Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: PaulSchneider on April 05, 2010, 04:03:44 pm
-
Hi guys,
I just have a little theoretical question about stitching with a view camera ... Can one easily combine normal stitching by shifting and nodal point stitching with a nodal point adapter in order to get really gigantic panoramas? Or are there any caveats?
One main advantage of using view cameras is to my understanding the capability of easy, seamless stitching by way of shifting the digital back. Coming from 35mm though, I read a lot about panoramic heads - such as the one from manfrotto - as another viable way of creating huge panoramas.
My thought now is to combine these two methods. I.e. by combining multiple 2 x 2 shifted images that were sequentially created side by side with the help of the panoramic head.
Is there anywhere a list where I can see the nodal points of the Rodenstock/Schneider lenses?
Kind regards
Paul
-
My understanding is that there is far greater lens distortion from nodal point panos. So if you are looking for ultimate image quality, I wouldn't use that method.
-
Hi guys,
I just have a little theoretical question about stitching with a view camera ... Can one easily combine normal stitching by shifting and nodal point stitching with a nodal point adapter in order to get really gigantic panoramas? Or are there any caveats?
One main advantage of using view cameras is to my understanding the capability of easy, seamless stitching by way of shifting the digital back.
Paul
What exactly are you contemplating doing?
You can easily stitch pairs of exposures vertically using rear rise/fall, and then pan and cylinder stitch the pairs...
What camera are you using?
I had thought that, with a Sinar, I could use the rail to put the nodal point above the tripod pivot, but the standard would probably be in the wrong place, but then I have a base plate for use with two rail clamps.
Most pano heads, I would have thought, would not take the weight of a view camera.
-
FWIW I used Blad / Leaf combo for shooting large panos on an RRS pano head and the results were pretty good. Now, shooting 5dmk2 and using the same head, the results are in my opinion stunning. I can shoot higher ISO and get less mirror movement. I shoot on average 7gig landscapes with 400mm lenses and at the other end stitch interiors with a 24mm. I rarely encounter stitching errors and am presented with huge, clean images. I know this doesn't really answer your large format question but I couldn't imagine creating these images with my P2 which unfortunately hasn't been out of the case in a long time... You can see downsampled versions here (http://www.craig-stephen.co.uk)
-
I couldn't imagine creating these images with my P2 which unfortunately hasn't been out of the case in a long time...
What could be unimaginable about pano-stitching with a P2?
-
What could be unimaginable about pano-stitching with a P2?
When I said "I couldn't imagine creating these images with my P2" I was referring to the panos I linked the OP to - if you care to take a look I would imagine you would come to the same conclusion.
-
Hi guys,
I just have a little theoretical question about stitching with a view camera ... Can one easily combine normal stitching by shifting and nodal point stitching with a nodal point adapter in order to get really gigantic panoramas? Or are there any caveats?
One main advantage of using view cameras is to my understanding the capability of easy, seamless stitching by way of shifting the digital back. Coming from 35mm though, I read a lot about panoramic heads - such as the one from manfrotto - as another viable way of creating huge panoramas.
My thought now is to combine these two methods. I.e. by combining multiple 2 x 2 shifted images that were sequentially created side by side with the help of the panoramic head.
Is there anywhere a list where I can see the nodal points of the Rodenstock/Schneider lenses?
Kind regards
Paul
Paul,
Having explored stitching with MFDB on a viewcamera with sliding back, the fact is that it is much work. A viewcamera also have tilt and swing. Combining this with the tolerances you need to maintain focus is a tricky thing in my book. Simpler seem to be to use a technical camera with only shift or shift and rise/fall --- and with sensor plane and lens plane perfect adjusted. This is something I was told by Schneider, and I believe they are right. For digital thus it seems the viewcamera belongs in studio for product shots and thethered shooting, for ultimate lens movement shots.
Picture this; the view camera image plane is what size? The sensor is what size? Now, the tolerances you adjust a view camera film image should be replicated in the small reduced scale of the sensor.
Granted, doing what you ask on a technical camera would be a whole lot simpler. If same time lens movements are required, the Sinar arTec might be the ticket. Or, simply use a tilt/shift lens on a Canon or Nikon.
Regards
Anders
-
When I said "I couldn't imagine creating these images with my P2" I was referring to the panos I linked the OP to - if you care to take a look I would imagine you would come to the same conclusion.
I have had a look, and I have not come to the same conclusion.
I can appreciate that with a cheap view cameras, aligning the standards accurately might be difficult, but not with a P2... and you can check the focus with live view, what is unimaginable? (we are in the Digital backs and Large sensor photography forum).
-
Hi guys,
I just have a little theoretical question about stitching with a view camera ... Can one easily combine normal stitching by shifting and nodal point stitching with a nodal point adapter in order to get really gigantic panoramas? Or are there any caveats?
Something tells me the limitation is geometric not the method used or combined methods used...
You should join and post this question on the APP forum or the PTGui forum.
My thought now is to combine these two methods. I.e. by combining multiple 2 x 2 shifted images that were sequentially created side by side with the help of the panoramic head.
Is there anywhere a list where I can see the nodal points of the Rodenstock/Schneider lenses?
Do you mean rotating the nodal point head and then stitching by shift after rotating? In theory it should work or be no different to just rotating further with the head. Where you will find limitations will be in the 'type' of stitch that the maths behind the stitch (related to the field of view (FOV)) allows you to do.
Smaller distances (FOVs) covered will permit planar or perspective stitching, larger FOV's will require cylindrical or spherical stitching. Lens distortion is another matter all together and will be introduced and multiplied the more frames that are stitched. It can also be controlled with various bits of software - some of which can be used as plug-ins in the more advanced stitching programs...
Murray
-
I'll throw in here, and say that I wouldn't be inclined to create stitched panos with my 679cs, but it's a joy to do on the Silvestri Bicam in direct mode (ie no bellows, and the lens mounted directly to the camera).
The 679 doesn't have enough direct shift, and I suspect that this may be similar with other view cameras, pending the image circle of the lens and the sensor size (ie you may not need a lot of movement if the image circle only just covers the sensor in the first place); the Bicam goes 17mm in every direction on the rear standard (when using the sliding backs), so also useful would be stuff like Alpas, which probably allow even more movement. (My technique is 3 across at rise 17 and 3 across at fall 17 (or just 3 across with no rise / fall)).
My small experience with doing nodal point panos with the Manfrotto a (long) while back was not anywhere as good (but there are probably others who have struggled and perfected that technique)
Hope this helps
-
Do you mean rotating the nodal point head and then stitching by shift after rotating? In theory it should work or be no different to just rotating further with the head. Where you will find limitations will be in the 'type' of stitch that the maths behind the stitch (related to the field of view (FOV)) allows you to do.
Murray
Stitching software can deal with horizontal shifting or horizontal panning, but not both in the same picture/area, but you can use rise and fall (vertically) to create pairs or sets of images for stitching vertically, and pan (horizontally) to create other sets for cylinder stitching.
For Port Isaac and similar harbours I intent to use cylinder stitching for the "amphitheater" inland end of the harbour, and shift-and-stitch and move-and-stitch for the estuary, so that the harbour would then end up looking like an Island, but this is an order of magnitude (or two) more complicated.
-
I have had a look, and I have not come to the same conclusion.
I can appreciate that with a cheap view cameras, aligning the standards accurately might be difficult, but not with a P2... and you can check the focus with live view, what is unimaginable? (we are in the Digital backs and Large sensor photography forum).
Really? I'm shooting either 300 or 400mm and averaging 400 shots per stitch. How would that work out on LF? Do you have examples you can show me, I'd be very interested in the comparison. Sorry I dont follow "we are in the Digital backs and Large sensor photography forum"
-
Really? I'm shooting either 300 or 400mm and averaging 400 shots per stitch. How would that work out on LF? Do you have examples you can show me, I'd be very interested in the comparison. Sorry I dont follow "we are in the Digital backs and Large sensor photography forum"
I have not used the technique yet, but, using a MFB with double the resolution, it should be easy to get similar or better results with half as many shots...
For the same field of view, use the same focal length, regardless of format... the lens I would like to get for stitching is the Schneider Fine art Gold 1100, but I have a selection of lenses up to 900mm, including a Novoflex/Leica 400mm which I think would be fully up to making the most of 6 micron pixels...
So many people think that panoramic and landscape photography is all about wide-angles, so it is nice to hear than some one else appreciates long focus lenses.
-
Jon, Dick, anyone else exprienced with the "stitching with view cameras" topic. Does anyone have images of moved stitches they would care to share?
I am about to try it (e.g. 3 across shift with front tilt and rear fall). In a 2 across, I want to try opposite rear swings on either side of a close perspective symmetrical subject. I can't find material in the usual sources to provide starting points. I have the feeling I will spend a month fooling with the view camera only to find myself boggled.
-
Jon, Dick, anyone else exprienced with the "stitching with view cameras" topic. Does anyone have images of moved stitches they would care to share?
I am about to try it (e.g. 3 across shift with front tilt and rear fall). In a 2 across, I want to try opposite rear swings on either side of a close perspective symmetrical subject. I can't find material in the usual sources to provide starting points. I have the feeling I will spend a month fooling with the view camera only to find myself boggled.
I think you are talking about camera movements, rather than moving the camera?
Someone did post a picture involving moving the camera - the photo was of the building along a river or estuary in Scandinavia somewhere.
...are you contemplating rectilinear stitching? ... you are looking at a rectangular building, and you want it to look rectangular on the print?
...or are you wanting perspective in the print of a rectangular object?
... either way you would use front tilt up (for a building from ground level) and rear fall
use front yaw (sideways tilt) for control over the position of the plane of sharpest focus, and rear yaw for perspective in the horizontal.
Different people and books use different terms for the different movements of a view camera, and it can be very confusing... but , if you do not already have it, I suggest you download Harold Merklinger's "Focusing the view camera" (http://www.trenholm.org/hmmerk/)
¿clear as mud?
If you shift-and-stitch, you move the back in a plane (as the focused image stays in the same plane), and, apart from that it is just the same as using the view camera for single shot.
you set the front rise/fall in one setting and leave it there,
If you want verticals parallel, keep the sensor plane vertical.
I would love to set this up with you, but I expect you are on a different continent?
-
My understanding is that there is far greater lens distortion from nodal point panos. So if you are looking for ultimate image quality, I wouldn't use that method.
Jonathan,
Your understanding is not correct.
There are different ways to project a facetted cylinder on a plane, and when using planar projection you get the exact same result you would have gotten with a wide lens, with the same limitations in terms of area that can be covered.
The difference being that you can use a longer lens to reach a much higher resolution.
Regards,
Bernard
-
Paul,
Having explored stitching with MFDB on a viewcamera with sliding back, the fact is that it is much work. A viewcamera also have tilt and swing. Combining this with the tolerances you need to maintain focus is a tricky thing in my book. Simpler seem to be to use a technical camera with only shift or shift and rise/fall --- and with sensor plane and lens plane perfect adjusted. This is something I was told by Schneider, and I believe they are right. For digital thus it seems the viewcamera belongs in studio for product shots and thethered shooting, for ultimate lens movement shots.
Anders,
Thanks for sharing your results on this. I am almost tempted to write "I told you".
One more step and you will recognize the superiority of spherical stitching.
Cheers,
Bernard
-
Hi guys,
I just have a little theoretical question about stitching with a view camera ... Can one easily combine normal stitching by shifting and nodal point stitching with a nodal point adapter in order to get really gigantic panoramas? Or are there any caveats?
Paul,
I would stick to spherical stitching alone, it will cover all your needs really easily and fast. These software make some assumptions about the sensor location relative to the rotation point, and combining shifting with rotation will confuse them.
These is also no value in my book.
I have been very happy about RRS pano kit, but there are many alternatives.
Now the next question is what sensor to use to do such stitching. As I have written in the past, I am firmly convinced that high end DSLR are a much better platform to do these things. The only problem is that not all of them deliver the DR you can find in MFDB, but some do get extremely close if not ahead.
Cheers,
Bernard
-
Anders,
Thanks for sharing your results on this. I am almost tempted to write "I told you".
One more step and you will recognize the superiority of spherical stitching.
Cheers,
Bernard
Bernard,
Well... but I did not state that I prefer stitching on nodal point.
For ultimate image quality, thus also correct perspective one need to do flat stitching. This applies primarily to wide lenses. Yes, it is correct that CS4 and other programs are very good to recalculate perspective to "near equal" flat, and no they are physically unable to do it perfectly. With long lenses one will see little difference, because of the distance used to the subject versus the movement of sensor around nodal point. For wide field of view, whether of architecture or landscapes, flat stitching is required for the ultimate perspective.
What appear tempting would be a technical camera that could flat stitch from 2-3 shots, and of similar proportion and lens perspective to classical 617 with 90mm lens or 612 with 58mm lens. That would equal focal length of 57mm and 42mm respectively using a 48x36mm sensor, and 39mm and 28mm lens on full frame DSLR (when relating to the short side of both formats/sensors).
Now someone in above mentioned using 300-400mm lens. That is complete different obvious. Stitching around a nodal point seem simplest!
For wide it is different. This is why top end architectural photographers buy the likes of Sinar arTec. For architecture they go wider than the classical focals I mentioned. Are 90mm on 617 and 65 on 612 classical? Well... at least those are ones that appeal to my eye . Some like 72mm on 617 etc . For architecture a Schneider Digitar 35mm XL and 48x36mm sensor would yield the equal view of a 54mm lens on a 617. That is pretty wide! If 3:1 proportion that would equal a 12mm lens on width of the 617 format. And... made with just 2-3 shots.
MFDB? The sensors at low ISO deliver higher image quality.
To OP, why combine stitching around a nodal point and flat stitching? If already accept the "near flat" calculation of CS4 or similar, then why not only do stitching around nodal point? If on other hand the point is to stitch very wide perspective, as I mentioned above a technical camera may tbe simplest and preferred for image quality.
A side note: The important is not only the image, but how one see it. A groundglass on a technical camera is one way. Or a panoramic viewfinder. The last can also be used on a DSLR. Seeing many stitched panoramas, frankly most do not have very good composition, because the photographer attempts to see the view in their mind. A viewfinder can help. Or... simply a zoom lens set on wide first and interpreting the scenes as split in half around horizontal, and thereafter zooming in to twice the focal for image capture, frame by frame around nodal. This is last technique was used by Ben Rubenstein on Getdpi. His stitched around nodal panos are very well composed.
Regards
Anders
-
Bernard,
Well... but I did not state that I prefer stitching on nodal point.
For ultimate image quality, thus also correct perspective one need to do flat stitching. This applies primarily to wide lenses. Yes, it is correct that CS4 and other programs are very good to recalculate perspective to "near equal" flat, and no they are physically unable to do it perfectly. With long lenses one will see little difference, because of the distance used to the subject versus the movement of sensor around nodal point. For wide field of view, whether of architecture or landscapes, flat stitching is required for the ultimate perspective.
Anders,
I have published in the past several examples of flat stitch (9 images shot with a 100mm lens) vs wide shot (one image shot with a 24mm), and the only reason why they are not totally identical is that the 24mm lens has 2.5 points of barrel distorsion. The stitch is more faithful to the actual scene.
I would definitely not use CS4 as a reference for these applications, PTgui and Autopano pro are much better.
Could you please point at evidences supporting your statement above? They are not backed up by either experience or mathematical theory.
Thank you.
Cheers,
Bernard
-
For ultimate image quality, thus also correct perspective one need to do flat stitching.
Anders,
Could you please point at evidences supporting your statement above? They are not backed up by either experience or mathematical theory.
Thank you.
Cheers,
Bernard
Bernard...
Much data is lost in rectilinear stitching, as the distorted images have to be cropped back to rectangles, but you can compensate for this by using more images.
(In pano-cylinder-stitching using long lenses, there may be little or no distortion or loss of data.)
No DSLR or other camera with an anti-aliasing filter approaches the "per pixel" quality of any medium format digital back, let alone the best modern ones, but, you could use your technique with a top quality camera.
I know your technique works, and you have photos to prove it, but for ultimate image quality, and correct perspective, one needs to flat stitch with a MFDB and a quality lens.
This is because the stitching software has to distort and re-sample the image, and this can only degrade the image, causing loss of resolution.
I am aware that the better stitching programs do a good job, and the loss of resolution is minimal, and would be difficult to see or quantify in a high-res image.
I think that even the best £5,000 ultra-wide angles are not distortion-free, and if you have to correct distortion in post-production, then you are re-sampling the image, and this would result in the same type of loss of resolution that you get with pano software, so, in extreme circumstances, a rectilinear pano can produce results comparable with what can be achieved with a less than ideal ultra-wide angle lens. (and I have to admit that i do not think they ever have produced an ideal ultra-wide, and never will.)
One problem is that there are AFAIK, there are no digital long lenses for Medium format digital view cameras, but the Schneider Fine Art Gold lenses might be useful as they have enough image circle for many Giga-pixels, so you could overcome any lack of lp/mm by down rezzing.
If you can produce a pano from 9 pictures taken with a 100mm lens that is equivalent to a single shot taken with the same camera with a 24mm lens, then the only benefit is that you could do without the 24mm lens, for subjects that do not move, in circumstances where you have the time to set up your kit and take 9 pictures!
I appreciate that cylinder / spherical stitching is different, and would, for me, be the technique of choice for panoramas of curved subjects.
-
For ultimate image quality, thus also correct perspective one need to do flat stitching.
Bernard...
Much data is lost in rectilinear stitching, as the distorted images have to be cropped back to rectangles, but you can compensate for this by using more images.
(In pano-cylinder-stitching using long lenses, there may be little or no distortion or loss of data.)
No DSLR or other camera with an anti-aliasing filter approaches the "per pixel" quality of any medium format digital back, let alone the best modern ones, but, you could use your technique with a top quality camera.
I know your technique works, and you have photos to prove it, but for ultimate image quality, and correct perspective, one needs to flat stitch with a MFDB and a quality lens.
This is because the stitching software has to distort and re-sample the image, and this can only degrade the image, causing loss of resolution.
I am aware that the better stitching programs do a good job, and the loss of resolution is minimal, and would be difficult to see or quantify in a high-res image.
I think that even the best £5,000 ultra-wide angles are not distortion-free, and if you have to correct distortion in post-production, then you are re-sampling the image, and this would result in the same type of loss of resolution that you get with pano software, so, in extreme circumstances, a rectilinear pano can produce results comparable with what can be achieved with a less than ideal ultra-wide angle lens. (and I have to admit that i do not think they ever have produced an ideal ultra-wide, and never will.)
One problem is that there are AFAIK, there are no digital long lenses for Medium format digital view cameras, but the Schneider Fine Art Gold lenses might be useful as they have enough image circle for many Giga-pixels, so you could overcome any lack of lp/mm by down rezzing.
If you can produce a pano from 9 pictures taken with a 100mm lens that is equivalent to a single shot taken with the same camera with a 24mm lens, then the only benefit is that you could do without the 24mm lens, for subjects that do not move, in circumstances where you have the time to set up your kit and take 9 pictures!
I appreciate that cylinder / spherical stitching is different, and would, for me, be the technique of choice for panoramas of curved subjects.
Dick,
Thank you.
Bernard,
Puh.
Allow me to be frank and without the meaning of being impolite; I have to admit that it is a tad tiring when one repeated is being requested on this site to justify/defend with proof of what one state, and to in deeper and deeper justify ones experience or what tool one chosen and have preference to use. It is permitted and positive with different experience and preference. I respect that. It is good if we can communicate those and leave it at that and not deeper and deeper interrogations. I have already written my reply in above. If you are interested I encourage you to try (more) flat stitch and see for yourself. It is good to retry things. In the end... lets just leave it as a preference thing.
I hope that my writing has been informative, also to others. And I hope others clearly see that there are options; also that the latest craze of tech method in software is not photography, only yet another tool. In no way do I seek to mathematically or scientifically justify my approach. Flat stitch is nothing much different than rise/fall and shift has been since Ansel Adams times and long before. The only thing is one use several sensor locations within an imagre circle, instead of one. Photography is visual art, not pixel peeping. My eye tells me that images with wide lens are not as truthful as round stitches, and it is frank a heck easier to flat stitch only 2-3 shots rather than round stitch many more. The Mamiya 50mm shift is a temptation for me. That is one simple way; any shift lens. If I recall there is a used Nikon 28mm such that could be a good option for anyone with a Nikon or (adapter) Canon DSLR. Such lens also give rise and fall, which is useful for perspective corrections.
The other aspect is the type of camera / lens one prefer. I prefer medium format and lean towards large format. It makes the process more slow, or... better planned. I also did not imply that CS4 was the only or best software for stitching.
What I do encourage is anyone to try what they feel is within their interest and eye, and of course based on $$.
Something I found very interesting lately and can much encourage anyone with interest in panoramics is using 617 or 612 camera. Mine is a Fotoman 617 with a 90mm lens. It yields images on Velvia 50 measuring 168mm x 56mm - - - each in ONE SHOT, ONE CLICK = finished. It also lends one to see in a very different format, with proportions 3:1. It is dramatic, powerful.
Regards
Anders
-
If you are interested I encourage you to try (more) flat stitch and see for yourself. It is good to retry things. In the end... lets just leave it as a preference thing.
Anders,
I guess that the difference between you and me is that I did try both a lot and selected the technique that delivered the best results.
Cheers,
Bernard
-
A side note: The important is not only the image, but how one see it. A groundglass on a technical camera is one way. Or a panoramic viewfinder. The last can also be used on a DSLR. Seeing many stitched panoramas, frankly most do not have very good composition, because the photographer attempts to see the view in their mind.
This is not a side note, it's absolutely central. Most stitched panos are boring as all hell for the viewer, even if the photographer enjoyed making them. I'm sure a lot of this has to do with not having a satisfactory means of composition. A groundglass is unsurpassed for considered, deliberative work. (A series of Polaroids only makes things better but maybe I'm being too recherche).
-
This is not a side note, it's absolutely central. Most stitched panos are boring as all hell for the viewer, even if the photographer enjoyed making them. I'm sure a lot of this has to do with not having a satisfactory means of composition. A groundglass is unsurpassed for considered, deliberative work. (A series of Polaroids only makes things better but maybe I'm being too recherche).
Why blame some particularities of a technique instead of the photographers who authored those many boring images?
I have a hard time believing that any talented photographer actually needs a groundglass to compose a good image.
Cheers,
Bernard
-
[quote name='Dick Roadnight' date='Apr 6 2010, 05:49 PM' post='358557']
I think you are talking about camera movements, rather than moving the camera?
Yes - but I am trying to do stitching with tilt or swing movements.
Someone did post a picture involving moving the camera - the photo was of the building along a river or estuary in Scandinavia somewhere.
Photogrammetry, I think it is called. Once knew someone who did it along the entire hull of a sunken battleship.
...are you contemplating rectilinear stitching? ... you are looking at a rectangular building, and you want it to look rectangular on the print?
...or are you wanting perspective in the print of a rectangular object?
Not usually rectangular, more cigar shaped with thingies sprouting out. Aircraft. Yes, perspective is the big issue.
I just read Simmons and Stroebel and was feeling quite proud of myself. Then I checked out your Merklinger reference. Aaargh - complicated. His tutorial is rather neat, though.
-
Anders,
I guess that the difference between you and me is that I did try both a lot and selected the technique that delivered the best results.
Cheers,
Bernard
Bernard,
Kindly stop. What you are posting is along lines with
1. Wrong and false interpretations to drag me deeper and deeper into an argument. As I have stated in above I have done round stitching. I did not like it. I found it severe inferior to flat stitching, in particular for short lenses. While I like the methodology of flat stitching on a viewcamera, it is too time consuming both in setup and in zeroing tilts and swings. Instead I feel the solution is a technical camera simply because lens plane and sensor planes remain perfect parallel.
2. I find your posts beginning with the likes of offending and dragging me into argument
"Thanks for sharing your results on this. I am almost tempted to write "I told you". One more step and you will recognize the superiority of spherical stitching."
"Could you please point at evidences supporting your statement above? They are not backed up by either experience or mathematical theory."
and in the recent thread on low ISO "Have you used a d3x since we last discussed this? At that time you had not if my memory serves me well." albeit I there really tried very hard and long to write something diplomatic in return.
In frankness, your replies in line with above and over the years, indeed are same as some select others in the crowd on LL who seemingly seem to seek to justify that their "little thing" is being greater and better to whatever someone else write of THEIR experience and learnings with something that frank is superior in results. In fact it reminds me of a post Jack Fletcher made a few months back to another member Ray in line with how he had observed his behavior of dragging people into argument deeper and deeper to justify and prove that the method or gear they use is better than his, and which in end he always failed to accept. That feels psychotic. Please stop.
This forum is a tremendous resource. While I honest have impression that rotational stitching is pure crap compared to flat stitching, I have tried to be diplomatic not to state so, simply because I recognize that people can have different opinion. Instead I have tried my best to be fair and simply point at options and some pros and cons for people to form their own opinion.
It is a pleasure of informing of things we learnt as photographers, but that is a heck alot more than the round stitching and D3X that you have written so much to justify in posts of all kinds over the last two years. All across this forum it seems that the mere mentioning of "sensitive" things such as viewcamera/panoramic should be argued by certain folks for the "pros" for round stitching, film should be argued for that digital is superior (also when in my case I shoot both and like both), MFDB should be argued DSLR instead... now come on Bernard and others, this is complete B.S. nonsense and silly. Fine, in my case I use a quality digital back and shoot exclusive low ISO, thus why should I waste my time and go out try a D3X which is severe inferior for MY use?? Or... in this thread, why should I waste my time and go out and try more round stitching when I see it is a crap compared to flat stitching?? With flat stitch 2-3 shots only. Heck, 617 is truly a pure breeze, ONE SHOT, finished.
Regards
Anders
-
This is not a side note, it's absolutely central. Most stitched panos are boring as all hell for the viewer, even if the photographer enjoyed making them. I'm sure a lot of this has to do with not having a satisfactory means of composition. A groundglass is unsurpassed for considered, deliberative work. (A series of Polaroids only makes things better but maybe I'm being too recherche).
Agreed, very exactly.
But... was trying avoid... you know what...
. . .
Photography is about SEEING. Not having the tool to accurately compose an image is like a painter painting in the air, versus one painting in a set frame. Then... who have best chance to arrive at good composition???
Regards
Anders
-
It's all about showing something unique and real. Distorted panos that show nothing are pointless even if they take a lot of work.
Check out T.C. Wohlbruck's panographs from early 1900's .. if you can create something with as much character, than you know you are creating something special.
-
Bernard,
Kindly stop. What you are posting is along lines with
1. Wrong and false interpretations to drag me deeper and deeper into an argument. As I have stated in above I have done round stitching. I did not like it. I found it severe inferior to flat stitching, in particular for short lenses. While I like the methodology of flat stitching on a viewcamera, it is too time consuming both in setup and in zeroing tilts and swings. Instead I feel the solution is a technical camera simply because lens plane and sensor planes remain perfect parallel.
2. I find your posts beginning with the likes of offending and dragging me into argument
"Thanks for sharing your results on this. I am almost tempted to write "I told you". One more step and you will recognize the superiority of spherical stitching."
"Could you please point at evidences supporting your statement above? They are not backed up by either experience or mathematical theory."
and in the recent thread on low ISO "Have you used a d3x since we last discussed this? At that time you had not if my memory serves me well." albeit I there really tried very hard and long to write something diplomatic in return.
In frankness, your replies in line with above and over the years, indeed are same as some select others in the crowd on LL who seemingly seem to seek to justify that their "little thing" is being greater and better to whatever someone else write of THEIR experience and learnings with something that frank is superior in results. In fact it reminds me of a post Jack Fletcher made a few months back to another member Ray in line with how he had observed his behavior of dragging people into argument deeper and deeper to justify and prove that the method or gear they use is better than his, and which in end he always failed to accept. That feels psychotic. Please stop.
This forum is a tremendous resource. While I honest have impression that rotational stitching is pure crap compared to flat stitching, I have tried to be diplomatic not to state so, simply because I recognize that people can have different opinion. Instead I have tried my best to be fair and simply point at options and some pros and cons for people to form their own opinion.
It is a pleasure of informing of things we learnt as photographers, but that is a heck alot more than the round stitching and D3X that you have written so much to justify in posts of all kinds over the last two years. All across this forum it seems that the mere mentioning of "sensitive" things such as viewcamera/panoramic should be argued by certain folks for the "pros" for round stitching, film should be argued for that digital is superior (also when in my case I shoot both and like both), MFDB should be argued DSLR instead... now come on Bernard and others, this is complete B.S. nonsense and silly. Fine, in my case I use a quality digital back and shoot exclusive low ISO, thus why should I waste my time and go out try a D3X which is severe inferior for MY use?? Or... in this thread, why should I waste my time and go out and try more round stitching when I see it is a crap compared to flat stitching?? With flat stitch 2-3 shots only. Heck, 617 is truly a pure breeze, ONE SHOT, finished.
Anders,
I appreciate your attempt to be fair, but basing your advise on facts instead of opinions would help the original poster more. I am just reacting to emotional and innacurate information. You don't need to be diplomatic, please just try to be rigorous in your statements since you are trying to help.
My final comment will be about facts:
- It is impossible to distinguish a stitch planar projection from a single image captured with a wide angle lens as long as the angular coverage is the same,
- Those cases where cylindrical panos exhibit geometric distorsions like curved lines correspond to cases where a single lens would not be able to reach the same angular coverage (with or without flat stitching). If this is a problem then just stick to reasonnable angular coverage by using a long enough lens, it will at worst limit you to what you could have done with a single lens, but with much higher resolutions if needed,
- Cylindrical stitching can be done without having to care about the accuracy of the positioning between sensor and lens since the camera itself (including sensor and lens) is moved. It can therefore be executed fast which extends considerably the range of applications that can be handled with cylindrical stitching (),
- The equipment needed to do cylindrical stitching is indendant from the camera used. It is therefore a long term investement that is flexible,
- Cylindrical stitching can be done with whatever camera, regardless of its sensor size and I know more than a few people doing it with MFDB. I am not sure why you comment about the D3x in this thread.
About the D3x, the only thing I did in that other thread was to ask you whether you had actually used a camera you were commenting negatively about. Here again, I don't need your diplomacy, just the honnesty of sticking to what you know from first hand experience.
Cheers,
Bernard
-
Agreed, very exactly.
Photography is about SEEING. Not having the tool to accurately compose an image is like a painter painting in the air, versus one painting in a set frame. Then... who have best chance to arrive at good composition???
Anders,
I used to play indoor soccer and once injured the ligaments of my right knee. Because of that it can get painful after 6 or 7 hours walk when the elevation difference is more than 1000m downwards. That is my own issue and limitation. I have never consider telling people around me that they should pick their hiking trails so as to avoid walking more than 1000m downwards, nor that buying knee support is a must when hiking.
You are unable to compose without a viewfinder. Fine. That is not biggie and there are ways around this. There might or might not be many photographers like you. I don't know of any stats covering this aspect of human activity.
Now... why do you feel that you have to advise the world not to consider some shooting techniques because you own inabilities are a poor match with their characteristics?
Cheers,
Bernard
-
Why blame some particularities of a technique instead of the photographers who authored those many boring images?
I have a hard time believing that any talented photographer actually needs a groundglass to compose a good image.
I agree. Heck, I don't even need a camera to see if a useful composition is present and the light(ing) is adequate, even before I take my camera out of the bag and fit the lens that I suspect will suit the job best. Perhaps some people need more help, could be. One of the great things about stitching is that we are no longer restricted by the default aspect ratio of the image format. We don't crop to adjust but we can add to adjust (or we can conform to a desired output format).
Cheers,
Bart
-
No DSLR or other camera with an anti-aliasing filter approaches the "per pixel" quality of any medium format digital back, let alone the best modern ones, but, you could use your technique with a top quality camera.
I know your technique works, and you have photos to prove it, but for ultimate image quality, and correct perspective, one needs to flat stitch with a MFDB and a quality lens.
This is because the stitching software has to distort and re-sample the image, and this can only degrade the image, causing loss of resolution.
I am aware that the better stitching programs do a good job, and the loss of resolution is minimal, and would be difficult to see or quantify in a high-res image.
Hi Dick,
IMHO there is not a single best solution for all circumstances, it depends on those circumstances. I just want to mention some considerations.
Flat stitching quality depends on the quality of the edges of the image circle. Therefore image quality is not as good across the image, unless one closes down the aperture enough so that the center suffers a bit from difraction and the edges improve due to reduced residual aberrations, and everything evens out a bit. Flat stitching does allow for simple postprocessing.
Rotating the camera through it's no-parallax point always uses the optically best part of the lens' image circle, and with some overlap even the edges are not used at all, so lens (edge) quality becomes less important (as long as it's good in the center). Your argument about resolution loss due to resampling only, and I repeat only, applies to those areas where it leads to interpolation in the sense of stretching the data beyond its sampling density. This is usually only the case where an extreme angle of view is chosen that probably would exceed the FOV of a single lens. Besides, shooting panorama tiles with a slightly longer focal length will even compensate for that, delivering superior resolution which usually requires downsampling rather than upsampling. It's just a bit more work/time to composite the full image, although as a bonus it automatically corrects for lens distortions and slight keystoning errors.
There's a place for both techniques, although I rotate more than I do flat stitching but that's probably due to the freedom in resolution and aspect ratio.
Cheers,
Bart
-
Hi Dick,
IMHO there is not a single best solution for all circumstances, it depends on those circumstances. I just want to mention some considerations.
Flat stitching quality depends on the quality of the edges of the image circle. Therefore image quality is not as good across the image, unless one closes down the aperture enough so that the center suffers a bit from difraction and the edges improve due to reduced residual aberrations, and everything evens out a bit. Flat stitching does allow for simple postprocessing.
Rotating the camera through it's no-parallax point always uses the optically best part of the lens' image circle, and with some overlap even the edges are not used at all, so lens (edge) quality becomes less important (as long as it's good in the center).
...Besides, shooting panorama tiles with a slightly longer focal length will even compensate for that, delivering superior resolution which usually requires downsampling rather than upsampling. It's just a bit more work/time to composite the full image, although as a bonus it automatically corrects for lens distortions and slight keystoning errors.
There's a place for both techniques, although I rotate more than I do flat stitching but that's probably due to the freedom in resolution and aspect ratio.
Cheers,
Bart
Apart from the bit I edited out about resolution loss due to re-sampling, I agree with everything you say.
Bernard says he can, using 9 shots with a 100mm produce a photo indistinguishable from a pic taken with a 24 mm lens on the same camera, but I would have expected that, using a 100mm lens on a D3X, one could produce a picture comparable to what you could produce with a P65+ and a wide angle with a single shot... but, I have a MFDSLR and a wide angle, so I would not bother, except for cylinder panos for very wide FOV.
I have a Schneider Super Angulon 47XL, which I hope I could use for FOVs up to about 100 degrees with shift and stitch... until I get one of those £5K digi UWAs.
-
All my work for the last 3 to 4 years has been stitching (alan-george.com if interested). I just recently switched from nodal point stitching with RRS and 5DII to single image circle stitching with a Cambo RS and PhaseOne P40+. For me the biggest difference is the pixel quality of the P40+ surpasses the 5DII by quite a bit. At 100% the 5DII (and other 35mm DSLRs) are always a bit fuzzy no matter how I attempt to sharpen them, but the P40+ with the digital lens at 100% are stunning. Blow me away. This translates for me into much larger prints.
-
All my work for the last 3 to 4 years has been stitching (alan-george.com if interested). I just recently switched from nodal point stitching with RRS and 5DII to single image circle stitching with a Cambo RS and PhaseOne P40+. For me the biggest difference is the pixel quality of the P40+ surpasses the 5DII by quite a bit. At 100% the 5DII (and other 35mm DSLRs) are always a bit fuzzy no matter how I attempt to sharpen them, but the P40+ with the digital lens at 100% are stunning. Blow me away. This translates for me into much larger prints.
Possibly the first time I've gone to a photographer's site from a forum link and actually liked some of the stuff there.
-
Hi,
My view is that excellent results can be achieved by both means.
We have excellent programs that can generate rotational panoramas with great ease. Rotational panoramas can be made with any gear and at great ease. I have shot rotational panoramas with both 12 and 560 mm on DSLRs.
In my view it's not necessarily the case that shift panoramas would be advantageous to rotational panoramas. On the other hand, shifted panoramas are made in camera while rotational panoramas are made in the computer. When I shoot rotational panoramas I always assume that cropping would be included. Very simple gear, like this:
[attachment=21381:Screen_s...00.21_PM.png]
Can achieve excellent results with programs like Autopano Pro (using Smart Blend)
It's like the old proverb, "F/8 and be there".
Best regards
Erik
Hi guys,
I just have a little theoretical question about stitching with a view camera ... Can one easily combine normal stitching by shifting and nodal point stitching with a nodal point adapter in order to get really gigantic panoramas? Or are there any caveats?
One main advantage of using view cameras is to my understanding the capability of easy, seamless stitching by way of shifting the digital back. Coming from 35mm though, I read a lot about panoramic heads - such as the one from manfrotto - as another viable way of creating huge panoramas.
My thought now is to combine these two methods. I.e. by combining multiple 2 x 2 shifted images that were sequentially created side by side with the help of the panoramic head.
Is there anywhere a list where I can see the nodal points of the Rodenstock/Schneider lenses?
Kind regards
Paul
-
Anders,
I'd say that Bernard has definitively a point when he say that excellent results can be achieved with standard equipment, like normal DSLRs. Now, the equipment that Bernard is using is not plain vanilla DSLR but top of the class Nikon equipment. Never the less, making the best use of what we have is generally a very good idea. It is quite obvious that equipment built for a set condition has some benefits over more generic equipment. On the other hand, the option that Bernard advocates, that is to use generic equipment in a flexible way also have a lot of advantages.
In a way, I would say that the person behind the lens is still the most important factor. In the digital era much of the image making moved in the light room, in the analogue time it used to be the dark room. Ansel Adams, who is often mentioned in this context, had a very scientific approach to image making. What he achieved by using advanced science and art in the chemical dark room is much more achievable in the software based "Light room".
Our hobby is a bit fixed on technology. Bernard say essentially: "You can do it with a DSLR", in my view it is a good piece of information. He doesn't really says that you cannot do it better with an MFDB.
Best regards
Erik
Bernard,
Kindly stop. What you are posting is along lines with
1. Wrong and false interpretations to drag me deeper and deeper into an argument. As I have stated in above I have done round stitching. I did not like it. I found it severe inferior to flat stitching, in particular for short lenses. While I like the methodology of flat stitching on a viewcamera, it is too time consuming both in setup and in zeroing tilts and swings. Instead I feel the solution is a technical camera simply because lens plane and sensor planes remain perfect parallel.
2. I find your posts beginning with the likes of offending and dragging me into argument
"Thanks for sharing your results on this. I am almost tempted to write "I told you". One more step and you will recognize the superiority of spherical stitching."
"Could you please point at evidences supporting your statement above? They are not backed up by either experience or mathematical theory."
and in the recent thread on low ISO "Have you used a d3x since we last discussed this? At that time you had not if my memory serves me well." albeit I there really tried very hard and long to write something diplomatic in return.
In frankness, your replies in line with above and over the years, indeed are same as some select others in the crowd on LL who seemingly seem to seek to justify that their "little thing" is being greater and better to whatever someone else write of THEIR experience and learnings with something that frank is superior in results. In fact it reminds me of a post Jack Fletcher made a few months back to another member Ray in line with how he had observed his behavior of dragging people into argument deeper and deeper to justify and prove that the method or gear they use is better than his, and which in end he always failed to accept. That feels psychotic. Please stop.
This forum is a tremendous resource. While I honest have impression that rotational stitching is pure crap compared to flat stitching, I have tried to be diplomatic not to state so, simply because I recognize that people can have different opinion. Instead I have tried my best to be fair and simply point at options and some pros and cons for people to form their own opinion.
It is a pleasure of informing of things we learnt as photographers, but that is a heck alot more than the round stitching and D3X that you have written so much to justify in posts of all kinds over the last two years. All across this forum it seems that the mere mentioning of "sensitive" things such as viewcamera/panoramic should be argued by certain folks for the "pros" for round stitching, film should be argued for that digital is superior (also when in my case I shoot both and like both), MFDB should be argued DSLR instead... now come on Bernard and others, this is complete B.S. nonsense and silly. Fine, in my case I use a quality digital back and shoot exclusive low ISO, thus why should I waste my time and go out try a D3X which is severe inferior for MY use?? Or... in this thread, why should I waste my time and go out and try more round stitching when I see it is a crap compared to flat stitching?? With flat stitch 2-3 shots only. Heck, 617 is truly a pure breeze, ONE SHOT, finished.
Regards
Anders
-
Hey guys!
Thy for all chiming in ... can't wait to get my MFDB to try stitching out ...
Just saw this: http://gigapansystems.com/gigapan-products...oduct-page.html (http://gigapansystems.com/gigapan-products/gigapan-epic-pro-product-page.html)
Anyone thinking of trying one of these gigapan-machines out with a MFD-camera?
Regards
Paul
-
Hey guys!
Thy for all chiming in ... can't wait to get my MFDB to try stitching out ...
Just saw this: http://gigapansystems.com/gigapan-products...oduct-page.html (http://gigapansystems.com/gigapan-products/gigapan-epic-pro-product-page.html)
Anyone thinking of trying one of these gigapan-machines out with a MFD-camera?
Regards
Paul
Could be very useful, when they support the H4D. (Cheap substitute for a Hex-stitch back, which, with a 60 Mpx sensor, would be high enough res for most jobs.)
These systems are obviously limited by the lens, and if the lens/sensor will not resolve bricks at 1km, then the system will not.
...but a D3X with a good lens might produce as good a result ... and taking ¿3 or 10? times as many pictures would not be a problem if it was fully robotic/automatic.
If everybody asks them when they are going to support the H4D, they might move it up the list.
The "trouble" with this is that it will allow anyone to produce very high res, high quality panos without spending mega-bucks, and without specialist knowledge.
Re-edit... most of the pictures are awful (due to variations in light between the exposures) so it seems that it does need luck, or a good day, or specialist skill to produce a good picture. ...it would be nice to have a system which automatically adjusts the colour temperature for individual exposures, but I think that would be impossible with a gray card in each picture!
Using a higher res camera (e.g. H4D-60) would reduce the total time for the stitch, making it easier to get the light the same, and, as there would be fewer pictures, it would be easier to adjust them for consistent colour.
-
The "trouble" with this is that it will allow anyone to produce very high res, high quality panos without spending mega-bucks, and without specialist knowledge.
Hi Dick,
Fortunately, resolution alone doesn't make high quality images/panos. Especially for wide angles of view, dynamic range is also important for a technically good quality. But even more important is the photographer's vision, choice of subject/composition, timing, lighting, etc., etc.
Tools that help to acquire technically good images will help freeing up time to spend on the creative aspects of image creation.
Cheers,
Bart
-
Hi,
Programs like Autopano are very good at taking care of colour differences. Light is probably pretty constant between exposures so it's OK to use a single gray card shot. Don't think that a larger format would help on the issue, anyway.
Best regards
Erik
Could be very useful, when they support the H4D. (Cheap substitute for a Hex-stitch back, which, with a 60 Mpx sensor, would be high enough res for most jobs.)
These systems are obviously limited by the lens, and if the lens/sensor will not resolve bricks at 1km, then the system will not.
...but a D3X with a good lens might produce as good a result ... and taking ¿3 or 10? times as many pictures would not be a problem if it was fully robotic/automatic.
If everybody asks them when they are going to support the H4D, they might move it up the list.
The "trouble" with this is that it will allow anyone to produce very high res, high quality panos without spending mega-bucks, and without specialist knowledge.
Re-edit... most of the pictures are awful (due to variations in light between the exposures) so it seems that it does need luck, or a good day, or specialist skill to produce a good picture. ...it would be nice to have a system which automatically adjusts the colour temperature for individual exposures, but I think that would be impossible with a gray card in each picture!
Using a higher res camera (e.g. H4D-60) would reduce the total time for the stitch, making it easier to get the light the same, and, as there would be fewer pictures, it would be easier to adjust them for consistent colour.
-
Hi,
Programs like Autopano are very good at taking care of colour differences. Light is probably pretty constant between exposures so it's OK to use a single gray card shot. Don't think that a larger format would help on the issue, anyway.
Best regards
Erik
It seems that the Gigapan systems software is useless, but I suppose you could use their hardware with quality software.
For really good colour matching, the pano software would need to be a plugin to the raw processor, but that would be tooo much to ask, (especially for manufacturer-specific programs like phocus) so you adjust the white balance in each picture, so that the colour of e.g. the sky, building, sea... changes gradually from one side of the picture to the other?
-
guys, ... having said that the advent of Gigapan-systems can be seen as a democratization of high-res photography - are there, in your opinion, still scenarios where a high end view camera setup with a MFDB can produce higher quality imagery than let's say 5D MK II with a gigapan epic pro and a 300mm lens?
For example, take Andreas Gursky's famous picture of the Montparnasse residency in Paris:
http://www.moma.org/interactives/exhibitio...arnasse_pop.jpg (http://www.moma.org/interactives/exhibitions/2001/gursky/images/montparnasse_pop.jpg)
His picture is amazing because of its high-resolution and sheer awesomeness in the eye of the beholder:
http://www.creativepooldesign.de/mood_art/.../gursky_14b.jpg (http://www.creativepooldesign.de/mood_art/artimages/gursky/gursky_14b.jpg)
It is one of my goals to reproduce that image this summer with my newly ordered H4D-60 and a view camera setup from camera (I'm pondering the 43 Schneider XL for that matter); but since I saw this new gigapan head I'm wondering whether one cannot achieve even superior quality (than what I could probably produce with the MFDB-setup) just with a 6000$ setup consisting of a 5D, 300mm and the pan head.
Or are there still reasons why the high-end setup might yield better results? (let's exclude the convenience factor for a sec)
Regards
Paul
-
It seems that the Gigapan systems software is useless, but I suppose you could use their hardware with quality software.
For really good colour matching, the pano software would need to be a plugin to the raw processor, but that would be tooo much to ask, (especially for manufacturer-specific programs like phocus) so you adjust the white balance in each picture, so that the colour of e.g. the sky, building, sea... changes gradually from one side of the picture to the other?
Not sure whether they support Hassy files, but autopano pro does also support raw files.
Besides nothing prevents you from working in manual more with a fixed white balance when you capture the images with gigapan or any of the other available robotic heads for that matter. Such heads have been available for years, the gigapan is just cheaper.
If you work in manual mode, you will have zero exposure issue. The only potential problem is light fall off in the corner of your lens, so it is always better to use an aperture that minimizes light fall off. I know that my Zeiss 100mm f2.0 has nearly negligible light fall off around f8, which is one of the reasons why it is such a great stitching lens.
Cheers,
Bernard
-
guys, ... having said that the advent of Gigapan-systems can be seen as a democratization of high-res photography - are there, in your opinion, still scenarios where a high end view camera setup with a MFDB can produce higher quality imagery than let's say 5D MK II with a gigapan epic pro and a 300mm lens?
For example, take Andreas Gursky's famous picture of the Montparnasse residency in Paris:
http://www.moma.org/interactives/exhibitio...arnasse_pop.jpg (http://www.moma.org/interactives/exhibitions/2001/gursky/images/montparnasse_pop.jpg)
His picture is amazing because of its high-resolution and sheer awesomeness in the eye of the beholder:
http://www.creativepooldesign.de/mood_art/.../gursky_14b.jpg (http://www.creativepooldesign.de/mood_art/artimages/gursky/gursky_14b.jpg)
It is one of my goals to reproduce that image this summer with my newly ordered H4D-60 and a view camera setup from camera (I'm pondering the 43 Schneider XL for that matter); but since I saw this new gigapan head I'm wondering whether one cannot achieve even superior quality (than what I could probably produce with the MFDB-setup) just with a 6000$ setup consisting of a 5D, 300mm and the pan head.
Or are there still reasons why the high-end setup might yield better results? (let's exclude the convenience factor for a sec)
You don't need a gigapan to achieve high resolution. Many of my panos done manually are in the 200+ megapixels range. The example below was downsized from 200 megapixel to just 60 megapixels.
(http://farm3.static.flickr.com/2622/3833720762_a85d3c5038_b.jpg)
It is a high contrast scene that was capture without HDR techniques with a D3x. I suggest checking the dark areas on the right side of the temple that appeared darker to the eye than they are in this rendition. There is hardly any noise.
The link below points to a 60 megapixel version.
http://www.flickr.com/photos/bernardlangui...57618032772894/ (http://www.flickr.com/photos/bernardlanguillier/3833720762/sizes/o/in/set-72157618032772894/)
The gigapan is a cheap option, but people serious about really high res panos have been using devices like the claus rodeon head for quite a few years. http://www.dr-clauss.de/VRStationST_EN.htm (http://www.dr-clauss.de/VRStationST_EN.htm)
Cheers,
Bernard
-
Hi,
I seldom feel the need of doing much, just relying on Autopano Pro. Don't really have a lot of panos on the web.
http://echophoto.dnsalias.net/ekr/index.ph...a-and-stitching (http://echophoto.dnsalias.net/ekr/index.php/photoarticles/panorama-and-stitching)
This one is a bit extreme, shot with a Sigma 12-24/4.5-5.6 from free hand.
http://echophoto.dnsalias.net/ekr/index.ph...s-quick-a-dirty (http://echophoto.dnsalias.net/ekr/index.php/photoarticles/21-panoramas-quick-a-dirty)
Best regards
Erik
It seems that the Gigapan systems software is useless, but I suppose you could use their hardware with quality software.
For really good colour matching, the pano software would need to be a plugin to the raw processor, but that would be tooo much to ask, (especially for manufacturer-specific programs like phocus) so you adjust the white balance in each picture, so that the colour of e.g. the sky, building, sea... changes gradually from one side of the picture to the other?
-
You don't need a gigapan to achieve high resolution. Many of my panos done manually are in the 200+ megapixels range. The example below was downsized from 200 megapixel to just 60 megapixels.
(http://farm3.static.flickr.com/2622/3833720762_a85d3c5038_b.jpg)
It is a high contrast scene that was capture without HDR techniques with a D3x. I suggest checking the dark areas on the right side of the temple that appeared darker to the eye than they are in this rendition. There is hardly any noise.
The link below points to a 60 megapixel version.
http://www.flickr.com/photos/bernardlangui...57618032772894/ (http://www.flickr.com/photos/bernardlanguillier/3833720762/sizes/o/in/set-72157618032772894/)
The gigapan is a cheap option, but people serious about really high res panos have been using devices like the claus rodeon head for quite a few years. http://www.dr-clauss.de/VRStationST_EN.htm (http://www.dr-clauss.de/VRStationST_EN.htm)
Cheers,
Bernard
Bernard, what do you mean bay manually? You just stood there and made different exposures, trying to make them overlap by turning your body, without tripod? I must admit, the image looks great! What focal length is it? I Imagine doing multi-row panorams is difficult this way??
And to conclude, is there any image-quality advantage anymore in using a MFDB-system or are you basically saying a D3x, a long lens, a pan-head (good hands) is all one needs today?
Regards
Paul
-
Hi,
Bernard makes great pictures. He is also very good at the technical stuff and a great ambassador for Nikon. That said, achieving good results with panos is actually quite easy. Achieving excellence, like Bernard's work, may take some more effort.
The image here is http://echophoto.dnsalias.net/ekr/multimed...dafoss_720p.mov (http://echophoto.dnsalias.net/ekr/multimedia/TPOI_Godafoss_720p.mov) actually my second "pano", I really hooked up on that technique since than.
Best regards
Erik
Bernard, what do you mean bay manually? You just stood there and made different exposures, trying to make them overlap by turning your body, without tripod? I must admit, the image looks great! What focal length is it? I Imagine doing multi-row panorams is difficult this way??
And to conclude, is there any image-quality advantage anymore in using a MFDB-system or are you basically saying a D3x, a long lens, a pan-head (good hands) is all one needs today?
Regards
Paul
-
Bernard, what do you mean bay manually? You just stood there and made different exposures, trying to make them overlap by turning your body, without tripod? I must admit, the image looks great! What focal length is it? I Imagine doing multi-row panorams is difficult this way??
And to conclude, is there any image-quality advantage anymore in using a MFDB-system or are you basically saying a D3x, a long lens, a pan-head (good hands) is all one needs today?
Paul,
I used a spherical pano head from really right stuff to capture this image. Capturing a 200 megapixel image like this one with a 20+ megapixel DSLR takes perhaps 2 mins at most.
The result will be perfect if:
1. you locate your lens at its nodal point and use a spherical pano head enabling this (like the RRS I am using)
2. the subject is perfectly static
3. you use a good pano software like PTgui or Autopano pro/giga
If the subject is not perfectly static (wind, people, clouds, water,...) then you need to somehow blend the images smartly to make sure that the movement in the scene does not result in any discontinuity in the final image. Most softwares do this more or less automatically if the characteristic size of the subject is not too large relative to the total image. So it works overall very well for landscapes and large city scapes, but you sometimes have problems with moving foregrounds like grass... this problem will be present with both cyliundrical stitching and flat stitching by the way.
There is little doubt that it will be much easier and relaxing to take a single 60 megapixel image with a H4D60/P65+ compared to doing stitching though. I am not willing to spend that kind of money for a camera and find stitching to be fun for my applications, but I would probably see things differently as a working pro.
More samples here: http://www.flickr.com/photos/bernardlangui...57600916381270/ (http://www.flickr.com/photos/bernardlanguillier/sets/72157600916381270/)
Regarding your D3x vs MFDB question, there have been many dicussions around here. The general agreement is that:
- backs have better pixel quality thanks to the lack of AA filter,
- backs have still more DR, but the gap has been reduced significantly with the top DSLRs like the D3x.
Cheers,
Bernard
-
Not sure whether they support Hassy files, but autopano pro does also support raw files.
Besides nothing prevents you from working in manual more with a fixed white balance when you capture the images with gigapan or any of the other available robotic heads for that matter. Such heads have been available for years, the gigapan is just cheaper.
If you work in manual mode, you will have zero exposure issue. The only potential problem is light fall off in the corner of your lens, so it is always better to use an aperture that minimizes light fall off. I know that my Zeiss 100mm f2.0 has nearly negligible light fall off around f8, which is one of the reasons why it is such a great stitching lens.
Cheers,
Bernard
Thanks, Bernard.
...as you said, one of the nice things about 2 axis panos is that you only use the best bit of the lens, and Phocus is supposed to eliminate any trace of vignetting (with Hasselblad lenses).
The white balance problem is that, when the sun goes behind a cloud, you go from "daylight" to "cloudy", and it screws up the colour temperature even if you compensate the exposure.
I think my 10 kg Manfrotto tripod with #400 geared head, and the geared rear rise & fall of the Sinar P2/3 for the vertical (row adjustment) adjustment, and two rail clamps and a base board allowing me to use the monorail for nodal point adjustment, and apo-digitars should allow me to do three row cylinder panos without spending any extra cash?
-
I had and used a RRS “Ultimate Omni-Pivot Package” first while shooting 35mm landscape then my Mamiya AFD after I switched to medium format. While I got get images I nevertheless always had large portions of the stitched images lacking detail and/or pixels. The images were large enough to crop out the offending areas and still achieve a large enough image with the detail I wanted. Then I moved to a technical camera.
I’ve been shooting with a Cambo WRS/P45+ combination for close to two years now and remain constantly amazed with the total image available to me after stitching. I can ‘flat stitch” using the movements of the P45+ around the image circle and if I’m rock solid level end up with a stitched image that has way more usable footprint than anything I ever achieved using the RRS setup.
Almost anyone can take a 2-shot panorama handheld and get a good result – (hell, I’ve done them) likewise just about anyone can get a great stitched image using gear by RRS or similar so long as you take your time in setting everything up. Having experience in the above two concepts I feel safe to say that at least for me – I can get constant great results from using the movements of my Cambo; again so long as I take the time in the beginning to properly set everything up.
This has all been from my own personal experience in shooting landscape panoramas and stitching 2 to 6 captures to achieve the 1 image.
Don