HowTo? Calibrate fisheye lens on panohead?

257 views
Skip to first unread message

l_d_allan

unread,
Oct 13, 2012, 7:02:52 PM10/13/12
to pt...@googlegroups.com
I'm unclear how to calibrate a fisheye lens on a panohead. Does the offset vary with the amount of overlap?
 
I've ordered an 8mm Samyang fisheye for a Canon T3i 1.6x crop camera. Is the offset for NPP fixed, or does it vary depending on whether I have a 50%, 40%, 30%, 20%, etc. overlap?
 
I've read several articles on LPP (least parallax point) on fisheye lenses, contrasted to NPP (no parallax point) on rectilinear lenses. I'm fuzzy on just how to proceed.
 
With a rectilinear lens, my understanding is that the NPP doesn't change, regardless of the overlap. With a 14mm rectilinear Samyang on a full frame camera, the offset for NPP remains the same whether I have 6 frames for a 360 degree FOV pano in portrait orientation, or 7 frames, or 8 frames. I found the NPP with the camera+lens in Landscape mode for the greatest amount of "swing distance FOV" from left to right. Correct?
 
But with a fisheye, it seems like the best offset for LPP might be different with 3 frames for 360 degrees FOV vs 4 frames. Correct?
 
So is the "best practice" to standardize on a specific overlap, and find the LPP for that overlap? I suppose that would also mean having the camera+lens in portrait mode on the panohead when doing the panohead calibration? Correct, or am I "unclear on the concept"?
 
 
 
 

Geoff - Spherical Visions

unread,
Oct 14, 2012, 7:57:50 AM10/14/12
to pt...@googlegroups.com
Yes with FE lenses the NPP varies with angle, I have two settings recorded for my Sigma 8mm, 45° - 8 shots and 90° - 4 shots around, there being 11mm difference between the two settings. Rectilinear lenses are much easier to setup!  Yes you need to do the setup with camera and lens in the actual positions you will use, its not just angular rotation that is important but also alignment to the axis of rotation that is critical.

Do also consider with FE lenses that going for the minimum number of shots may not be the best solution, the difference between taking 3/4 and 6/8 shots around is pretty small in time 10-15 seconds, however having additional shots means that ghosting and movement is much easier to resolve and can save you hours of work in post processing!  I rarely use the 4 shot + Nadir option, much preferring 8 shots + Nadir with the Sigma 8mm.

Mike Dols

unread,
Oct 14, 2012, 2:08:14 PM10/14/12
to pt...@googlegroups.com
Check THIS link for a tutorial about a panohead calibration with a FE lens. Hope it helps.

Good luck!

PTGui Support

unread,
Oct 15, 2012, 4:04:46 AM10/15/12
to pt...@googlegroups.com
Hi Lynn,

I think you shouldn't worry about this too much. Yes, for fisheye lenses
there is no single no-parallax point but the deviation is very little.
Follow the calibration tutorials and adjust the head until parallax is
minimal.

You may try to overlap for a specific overlap percentage but keep in
mind that the seam line would have to be a perfect circle in the source
image, but this is not the case.

Joost
> --
> 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

l_d_allan

unread,
Oct 20, 2012, 8:21:15 PM10/20/12
to pt...@googlegroups.com
Thanks for the clarification (and encouragement).
 
The Samyang (Rokinon) 8mm fisheye arrived today, and the panohead calibration seemed to go smoothly. On a Canon Rebel T3i (1.6 crop factor), I've got it set up to 72 degree FOV, so a 360 degree pano is five images. With nominal effort, the first 360 degree pano has avg/max CP errors of 0.52 / 0.96.
Reply all
Reply to author
Forward
0 new messages