Luminous Landscape Forum
Equipment & Techniques => Beginner's Questions => Topic started by: jannatul18 on July 24, 2014, 05:06:10 am
-
Hello can you please let me know for panoramic pictures, where is the best place to pivot the camera? Please help.
-
Hello can you please let me know for panoramic pictures, where is the best place to pivot the camera? Please help.
Hi,
The best point is the center of the Entrance pupil (http://www.johnhpanos.com/epcalib.htm) (AKA No-Parallax Point or NPP) of your lens.
Cheers,
Bart
-
Hi,
The best point is the center of the Entrance pupil (http://www.johnhpanos.com/epcalib.htm) (AKA No-Parallax Point or NPP) of your lens.
Bart's right in theory, but my experience is that it really makes no difference. On a recent trip I made dozens of panoramas ranging from 4 to 12 shots simply by hand-holding the camera a pivoting my neck. They stitched together perfectly. Just be sure you have a good deal of overlap from one shot to the next.
-
See http://www.luminous-landscape.com/forum/index.php?topic=82928.msg670134#msg670134
-
Bart's right in theory, but my experience is that it really makes no difference. On a recent trip I made dozens of panoramas ranging from 4 to 12 shots simply by hand-holding the camera a pivoting my neck. They stitched together perfectly. Just be sure you have a good deal of overlap from one shot to the next.
I'll second your comment. Most of my panoramas are hand-held with 70-200 mm IS tele with no problems whatsoever. 1/4 or more overlap is the key. If I know I'll be stationary for a long while, close to my car, I'll use a tripod and shutter release cable. That allows me to simply pivot the camera in small increments without even looking through the viewfinder to judge how much overlap I'm using. So sometimes I may be overlapping 70%. Not a problem with that either for PS stitching.
-
Programs have gotten better, and you don't need absolute precision to stitch panoramas without very close objects. You do need to use the no-parallax point if you plan on having a close object in your panorama.
-
Programs have gotten better, and you don't need absolute precision to stitch panoramas without very close objects. You do need to use the no-parallax point if you plan on having a close object in your panorama.
Even then, sometimes I just move a foreground rock or two around so the stitching line doesn't matter.
-
Bart's right in theory, but my experience is that it really makes no difference.... I made dozens of panoramas ranging from 4 to 12 shots simply by hand-holding the camera a pivoting my neck. They stitched together perfectly.
Much will depend on the subject matter; For panoramas that include little foreground, you can get away with a lot of imprecision, but if you start to include objects closer to camera the importance of pivoting around the lens nodal point increases.
-
Hi,
That works as long as you don't have objects in the near foreground.
Best regards
Erik
Bart's right in theory, but my experience is that it really makes no difference. On a recent trip I made dozens of panoramas ranging from 4 to 12 shots simply by hand-holding the camera a pivoting my neck. They stitched together perfectly. Just be sure you have a good deal of overlap from one shot to the next.
-
Hi,
That works as long as you don't have objects in the near foreground.
Indeed, the question was, "Where is the best place to pivot the camera?"
Of course there are scenarios (and ways!) to perform the shooting sequence that reduces the risk of failure, but the best one is as indicated ...
Cheers,
Bart
-
Hi,
That works as long as you don't have objects in the near foreground.
How close? I have hand-held panos with objects as close as 10 feet and they are fine.
-
How close? I have hand-held panos with objects as close as 10 feet and they are fine.
It all depends on the scene. One can align foreground parallax, or background parallax, not both at the same time. So if e.g. the background is not critical (no predictable structures) or blurred, or chaotic, one can align foreground features without trouble. However, parallax can turn out to be impossible to fix, or at least very time consuming, so it's still better to prevent than (having) to cure.
With a bit of practice (like turning oneself around the entrance pupil, instead of rotating from the neck or torso), and a good amount of overlap and good software, and a suitable subject (few occlusions), and some luck, it is possible to achieve good handheld results. But that was not the question the OP asked. Besides, good technique also requires fewer images (with less overlap) to achieve a successful stitch, which will save more time, especially when focus stacking or exposure bracketing enters the picture, so to speak.
Cheers,
Bart
-
Bart's right in theory, but my experience is that it really makes no difference.
My impression is that Bart may see a thousand unacceptable mismatches in that really makes no difference picture :-)
-
My impression is that Bart may see a thousand unacceptable mismatches in that really makes no difference picture :-)
I agree. All of Bart's suggestions may be perfectly true, but do they make a difference in the final photos? Not in my experience. And, isn't that our final goal? A photograph that we and others will look at and enjoy? God save us from the pixel peepers who haven't any interest in (or clue about) the aesthetic value of a photo, but just stick their nose up against the photo and look for flaws. This is not meant as a criticism of Bart, but I do think that a lot of people on LuLa are too obsessed with technique and not concerned enough about composition, lighting, and the like.
-
My impression is that Bart may see a thousand unacceptable mismatches in that really makes no difference picture :-)
Your impression then, is most likely based on your lack of practical experience. My practical experience with panorama stitching (montage, back to the film days) goes back some 2 decades. Time spent in post-processing impossible feature matches is largely wasted time, if better technique could have prevented that waste of time (with no guarantee of success, unlike with proper technique).
I do admit that my experience as a part-time ISO certification team leader in (what then was) the world's largest supplier of photographic materials may bias my opinion towards a 'prevent GIGO', i.e. garbage-in-garbage-out, mentality. It's because 'quality management (https://en.wikipedia.org/wiki/Quality_management)' actually works.
Avoiding sloppy work at the input stage, will pay off at the end.
Cheers,
Bart
-
I agree. All of Bart's suggestions may be perfectly true, but do they make a difference in the final photos?
Hi Peter,
As I said, it depends on the subject matter (and experience to compensate), and the amount of time you choose to waste on avoidable post-processing attempts. Feel free to waste as much time as you wish, but good advice is something different.
Cheers,
Bart
-
…but do they make a difference in the final photos?
They can and the practical issue is that post-processing fixes to make sure they don't make a difference in the final photos takes time. (Yesterday I repeatedly reworked images to remove mismatched edges in the overlaps, by painting all but one version transparent.)
Often I can bodge the images into a composite and I wouldn't wish to discourage anyone from making composites without a pano head; but if the money's available…
…too obsessed with technique and not concerned enough about composition, lighting, and the like.
I take them to be obsessed with technique and concerned with composition, lighting, and…
-
Avoiding sloppy work at the input stage, will pay off at the end.
As I agree, I think you grabbed the wrong end of the stick.
-
As I agree, I think you grabbed the wrong end of the stick.
What stick?
Cheers,
Bart
-
http://dictionary.cambridge.org/dictionary/british/get-the-wrong-end-of-the-stick
-
http://dictionary.cambridge.org/dictionary/british/get-the-wrong-end-of-the-stick
Yes, that's exactly what I asked, what end of the stick, not the meaning of the phrase/expression ...
In addition (https://en.wikipedia.org/wiki/W._Edwards_Deming), for your edification (since you seem to like referencing other's achievements) ...
-
Yes, that's exactly what I asked, what end of the stick, not the meaning of the phrase/expression ...
Actually, "exactly what [you] asked" was -- "What stick? (http://www.luminous-landscape.com/forum/index.php?topic=91947.msg750048#msg750048)"
(Read "Bart may see a thousand unacceptable mismatches..." as "Peter may be ignoring a thousand unacceptable mismatches...").
-
I shot a pano sequence inside a cathedral recently. Had a lens/camera combo that I had never bothered to check for nodal point, so I just rotated about the tripod screw on the body. Looks fine with objects at appreciable distance, but I cant get the nearby pews to stitch well at all. So in that case, not considering nodal point ruined a really, really nice pano shot.
Edit: This would have helped tremendously:
http://gregwired.com/pano/pano.htm
-
I shot a pano sequence inside a cathedral recently. Had a lens/camera combo that I had never bothered to check for nodal point, so I just rotated about the tripod screw on the body. Looks fine with objects at appreciable distance, but I cant get the nearby pews to stitch well at all. So in that case, not considering nodal point ruined a really, really nice pano shot.
Hi,
Depending on the exact scene layout, maybe something is salvageable, but it will probably take a significant manual editing job to pull it off.
As I said earlier, one can either minimize parallax errors in the foreground, or in the background, but not both at the same time. So you can do one stitching operation, while only selecting background features to place the control points on. Then do another stitching operation based on only control points that are around the distance where the initial stitch started showing issues. Then another for a bit shorter distance, and so on. Then do a new stitching operation on all these intermediates, making sure that each image pair has control points at the transition zone between two optimized distances, while probably zeroing the distortion parameters since the distortions were already removed. That would basically only translate and magnify the various layers into optimal position.
In all these operations, also try a good blending program, such as SmartBlend, because it may find a better transition blend between the layers.
This will require a very good Panostitching application (e.g. PTGUI, or Hugin) that allows lots of manual control. Photoshop is probably not able to do what's needed, unless in addition you apply manual masks to each tile that needs to be aligned, prior to the varoius stitching operation runs. That would add even more manual effort.
Even then, there is no guarantee for success. But if time and effort are not an issue, it may be worth an attempt. This obviously is not something one wants to do when time is of the essence, like having to meet a deadline for a client (who is also not likely willing to pay for your wasted time).
Edit: This would have helped tremendously:
http://gregwired.com/pano/pano.htm
Yes, or something sturdier for heavier cameras, to reduce vibration blur. Usually, the optimal No-Parallax settings will result in a center of mass of camera/lens combination that is significantly offset from the center of the tripod, so vibrations may become magnified rather than damped. Also, not using the multi-row setup when a single row setup is sufficient, may help to get lower vibration risk.
Cheers,
Bart
-
…also try a good blending program, such as SmartBlend…
Oooo! That looks interesting!
-
Oooo! That looks interesting!
It is!
SmartBlend (http://wiki.panotools.org/SmartBlend) is really smart in performing miraculously effective blends, but the impossible still remains that, impossible. It is also no longer being developed it seems, so there's no 64-bit version available, and it therefore depends on also very clever front ends to feed it the optimized (for distortion and light fall-off) projection solution piece by piece.
There are others as well, like Enblend, but they all have (besides their strengths) some sort of drawbacks to offer as well. SmartBlend is an overall good performer, and also has some parameters that can help it to do even better for certain scenarios. So if a Stitching program is capable enough to allow to adjust the command line it sends to SmartBlend, most things will come out relatively well.
Cheers,
Bart
-
So if a Stitching program is capable enough to allow to adjust the command line it sends to SmartBlend…
Otherwise I suppose it might be possible (Hugin) to dump intermediate files and manually adjust the command line…?
-
Otherwise I suppose it might be possible (Hugin) to dump intermediate files and manually adjust the command line…?
Manual is always possible, but I seem to remember that Hugin can directly utilize SmartBlend, but maybe that has changed (I haven't used Hugin for a while). I just checked, and SmartBlend is mentioned in the help file, but I think one will have to Google for instructions on how to integrate (maybe change the preferences to the optional blender, and adjust parameters?).
Cheers,
Bart
-
Hi,
The best point is the center of the Entrance pupil (http://www.johnhpanos.com/epcalib.htm) (AKA No-Parallax Point or NPP) of your lens.
Cheers,
Bart
This being the 'Beginner's Questions' forum, Bart's answer is the best one. Yes, you can get around parallax issues in post, but the 'best' way is to avoid them in the first place. Since that varies from lens to lens, you'd really need to invest in a nodal head for your tripod. Several companies make them, at different price points, and then you'll have to experiment to find the right position for a given camera/lens combination.
Mike.