Heavy CA on a HDR spherical panorama !

27 views
Skip to first unread message

JPS

unread,
Dec 26, 2012, 10:43:28 AM12/26/12
to pt...@googlegroups.com
Hi ! Longtime I haven't done any 360x180, but since I got a D800 now, I wanted to see if the result would be much better (sharper) than what I used to get with my "old" D700... I use a Nikon 10.5 f/2.8 DX (shaved) and shoot 4 images plus one for the Nadir !

I load my RAW images (3 images bracketed -1, Ø, +1 EV) in PTGui and veryting goes well, except I notice a very heavy CA, -lot of red and some green-, wich I cannot correct, as it happens on the left as well as on the right of objects ! If I correct for one side, it increases on the other side...

Would it be better to load my RAW image in CaptureNX2 or Photoshop and correct each image' CA, then convert them to TIF before to load them in PTGui ? The drawback is that an allready heavy D800 RAW file is +/- 30 Mo., and once converted to TIF, it becomes a more than 100 Mo . !!! When one try to fill up PTGui with 15 files of 100 Mo., it takes aaaaaages to work them out !

What is the best way to get rid of CA in the final spherical pano ?

TIA, and..... HAPPY NEW YEAR !!
J-P.


Henrik

unread,
Dec 26, 2012, 11:30:16 AM12/26/12
to pt...@googlegroups.com
Did you not used to get CA with this lens on your old camera?

You will be best off to process your raw files before loading them in PTGui, and eliminate any artifacts there first.

I am not familiar with your Nikon software, sorry

All the best

Henrik

Sent from my iPhone
--
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 upload your files at a file sharing site (for example http://ge.tt/ ) and include a link in your message.
For more options, visit this group at http://groups.google.com/group/ptgui

Geoff - Spherical Visions

unread,
Dec 26, 2012, 4:14:35 PM12/26/12
to pt...@googlegroups.com
CA or more correctly TCA is a radial problem in that the focus point varies in a radial fashion so it is only possible to correct in the original images where the centre of the radial point for each image is known, you cannot correct after the images have been warped and blended.  See http://wiki.panotools.org/Chromatic_aberration
Use the Capture NX, software to correct the CA and any other pre-processing required on the RAW images and then output to TIF ready for PTGui.

Good luck!

JPS

unread,
Dec 27, 2012, 5:48:04 AM12/27/12
to pt...@googlegroups.com
Thanks Geoff !

....that's about what I was afraid of ! I must first get rid of the CA, THEN load the files in PTGUI !

Just thought of a possibility: the D800 RAW files being around 30 Mo. and once converted to TIF they become more than 100Mo..... Can't get rid of the CA (plus a few other corrections in CaptureNX2), then RE-SAVE AS RAW files, getting files about 40 Mo. "only" !!! ...so I would have much smaller files to work for PTGui !

Do you think that would make the process faster ?

:-)
J-P.

Geoff - Spherical Visions

unread,
Dec 27, 2012, 6:06:14 AM12/27/12
to pt...@googlegroups.com
Your workflow with RAW files in NX? would be to add the required corrections to the parameters and then to save the RAW files with there parameters (Remember with RAW you don't actually change the image merely list the parameters to be applied when the image is drawn on the screen) then you can save the resulting images as TIF for working on with PTGui at the end of the process once you have the equirectangular you can delete the TIF files as they can always be recreated from the smaller RAW files and parameters.
It probably isn't a good idea to load the RAW files into PTGui as what your doing is calling dcraw to do the conversion and then load the result into PTGui, dcraw doesn't do as good a job as NX will !
Another factor to remember is that the actual memory space (RAM) used won't alter its merely the size of the compressed file on disk that appears to change, once a given image is stored in RAM its the number of pixels*3*colour depth so pixels*3*16 in  most cases.

Reply all
Reply to author
Forward
0 new messages