Chromatic aberration in final output image

755 views
Skip to first unread message

Debabrata Roy

unread,
Aug 13, 2010, 11:36:16 AM8/13/10
to pt...@googlegroups.com

As a photographer I make partial panos for prints. But I am relatively new to this. My problem is, there is presence of chromatic aberration in every panos I create. I shoot RAW and directly use them in PTGui Pro. My understanding is, in the process of stitching PTGui fixes chromatic aberration.


May be I am doing something wrong, but not sure what I should be doing. Any help is greatly appreciated. I have attach an example (2 images, one detail & one full frame).


Thanks


Roy


http://img444.imageshack.us/img444/2519/detailq.jpg

http://img683.imageshack.us/img683/3595/detailfull.jpg

Willy Kaemena

unread,
Aug 13, 2010, 12:00:48 PM8/13/10
to pt...@googlegroups.com
My  strong recommendation is to open  your RAWs in Photoshop or any other program able to treat  C.A. Then  safe it as TIFF 16bit and  put these into PTGui
That  would give  you the highest possible quality and no CA

Willy


--
You received this message because you are subscribed to the Google Groups "PTGui" group.
To post to this group, send email to pt...@googlegroups.com
To unsubscribe from this group, send email to ptgui+un...@googlegroups.com
Please do not add attachments to your posts; instead you may upload files at
http://groups.google.com/group/ptgui/files
For more options, visit this group at http://groups.google.com/group/ptgui

John Houghton

unread,
Aug 13, 2010, 12:03:45 PM8/13/10
to PTGui Support
On Aug 13, 4:36 pm, Debabrata Roy <fotob...@gmail.com> wrote:
> I shoot RAW and directly use them in PTGui Pro. My
> understanding is, in the process of stitching PTGui fixes chromatic
> aberration.

I'm afraid you are mistaken. No corrections that you make in Adobe
ACR or other RAW converter will be applied by PTGui when stitching
from RAW. Generate corrected images in TIFF or other supported format
for stitching.

John

Roy

unread,
Aug 13, 2010, 3:18:16 PM8/13/10
to PTGui Support
Thank you both (Willy & John)

I was hoping I can avoid preprocessing in Photoshop/Lightroom/Bridge. :
(

Thanks for correcting me :)

Roy

Willy Kaemena

unread,
Aug 13, 2010, 4:25:43 PM8/13/10
to pt...@googlegroups.com
The "preprocessing" gives you the huge advantages when shooting in RAW ( exposure, WB,CA,noise, whatever went wrong or not the best way can be easily corrected in the processing/transformation of the RAW files, even it is possible to extract several exposures and treat them as it would be a set of bracketed pictures for HDR !. These are the reasons to shoot RAW!!

Willy

Roy

unread,
Aug 13, 2010, 11:27:14 PM8/13/10
to PTGui Support
Point well noted :)

Roy

Tom Sharpless

unread,
Aug 15, 2010, 12:39:17 PM8/15/10
to PTGui Support
I second Roy's request.

Fixing TCA is important for best resolution with short lenses. Doing
that while warping is just a matter of adding a little radial
correction depending on the color. PTGui should be able to do that (I
believe libpano always has supported this).

The hard part is finding the right parameters. Photoshop has a very
nice interactive tool for that, so there is no need for Joost to
invent any new gui -- just learn to interpret Photoshop's parameter
values correctly, and put the option of using them in the stitcher.

Regards, Tom

Joergen Geerds

unread,
Aug 15, 2010, 12:49:26 PM8/15/10
to PTGui Support
On Aug 15, 12:39 pm, Tom Sharpless <tksharpl...@gmail.com> wrote:
> The hard part is finding the right parameters.  Photoshop has a very
> nice interactive tool for that, so there is no need for Joost to
> invent any new gui -- just learn to interpret Photoshop's parameter
> values correctly, and put the option of using them in the stitcher.

wouldn't it be great if dcraw would honor/read/use the XMP file from
adobe? at least for some key parameters like WB, CA etc?
I think that it would be possible for ptgui to "interpret" the xmp
values while feeding the raw files through dcraw, but I think Dave
Coffin should simply implement that feature into dcraw, so joost
wouldn't have to deal with it.

joergen

Ken Warner

unread,
Aug 15, 2010, 12:54:44 PM8/15/10
to pt...@googlegroups.com
If you have Adobe stuff (Photoshop, Lightroom, etc) why would you
care about dcraw?

Erik Krause

unread,
Aug 15, 2010, 1:02:30 PM8/15/10
to pt...@googlegroups.com
Am 15.08.2010 18:39, schrieb Tom Sharpless:

> The hard part is finding the right parameters. Photoshop has a very
> nice interactive tool for that, so there is no need for Joost to
> invent any new gui -- just learn to interpret Photoshop's parameter
> values correctly, and put the option of using them in the stitcher.

I thought about a project to do that: Create artificial raw test images
(DNG in this case - already done) with defined CA, then use different
raw converters to correct it. This would f.e. enable PTGui to read XMP
files and pass the appropriate correction values to dcraw. A first step
would be a place in PTGui to enter the required values for the dcraw
command line...

BTW, dcraw CA correction values can be determined from normal images as
well: http://wiki.panotools.org/Tca_correct#Using_with_dcraw

--
Erik Krause
http://www.erik-krause.de

Joergen Geerds

unread,
Aug 15, 2010, 1:06:10 PM8/15/10
to PTGui Support
On Aug 15, 12:39 pm, Tom Sharpless <tksharpl...@gmail.com> wrote:
> The hard part is finding the right parameters.  Photoshop has a very
> nice interactive tool for that, so there is no need for Joost to
> invent any new gui -- just learn to interpret Photoshop's parameter
> values correctly, and put the option of using them in the stitcher.

some time ago we had a discussion about this here on the list. one
suggestion was to make the command line for dcraw visible and
editable... i.e. making it a input line in the lens parameter pane,
showing the default parameters that ptgui is already using, plus
letting the user enter specifics for the 3 CA parameters, which can be
derived from the XMP, maybe using some manual 1/x math (dunno how
exactly)... this would be a hack that would work only for projects
that use only one lens, which should be most projects anyway.

this hack would require almost no work from joost, and would put all
the customization power into the hands of the users who don't want to
develop their RAWs into tiffs (I love tiffs). If I were Joost, I would
even put a big disclaimer next to the dcraw parameter line saying
"proceed at your own risk, and don't ask me for support for dcraw
parameters, rtfm" :-)

joergen

Joergen Geerds

unread,
Aug 15, 2010, 1:18:32 PM8/15/10
to PTGui Support
On Aug 15, 12:54 pm, Ken Warner <kwarner...@verizon.net> wrote:
> If you have Adobe stuff (Photoshop, Lightroom, etc) why would you care about dcraw?

while I am the #1 Adobe fanboy on this list, there are enough people
that either don't want to use/own Adobe products, or can't afford
them... it's not up to me to judge them... I personally still think
that it was an unnecessary step to include dcraw into ptgui.

joergen

Tom Sharpless

unread,
Aug 15, 2010, 9:17:48 PM8/15/10
to PTGui Support
Hi Joergen

It would be good if PTGui exposed dcraw's parameters.

But what I was talking about was something simpler, that would work
for jpegs too: simply adjusting the magnifications of the red and blue
images slightly (which is basically all that dcraw or Camera Raw do
for TCA). That can be added to the normal stitching process at very
little cost or effort.

Regards, Tom

Erik Krause

unread,
Aug 16, 2010, 9:21:46 AM8/16/10
to pt...@googlegroups.com
Am 16.08.2010 03:17, schrieb Tom Sharpless:
> But what I was talking about was something simpler, that would work
> for jpegs too: simply adjusting the magnifications of the red and blue
> images slightly (which is basically all that dcraw or Camera Raw do
> for TCA). That can be added to the normal stitching process at very
> little cost or effort.

Currently you can use the Correct filter on Source Images tab (32 bit
versions only with installed panotools library of course). Suitable
correction parameters can be determined with tca_correct or PTShift.
However, correction after raw conversion is only good if the CA fringes
are much wider than the bayer pattern. Otherwise you get something like
http://wiki.panotools.org/TCA#Strange_Example

Reply all
Reply to author
Forward
0 new messages