R10 accessories - 4 shots, 6 shots, NPP setting
Results 1 to 15 of 15

Thread: R10 accessories - 4 shots, 6 shots, NPP setting

Share/Bookmark
  1. R10 accessories - 4 shots, 6 shots, NPP setting

    #1

    I've been looking at an R10 as a travel head, the rest of my pano gear is too heavy.

    Shooting a Sigma 8mm f3.5 on a Canon APS-C sensor, I've been happy with 6 shots around at +7.5-deg pitch on an NN5, plus nadir.

    I'd like to replicate this on the R10 with mini rotator. No problem, I see the 60-deg detent ring version.

    ***here is the question***
    However, the R10 quick reference guide refers to a .85 rail setting "for four shots around".
    Likewise, the R10 stop plate 102 product listing refers to "+7.4 4-shots GTP".

    Is this lens ring & stop plate going to be equally fine for 6 shots around? My brain tells me "yes", if the NPP is set it doesn't change with 4 or 6 shots. But then there are some comments from DennisS about a slight sideways offset with this head; would 6 shots tend to make this offset more prominent, making stitching more challenging?

    Thanks,
    Seth
  2. #2
    Users Country Flag
    Join Date: Apr 2008
    Location: Hong Kong
    Posts: 2,452

    6 shots will be slightly different from 4 shots, about -0.15 for the LRP settings. If you just need a stop plate, we can make one easily.

    nick



    Fanotec
    We listen. We try harder.
  3. #3

    Thanks Nick.

    For now I've ordered the F2106 Arca-Swiss style stop plates, I think those will serve me well as I'm getting used to the R10. Maybe I'll do the inserts in the future... but those stop plates look like good stuff too.

    I'm still a little bit in the dark - why is the 6-shot NPP different than the 4 shot? Does this have to do with the inconsistency of NPP at the edges of this fisheye, as demonstrated with laser pointers at some web site? The 6 shot moves the stiching seams to a different area of the image?
    Last edited by sbloombaum; 08-07-2014 at 12:28 AM.
  4. #4
    Users Country Flag
    Join Date: Apr 2008
    Location: Hong Kong
    Posts: 2,452

    Quote Originally Posted by sbloombaum View Post
    Thanks Nick.

    For now I've ordered the F2106 Arca-Swiss style stop plates, I think those will serve me well as I'm getting used to the R10. Maybe I'll do the inserts in the future... but those stop plates look like good stuff too.

    I'm still a little bit in the dark - why is the 6-shot NPP different than the 4 shot? Does this have to do with the inconsistency of NPP at the edges of this fisheye, as demonstrated with laser pointers at some web site? The 6 shot moves the stiching seams to a different area of the image?
    yes, you already got some idea.
    further reading
    http://michel.thoby.free.fr/Fisheye_...the-pupil.html



    Fanotec
    We listen. We try harder.
  5. #5
    Users Country Flag
    Join Date: Sep 2009
    Location: Los Angeles, CA
    Posts: 355

    Quote Originally Posted by sbloombaum View Post
    I'm still a little bit in the dark - why is the 6-shot NPP different than the 4 shot? Does this have to do with the inconsistency of NPP at the edges of this fisheye
    Seth,
    .
    You are correct that the NPP is different based on the number of shots around you want to use. It has to do with where the seam is in the picture. If you add more pictures, the seam moves towards the center of the image. This can effect where the NPP is and an adjustment may be necessary.
    .
    As to my previous post showing a modified R10 allowing side to side adjustment.
    .
    I approach panoramas as an engineer would. I do not have any stitching issues because my gear is calibrated pixel perfect. I do not know what the lens parameter tab in PTGui is and have never needed to enter individual parameters. Matter of fact, all I do is set the crop circle, auto generate points, add or remove control points as needed then stitch. Sometimes a pano needs more work in pre and post, but for the most part stitching is not an issue.
    .
    Since I take such an engineering approach to panoramas I want to get my rig as absolute spot on as possible.
    .
    The R1/R10 line of pano heads does not always position the center of the lens exactly above the center of rotation. I measured several rings and each was off center one way or the other.
    .
    This does not mean you will not get a good stitch. Modern stitching software does an excellent job of correcting an off calibrated rig.
    .
    I am very fortunate enough to have the tools and skills to fabricate whatever I need. When I was calibrating my monopod rig I noticed all the images were off center. I put the rotator on a tripod and verified the camera needed to shift sideways. I figured out how to modify the R10 to allow this. Hopefully if enough people ask Nick will implement side to side adjustment.
    .
    Side to side adjustment for the R1/R10 is not absolutely necessary. It would, however, help to dial in those last few millimeters to eliminate one more variable in rig calibration.
    .
    Dennis
  6. #6

    Hello Seth,

    This may help explain graphically what Dennis is saying.
    If you go to hugha.co.uk and scroll down to the button with an image of a Sigma 8mm lens (Determining the Nodal Point of a Lens), or use this URL - http://hugha.co.uk/NodalPoint/Index.htm and go near the bottom you will see diagrams of how the NPP moves relative to the angle of the light entering the lens.

    As Dennis says, the position to chose for setting up the lens is where the "join" is between shots.
    4 shots round and you are looking where the rays meet at 90 degrees - i.e. 45 degrees either side of the principal ray
    6 shots round and you are looking where the rays meet at 60 degrees - i.e. 30 degrees either side of the principal ray

    Also, with more shots round you get more overlap so can use the "Mask" tool in PTGui more easily and it is worth noting the the "distortion" caused by the way the NPP moves increases as the angle from the principal ray increases.

    Hope this explains the difference.

    Another consideration is how far away your subject is.
    It is a matter of scale so if your panorama has detail close to the camera then positioning is more critical than if all the detail is some distance away.

    Hugh.
  7. #7
    Users Country Flag
    Join Date: Sep 2009
    Location: Los Angeles, CA
    Posts: 355

    Quote Originally Posted by Hugh View Post
    Also, with more shots round you get more overlap so can use the "Mask" tool in PTGui more easily
    Hugh,

    I do not understand this statement. You should be able to use the masking tool just as easily if your pano has 4 or 400 shots. The shot count does not enter into the functionality of masking. I usually use viewpoint correction along with masking to patch the Nadir shot, no matter how many shots around.

    Dennis
  8. #8

    Hello Dennis,

    If you have more overlap then there is more image area to use the "Mask" tool in.
    For example , if you take 4 shots round you only have a small overlap then you are limited to the area you can "Mask" out or in, but if you take 8 shots round with the same lens with much greater overlap then you have much more area you can "Mask" out or in

    I have taken panoramas in very busy places with people moving about quite rapidly close to the camera and having a big overlap allows me to mask out people who appear in two (or sometimes more) images.

    I will post examples if you like.
    Last edited by Hugh; 08-08-2014 at 04:43 PM.
  9. #9
    Users Country Flag
    Join Date: Sep 2009
    Location: Los Angeles, CA
    Posts: 355

    Hugh,
    .
    So if I understand you correctly, it is not that the mask tool is any easier to use based on the number of shots, it is when you have moving subjects you want to remove, shooting more shots provides more material to work with when masking.
    .
    I always shoot 4 shots around when using the Sigma 8mm lens on a cropped sensor or the Tokina 10-17 on a full frame sensor. I shoot multiple images at the same position, but only at 4 positions. This gives me enough coverage to eliminate movement.
    .
    http://dlsphoto.net/Tutorials/DeletePeople/Compare.html
    http://dlsphoto.net/Tutorials/Crowd_Demo/CrowdDemo.htm
    .
    It is more your shooting technique than the number of shots taken around that will allow masking movement out of a panoramas. With monopod shooting, 4 images are easier to stitch than 6 or 8. For more images taken using a tripod you might need to deal with shadows as the sun races across the sky.
    .
    Dennis
  10. #10

    Hello Dennis,

    Yes, it is not that the more shots round makes the "mask" tool easier to use, but that the greater overlap gives me more scope for using it to eliminate or include detail (people).

    This is the fourth attempt at attempting to explain a little more as three times I have had to renter my user name and password and lost all that I had written so I'll try again tomorrow.

    Hugh.
  11. #11

    Thanks guys - appreciate the deeper explanations!

    Nick, upon setting the R10 up with the N6-60deg ring I first set to what I thought was +1.5 on the LRP. My mistake, that was actually 5+1.5 or 6.5.

    Then I noticed you'd actually written "-0.15" in your post above. (I'd been remembering "1.5")

    That's going to be .85-.15=0.7, right? I did a quick test at .7, and it autostitched in PTGui with the biggest control point error at 2.0315, only two points were above 2; I think this is pretty good?

    Is 0.7 the LRP setting you were recommending to start with? (n=6, Sigma 8mm f3.5 on a Canon APS-C)

    I must say, I love the design and engineering on this head. The accessory A-S style stop plate works easily and perfectly. The mini-rotator is fine in use, if a little loose when moving the tripod around.

    I just am so happy that this head will fit in my belt-pack camera bag, that's going to be great on an upcoming trip, it means I'll actually have it with me instead of the NN5 languishing in a suitcase

    Well done!

    DennisS wrote:
    "I figured out how to modify the R10 to allow this. Hopefully if enough people ask Nick will implement side to side adjustment... Side to side adjustment for the R1/R10 is not absolutely necessary. It would, however, help to dial in those last few millimeters to eliminate one more variable in rig calibration."

    I too have an engineering orientation, I want to make the tech work right with the least effort. Part of this was cutting my pano teeth on large virtual tour style projects, and needing to be able to replicate everything precisely for hundreds of panos.

    Would some shimming in the clamp be sufficient for minor side-to-side corrections? I guess that only gets you movement in one direction, towards the clamp screw.

    Higher on my wish list would be a lock for the mini rotator. I haven't dropped the camera yet, and I don't think I will, but when that thing starts swinging around with gravity while I'm moving the tripod or setting up an on-head nadir shot it's kind of disconcerting...
  12. #12
    Users Country Flag
    Join Date: Apr 2008
    Location: Hong Kong
    Posts: 2,452

    That's going to be .85-.15=0.7, right? I did a quick test at .7, and it autostitched in PTGui with the biggest control point error at 2.0315, only two points were above 2; I think this is pretty good?
    That is about right. -0.15 value is based on tests on R1. I usually do a bracketing with +/- 0.5mm and see what is best. :-)

    See how lens is made and why sometimes there is off-axis misalignment of lens rings.
    http://www.lensrentals.com/blog/2014...took-one-apart
    Basically, it is only an assumption that the lens barrel is symmetrical about lens axis. Shimming the lens ring will help but is really not needed. After all there is no single NPP for fisheye lenses! And the stitchers do not necessarily blend at mid point of the seam.

    nick



    Fanotec
    We listen. We try harder.
  13. #13

    Hello Dennis,

    Thank you for the links to your example.
    Unfortunately my computer told me that there was "No plug-in available to display this content" for the first one, but the second one
    http://dlsphoto.net/Tutorials/Crowd_Demo/CrowdDemo.htm
    worked just fine

    I don't know what was going on last night, but three times I was asked to login again with User Name and Password and each time I lost all my text.
    Perhaps I was taking too long or writing too much!

    I guess that our individual approaches to panoramic photography are related to our individual backgrounds, but when I started to explore this exciting form of photography I could not find the information I was looking for on the Internet so had to work things out for myself, as I am sure many of us who make panoramas have done.

    I have noticed from your posts that you are an Engineer with the skills and tools to turn your ideas into reality and fear that my first attempt would fall below your standard, but it did work and was built to rotate my D70 and Sigma 8mm around the NPP so I measured its positions first.
    http://hugha.co.uk/NodalPoint/Index.htm
    When I changed the camera body the rig became obsolete so I decided it was time to invest in Nodal Ninja products, and have not looked back.

    One early decision was how many shots round to take and fell back on my experience with photogrammetry where overlap is everything!
    To extract data from stereo photographs with an analogue plotter we would ensure that the image planes for the two photographs were in the same plane otherwise the operator would get considerable eyestrain when measuring from them and a 60% overlap was considered the optimum.
    With the advent of digital imaging it became possible to use convergent photographs and a 100% overlap became possible, so I felt that for panoramas a 50% overlap horizontally between images would give me a good solution.
    When information started to become available on the Internet I found that I took rather more images round than the recommendations and tables being published, but I found that having a greater overlap had three advantages in my view:
    1) There was a larger area for Control Point generation between images giving a greater spread of the points
    2) When the "Mask" option was introduced into PTGui I had more area to play with
    3) For the horizon I was using the more central parts of the images where the "distortion" was less.

    I have attempted to explain this on:
    http://hugha.co.uk/NodalPoint/The-Nodal-Point.htm
    and also in my User Guide for using 360° panoramas to add true colours to HDS Point Clouds:
    http://hugha.co.uk/Acrobat%20PDF%20F...int-Clouds.pdf
    which was written because many users were not even photographers let alone panoramic photographers.
    I found that in supporting these Customers that once they took more shots round the calls relating to stitching problems soon ceased.
    Perhaps it is a personal idiosyncrasy, but I find that more shots round with a larger overlap works for me so if any one asks I recommend 8 shots round instead of 6 or 4 shots round instead of 3.

    There is of course the problem that it is possible to take too many shots round giving too much overlap so that Control Points are generated for quite wildly separated images which I found gives PTGui a hard time.

    I note from your example that you take several "sets" to have a good selection of material to work with and I do the same.
    As usual you have produced a great panorama, but unlike Yellowstone and Yosemite I have visited Pompeii, but that does not diminish the enjoyment of looking around it virtually.

    In creating panoramas for corporate entertainment facilities I am faced with lots of people all moving about very close to the camera and often very limited space to work in so find the extra overlap useful when masking out movement that has taken place between shots.
    I do try and shot all the shots as quickly as possible, but there are occasions when I have to wait during the round.

    Clearly there is no way you could exclude your wife and two daughter, but eliminating people from panoramas does raise and interesting question.
    In the UK we introduced panoramic photography to many of the Police Forces, especially the Collision Investigation Units, as a means of adding true colour to Scanned (HDS) Point Clouds at accident and crime scenes, but recently they have been using the panoramas in their own right and several convictions have resulted because of a clearer understanding of the site from the panoramas by Judge and Jury, but it begs the question that the camera can lie because someone could claim they were there when the panorama was taken, but is excluded from the final result!
    It is of course possible to look at the original shots, but they may not be available in court.

    I was interested in your comment about getting a kick out of seeing the expressions on people's faces as they wonder what the heck you are doing because I find that in general people take no interest apart from the occasional person who does know what I am doing or the one occasion where I was asked not to take photographs because I suspect the couple concerned did not want a record of their being seen together!

    Hugh.
  14. #14
    Users Country Flag
    Join Date: Sep 2009
    Location: Los Angeles, CA
    Posts: 355

    Hugh,
    .
    You will need the DevalVR plug in for http://dlsphoto.net/Tutorials/DeletePeople/Compare.html
    .
    I am actually not an Engineer. Currently I write software for a living but have a CNC and manual machinist background. When I am out using my equipment and I think "this knob would be better if it were over here", I go home and move the knob location.
    .
    I do not like the idea of unscrewing a knob in order to move the upper arm of the M1 pano head. The internal pins hang if you do not open the assembly up enough. I have never liked the knob concept going way back to my first NN3.
    .
    http://www.dlsphoto.net/M1PanoHeadMo...perRotator.jpg
    .
    This modification is a bolt on mod that allows the upper assembly to stay fully supported as you rotate.
    .
    I am working on an idea for the R1 series that provides for fore/aft adjustment in the base plate of the ring clamp. We all know that not all camera/lens combinations are 100% the same. Published numbers are always a good place to start and for the most part work quite well. Having fore/aft adjustment for the R1 is just a place to dial out another variable when shooting panos. Providing fore/aft and left/right adjustments to the R1/R10 pano head rounds out the product.
    .
    Dennis
    Last edited by DennisS; 08-10-2014 at 10:21 AM.
  15. #15

    Thanks Dennis,

    I have downloaded the DevalVR plugin and enjoyed the view from the front of the Palace of Versailles, but would question whether the lady in the white hat and coat, grey skirt and black handbag had a twin?
    I have experienced this many times in crowded places and been grateful that I went round three times and had lots of overlap, but I have had situation virtually impossible to resolve, such as at Lord Cricket Ground where a lady seemed to be in every shot in a different place, but I could not get down to only one of her because of the close proximity and movement of all the other people.
    Every attempt to mask (both red and green) simply moved the problem around.
    I think my worst scenario was at he Cheltenham Gold Cup where I had a lady in a vibrant pink coat approaching me on one shot, then walking away after I had to wait while a dust cart went by.
    There are some I never did get 100%.
    http://www.keithprowse.co.uk/virtual...ton/index.html

    Having a CNC and manual machinist background to modify your kit to suit your needs clearly has its advantages.

    All the best, Hugh.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •