PDA

View Full Version : RD10 canon 7d sigma 4.5mm NPP



softgad
12-09-2010, 10:13 PM
HI
Still a newbie, I have taken out the rd4 rotator from my nn180 and installed the NN Ultimate R10 with 2.5 static, clamp my sigma 4,5mm following the R1 pdf file set up. According to the pdf file the NPP for 2.5 degree with three shots is 3.7 for my canon. I do have parallax problem with this setting. I will have to experiment but anybody can share values with me?

Thank you

softgad
12-09-2010, 10:34 PM
OOps figure out I was +2.5 degree and not -2.5 degree, I have reset the setting to 0 degree and will post later if NPP value are working for me

nick fan
12-10-2010, 12:54 AM
HI
Still a newbie, I have taken out the rd4 rotator from my nn180 and installed the NN Ultimate R10 with 2.5 static, clamp my sigma 4,5mm following the R1 pdf file set up. According to the pdf file the NPP for 2.5 degree with three shots is 3.7 for my canon. I do have parallax problem with this setting. I will have to experiment but anybody can share values with me?

Thank you

R10 setting is different from R1. Don't mix them up.

nick

softgad
12-10-2010, 10:53 AM
Thanks Nick, but cannot find any NPP value then for the RD10, any lead?

John Houghton
12-10-2010, 11:36 AM
The Sigma 4.5mm fisheye hasn't proved to be very popular and you may have some difficulty in finding anyone else using one on an RD10. You should just learn how to set up your pano head yourself. It's straightforward to do since all you can do is slide the camera forwards and backwards for any given angle of tilt. Just nudge the camera forward/backward a mm at a time and check for parallax until the parallax shift is eliminated (or at least minimized). See this tutorial (http://www.johnhpanos.com/epcalib.htm).

John

softgad
12-10-2010, 11:26 PM
I spent a few hours trying just about every setting possible on 0 degree with three shoots. I have tried from the minimum to the maximum NPP value with each mm option. The best result is at 0.7 but still not perfect. I will like to share with you that I obtain perfect result when I shoot 4 pictures at 0.7 setting on my sigma 4.5mm with canon 7d RD10. For someone not looking for parallax problem at the upper level of the picture almost all setting works at 3 shoots! which mean intuitively that outside pictures with a pole will be quite forgiving

John Houghton
12-11-2010, 01:05 AM
If you are shooting on a tall pole, then a few mm from the ideal NPP position will make no visible difference to the stitching. I do wonder, however, if you are distinguishing between parallax errors and stitching errors in the upper level of the panorama image. Stitching errors are often due more to poor stitching than to parallax effects. What stitching software are you using? Adjusting your head to the NPP should take no more than 10 minutes or so.

I found some R10 settings with Google search at this page (http://www.panophoto.org/forums/viewtopic.php?f=230&t=5252)

John

softgad
12-11-2010, 11:38 AM
HI John merci, actually you may be right that the problem is more related to the software but I am using the current version of PtGui but I am using a tripod for this test. Also the NPP position according to Nick is for the R1 and not for the R10. So the result reflect how happy is the software (without any manipulation) with the position of the NPP. The misalignment is at the seam for all NPP position except at the setting I have given. So my data could be wrong if I learn better to correct them in the software. Do you guys do all manually the stiching?

John Houghton
12-11-2010, 12:03 PM
It's important to stitch the images manually and not rely on PTGui in automatic mode. Only that way can you be sure that you have control points assigned in the necessary places and are optimizing all the appropriate parameters. Automatic control points generation has a strong tendency to bunch control points in the middle of the seams and ignore the outer edges of fisheye images. To get the images to align at the top and bottom, you need control points in those areas, and if there are none there, you should assign some manually. You also need to optimize the horizontal and vertical shift parameters (on the Optimizer tab: Minimize lens distortions - select heavy + lens shift). If you can get the maximum distance down to less than 2 or 3 AND you have control points nicely spread along the whole length of the seams, then you will get excellent alignment.

softgad
12-12-2010, 04:17 PM
This is good info, then with the prescribed 3.4NPP with 0 degree (3 shoots), with additions of control points at the upper level and bottom level + advanced setting at your recommendation (Minimize lens distortions - select heavy + lens shift) for the optimizer, I have no parallax problem at all. The lens has its advantages, quick to produce and is best used when a lot of actions/people are present of for some small bathroom space, but to be frank when you are alone in a landscape or even city streets, the lens is far from being ideal. I still prefer in this context my 17-40mm lens with nn5 by far far away distance :)

John Houghton
12-13-2010, 12:12 AM
This is good info, then with the prescribed 3.4NPP with 0 degree (3 shoots), with additions of control points at the upper level and bottom level + advanced setting at your recommendation (Minimize lens distortions - select heavy + lens shift) for the optimizer, I have no parallax problem at all.)

You didn't have parallax problems before. What the above actions have eliminated is poor alignment, i.e. stitching errors.

John