several tests, varying the exposure (moving
the brightest peak in the histogram more towards the center), and I
still see purple fringing in almost every wide angle shot with
bright skies and trees. Of course this often goes away when the
picture is shot so that the sun or bright sky light source is
behind me.

Am I seeing normal behavior for this lens?

Incidentally, I have been posting for a month and reading forum
contributions since November and have appreciated the contributions
each of you have made to this site.

Regards,

Pat

Click to expand...

Aha! I think Rick and Tony will agree! Whatever it might be in the other shots, that is a description of classic CA - it should occur towards the edges of the frame, mostly wide open in bright sunlight, and each purple fringe has it's matching fringe!

If it's any consolation, the Minolta lens is a lot less subject to it than many - I've seen some real good examples from Canon primes costing several thousand!

You mainly just have to be aware of the situations which can cause it, I am afraid, and try stopping down in those situations etc.
The PF action I gave will aslo help.
--
Regards,
DaveMart
Please see profile for equipment

fotog said:

I agree. My error. I should have made my point in a more subtle
way. In the context of friends, I can now see that your post was
not intended to flame. At the same time I think you would agree
that there are lots of newbies like myself, some of whom post and
many who lurk, and they might not appreciate your subject line if
they had just ordered the camera.

I spent quite a bit of time lurking here on several forums, (Fuji,
Sony, Minolta) and a trend I have seen too often is to bash the new
cameras, and even their prospective users. Sometimes in subtle
ways, others more overt. New, enthusiastic faces appear to announce
their new purchase. And then they don't post again.

I have ordered an A2, and I can't wait to receive it. Then I will
do a head-to-head with my Fuji S7000, and post my conclusions.

I have been involved in photography for over 20 years, from 35 up
to 4x5. But I have only a year's experience with digtal. Even so, I
hope to be a valuable addition to these forums.

--



Beauty is in the eye of the Bee-holder...

Click to expand...

I've got the 10D, and whenever I am talking about it never fail to mention that the AWB is pants.

That's just the way it is, and whether I own it or not, it will still be bad in that respect.
Canon can build decent AWB and have merely chosen not to.

All these 8MP sensor cameras will do a fine job for many, and in skilled hands and the right light will give shots with great detail which can be blown up hugely.

Unfortunately, in the rush for more megapixels all the manufacturer's have lperhaps left behind the needs of much of their market, as many will wnat a more forgiving camera and rarely if ever go to very large sizes.

Then there are people like me who need a second camera, and could do with a more flexible package as if they are going to go to huge size will use a DSLR anyway.

The thing which has made this a great forum though, is that people can relax here and express their thoughts withour fearing that they will be judged too harshly, and until it becomes glaringly and repeatedly apparent that there is malice aforethought in their comments will not be judged or flamed.

Some forums here are virtually unusable because of the personal animus there, and the chains of accusation and counter accusation.

It all contributes to an unpleasant place to hang out, and a nasty, nasty atmosphere.

I would hate this forum to become such a place, and I'm sure we will just bumble along, rubbing along together, with a great deal of tolerance for each others foibles.
Congrats on your new camera. I am sure you will have great time
--
Regards,
DaveMart
Please see profile for equipment

koo22 said:

William R Stirling said:

David Martin said:

See:
http://www.shaystephens.com/CA_Fixes.atn
Just duplicate the layer, apply the action, then erase back where
it is not needed
--
Regards,
DaveMart
Please see profile for equipment

Click to expand...

Sure wish someone would point me to an A2 image with CA. I haven't
been able to find one yet.

--
cheers!
Rick Stirling
http://clanstirling.net/gallery/

Click to expand...

--
Galleries: http://www.koo22photos.com/-/koo22photos/default.asp

Click to expand...

It's got rather different causes, Bill was quiter right,I was too hasty in labelling it as CA, so it is important because avoiding it is slightly different.

If it is just PF caused by gross underexposure, and the sensor is implicated, all you have to do is be more careful to nail the exposure, and watch for overexposed areas.

Avoiding CA is a bit more tricky, you try to stay away from wide open if bright light and it tends to be worse at very WA and telephoto
--
Regards,
DaveMart
Please see profile for equipment

The A2 is no worse than the A1 or the D7 series.

The older cameras are not devoid of chromatic aberration or purple fringing. They merely control it well.

Jawed

Jawed said:

The A2 is no worse than the A1 or the D7 series.

The older cameras are not devoid of chromatic aberration or purple
fringing. They merely control it well.

Its not clear to me how you can be so sure of that . Surely we havn't seen enough examples yet. The A2 may have the same lens but the question mark is over the sensor!--

Click to expand...

Keith-C

Simple really. The so-called "bad" results produced by the A2 are no worse than the A1 or D7 series cameras in the same circumstances.

The older cameras produce bad results too. For some reason the F828 is in a league of its own when it comes to purple fringing. The F828 has pretty good chromatic aberration performance, overall.

Some of the A2 pix already presented are the worst case possible. Very bright sky against dark branches. The older cameras would behave the same with the same subject.

Anyway, the comparisons will soon be in and everyone will be able to relax. If I was buying an A2 I wouldn't be worried about this.

Jawed

KEITH-C said:

Jawed said:

The A2 is no worse than the A1 or the D7 series.

The older cameras are not devoid of chromatic aberration or purple
fringing. They merely control it well.

Its not clear to me how you can be so sure of that . Surely we havn't seen enough examples yet. The A2 may have the same lens but the question mark is over the sensor!--

Click to expand...

Keith-C

Click to expand...

For those who don't have access to PS actions, I've written a fringe-reducing plugin that may be helpful. See
http://forums.dpreview.com/forums/read.asp?forum=1009&message=7606575
for download details.

It operates in a somewhat different fashion from Shay's action (which I haven't had the opportunity to test) so it may work in some cases where the action doesn't.

Tom

David Martin said:

See:
http://www.shaystephens.com/CA_Fixes.atn
Just duplicate the layer, apply the action, then erase back where
it is not needed
--
Regards,
DaveMart
Please see profile for equipment

Click to expand...

Haven't tried it, but I'm guessing the CA fix in raw works like the same in PanoTools - i.e. it assumes conventional transverse CA, where one channel (e.g. blue) focuses at a greater magnification than others, and can be realigned by a slight scaling of the image for that channel.

Unfortunately, this is not what's happening with PF/CA on the 828, and the blue fringes on the branches in the sample posted above look very similar to the 828 results. As far as I can tell there are two separate phenomena (1) a fringing effect due to some kind of opto/electronic effect around blown highlights, which can occur anywhere in the frame affecting both red and blue channels, hence the purple appearance from white highlights and blue fringing against skies where only blue is overexposed, and (2) a CA-type effect in the blue channel, occuring in the corners, mostly at wide angle, but as a secondary displaced image, while the primary image in the blue channel is still sharp and aligned with the other channels.

Personally, I find that only type (1) is really troublesome, as type (2) can be avoided by stopping down if you have high contrasts in the corners at wide angle - an infrequent occurrence for me. I avoid type (1) by minimizing overexposed highlights - particularly where they are small and lie against a dark background (bright water drops on a leaf for example, or sticks against sky). Neither is debilitating. If these effects are really less on the A2, it will be fantastic, but even if they're the same as on the 828 it will be a great camera.

Tricia said:

David Martin said:

See:
http://www.shaystephens.com/CA_Fixes.atn
Just duplicate the layer, apply the action, then erase back where
it is not needed

Click to expand...

The RAW converter in Photoshop CS has CA corrections for R/C and
B/Y. I'd check it out, but I don't have any RAW photos with CA. If
anyone finds one, let me know and I'll see how well it works.

--
Tricia

Click to expand...

But for those who can understand such things, the following post may be interesting (WARNING: Technical content - the file size is way too big for my brain, and causes overflow)
http://forums.dpreview.com/forums/read.asp?forum=1009&message=7762893

It does seem to my limited understanding as though we may be approaching some practical limits to pixel size, and we are having to make increasingly heavy trade-offs to increase resolution,

TomFid said:

Unfortunately, this is not what's happening with PF/CA on the 828,
and the blue fringes on the branches in the sample posted above
look very similar to the 828 results. As far as I can tell there
are two separate phenomena (1) a fringing effect due to some kind
of opto/electronic effect around blown highlights, which can occur
anywhere in the frame affecting both red and blue channels, hence
the purple appearance from white highlights and blue fringing
against skies where only blue is overexposed, and (2) a CA-type
effect in the blue channel, occuring in the corners, mostly at wide
angle, but as a secondary displaced image, while the primary image
in the blue channel is still sharp and aligned with the other
channels.

Personally, I find that only type (1) is really troublesome, as
type (2) can be avoided by stopping down if you have high contrasts
in the corners at wide angle - an infrequent occurrence for me. I
avoid type (1) by minimizing overexposed highlights - particularly
where they are small and lie against a dark background (bright
water drops on a leaf for example, or sticks against sky). Neither
is debilitating. If these effects are really less on the A2, it
will be fantastic, but even if they're the same as on the 828 it
will be a great camera.

Tricia said:

David Martin said:

See:
http://www.shaystephens.com/CA_Fixes.atn
Just duplicate the layer, apply the action, then erase back where
it is not needed

Click to expand...

The RAW converter in Photoshop CS has CA corrections for R/C and
B/Y. I'd check it out, but I don't have any RAW photos with CA. If
anyone finds one, let me know and I'll see how well it works.

--
Tricia

Click to expand...

Click to expand...

--
Regards,
DaveMart
Please see profile for equipment

I've added a 2nd plugin that geometrically corrects blue fringes in the corners of non-overexposed images. See
http://www.sd3.info/pf828/

It might be more appropriate for the sample that lead this thread.

I've shot with my A1 for a few months and find CA to be hardly noticable in the worst circumstances. When it is noticeable it is the wide angle variation (radial) than can be easily corected by a few programs. That's completely different from what I've seen with the F828 (tried 3 different ones). The A2 is more prone to edge CA at any focal length in comparison to the A1, but its CA - not the ugly PF that can appear anywhere in the picture of the F828. There is a big difference when it comes to post processing and final image quality.

Paul

Jawed said:

The older cameras produce bad results too. For some reason the
F828 is in a league of its own when it comes to purple fringing.
The F828 has pretty good chromatic aberration performance, overall.

Some of the A2 pix already presented are the worst case possible.
Very bright sky against dark branches. The older cameras would
behave the same with the same subject.

Anyway, the comparisons will soon be in and everyone will be able
to relax. If I was buying an A2 I wouldn't be worried about this.

Jawed

KEITH-C said:

Jawed said:

The A2 is no worse than the A1 or the D7 series.

The older cameras are not devoid of chromatic aberration or purple
fringing. They merely control it well.

Its not clear to me how you can be so sure of that . Surely we havn't seen enough examples yet. The A2 may have the same lens but the question mark is over the sensor!--

Click to expand...

Keith-C

Click to expand...

Click to expand...