I've been using the R1 for a couple days now on my D800 and shaved 10.5. I'm really happy with the results for the most part. I'm getting great spherical panoramas with no stitching issues, but I need to get rid of the tripod in the shot. What's the best way to do this with my setup? I'm currently using PTGUI.
When the ground beneath the tripod is relatively plain, you can often simply clone out the tripod using Photoshop. It's best to extract a rectilinear view of the nadir area from the stitched equirectangular panorama to facilitate this. The easiest way is to generate 6 cube faces using a program such as Pano2VR or Pano2QTVR, edit the nadir face, and reassemble the cube faces back into an equirectangular image. Or you can use the Photoshop plugin by Rune Spaans: http://www.superrune.com/technical/s...eretocubic.php .
To replicate the true surface at the nadir, you need to take a shot of the nadir area with the tripod shifted out of the way and the camera at or close to its original position. This can be done in a variety of ways. Outdoors, you can simply hand hold the camera pointing down with outstretched arms (on or off the tripod). Indoors where the poor light levels dictate relatively long exposures, you need to use the tripod. If you release the tilt knob of the R1, you can tilt the camera down to its maximum extent. Then lean the tripod forwards with your foot on the bottom of the (lengthened) rear leg to prevent it toppling over, as illustrated in this article: http://www.rosaurophotography.com/html/technical6.html . An easier solution is to use the Nodal Ninja nadir/zenith adapter: http://shop.nodalninja.com/ultimate-...adapter-f6112/ .
The nadir image should be stitched into the panorama using viewpoint correction (PTGUI Pro only). This corrects perspective differences in the nadir shot that arise from not positioning the camera precisely with the original no-parallax point. While viewpoint correction will work accurately even for quite large displacements of the camera, it's best to get the camera reasonably close to the original position whenever possible. See these tutorials: http://www.ptgui.com/examples/vptutorial.html http://www.johnhpanos.com/ptgvpt.htm
Thanks for your reply. I'm still having issues with this though. My panoramics are coming out great, except for the ugly tripod in it! Do you think it could be a problem of the floor not having distinguishable control points. Some of the shots I've had problems with were laminate floors that are very plain. Do you think the Nadir adapter would help?
It's not quite clear what the problem is that you are having. The nadir adapter might help, but you still need control points to align the nadir image - assigned manually, if necessary. If the floor appears lacking suitable features for control points, you can sometimes find small spots (dust, scratches etc.) that are visible when zoomed right in. Otherwise, you can scatter some temporary features around the nadir - e.g. five or six small coins or adhesive spots. These can be cloned away from the final stitched panorama image. If you cannot get the nadir properly stitched, upload a set of (jpeg) images somewhere, and I and no doubt others will take a look. You can use a free file sharing site such as http://www.ge.tt/ if you don't have your own private web space.
Thanks John. I'm not sure if the problem is me being unfamiliar with PTGUI Pro, or if I did something wrong when taking the pictures. Here's a link to one of the panoramics I did that I'm having stitching issues with. Again, thank you for your help.
So what you have to do is to reset the old NPP to the whole system before shooting the Nadir shot. So you have to raise your center column and shift it sideways to take your nadir shot.
Would be easier to take a handheld shot or move aside with the tripod, use R1 built in Nadir Shot Position. So place a coin below your tripod, move the R1 forward till it touches the the flat surface. Now move the tripod backwards till you can focus to the coin and shoot your Nadir Image. For this method you have to use the Full Viewpoint Correction described in PTGui.
In case you are not using a pole might be a good idea to have a look to the Compartment "Custom Heads" to see the R1-S. Of course it is a little bit bigger, smaller than NN3, but has the opportunity to shoot Zenith and Nadir1 and Nadir2 Shots in NPP with Nadir Adapter.
Mark, I am following the steps detailed in my tutorials, viz: align the horizontal row of images first, making good any deficiencies in the control points assignments to get a good spread. Be sure to include the lens shift parameters in the optimization. Then roughly align the nadir using a few manually placed control points, and then check the viewpoint option on the nadir before repeating the optimization. The optimizer report should then be as good as it was before the nadir was added. Mask away those parts of the nadir image that are not part of the flat floor.
Thanks again for your help John. I'm tying a new one today and having problems again. When I follow your steps everything looked great except a little piece of tripod was showing in from one my non-nadir shots. I masked that away but everything went haywire after that. The images were all rotated and optimizer said it was a bad fit.
I'm sorry to keep bugging you with this but I feel like I'm missing something. The nadir adapter is on it's way from Nodal ninja so hopefully that helps.
Might be something with your calibration. The first pic is made of 4 shots just by deleting bad control points. You see a little misalignment and might be you should move a little bit to the right. The second pic is made with John's stitch by deleting the masks and stitching to see the output.
When I follow your steps everything looked great except a little piece of tripod was showing in from one my non-nadir shots. I masked that away but everything went haywire after that.
When "everything looked great", I assume you had aligned the images with the optimizer but there was a bit of tripod visible in the output image. If you then just mask the relevant image to remove the tripod bit, this will not affect the alignment of the images in any way. If the viewpoint nadir has at this stage been included and aligned, all you need do is generate the output. If you are at the stage where you need to add the nadir and align it to the already aligned images, you can uncheck the y,p,r boxes of all the images except the nadir before running the optimizer. The images cannot then go haywire as they are locked in position. If at any stage things do go haywire, just run Undo (ctrl+Z) to restore the project to the previous sane state and try to identify what it was that you did to make things go awry before having another go with the optimizer. (Never run the Align Images function twice, BTW).
I'm back! After a few months of not shooting 360's I'm back at it and having issues again. I really appreciate your help! I can stitch my 4 images into a nice 360, but when I include the nadir shot the optimizer says it's a very bad match. Any suggestions? Thank you!
All the suggestions one would like to make are already in the earlier posts of this thread. For the nadir shot, you should only have control points on the flat surface of the floor/ground. Without viewpoint correction selected, you will expect to get a "bad" optimization. Accept this result and then select VP for the nadir image only. A further optimization should now give a good result. If this doesn't work, upload jpeg copies of the images and your saved project file to your personal web space or a free file sharing site such as www.ge.tt , and post a link to them here.