Ad
Ad
Ad
Pages: « 1 [2] 3 4 »   Bottom of Page
Print
Author Topic: progessive stretch  (Read 12538 times)
marc gerritsen
Sr. Member
****
Offline Offline

Posts: 299


WWW
« Reply #20 on: May 15, 2010, 11:02:55 PM »
ReplyReply




good shot and indeed agree 100% to not correct
m
Logged

marc gerritsen
Sr. Member
****
Offline Offline

Posts: 299


WWW
« Reply #21 on: May 15, 2010, 11:04:39 PM »
ReplyReply

Quote from: Hening Bettermann
When looking for a software that could correct perspective by the numbers (of degrees of camera tilt), I found Digital Photo Shifter:

http://www.deraltenburger.de/Win-Tools/Index.html

It can do what I wanted, with or without proportional compressing of the upper part of high vertical objects. (The latter is what you want, if I got it right).

The tool comes in different versions:
1-as a standalone Windows app,
2-as a plug-in for PS
3-as a plug-in for FixFoto,
4-as a plug-in for the Windows version of PhotoLine

http://www.pl32.com/

Only #4 offers 16 bit color depth. - German only...and I found the surface rather confusing.

Good light - Hening.

will investigate, too bad only for pc so will try tomorrow at the office
viele dank
m




Logged

Dick Roadnight
Sr. Member
****
Offline Offline

Posts: 1730


« Reply #22 on: May 16, 2010, 03:26:42 PM »
ReplyReply

Quote from: CBarrett
While I find the decreasing window size in the shot of the building unsettling ( I prefer the "stretch the photo vertical" version).  I've really got to applaud not just the fact that Marc is so creatively inventive, but that he is willing to share the results with us.  That building could obviously not have been shot from that distance with a perspective control camera (or lens) while capturing the full height.

Here comes an interesting dilemma...  The interpolation introduced in some applications of this technique might send some of us pixel peepers running for the hills, but if you're client is only ever going to use it on the web and in powerpoints AND the shot simply cannot be achieved with a plumb and level camera.... then get the friggin shot however you can and fix it!

For 99% of my work the image circles of my lenses accomplish what I need to do while keeping my camera quite parallel to mother earth, but 1% of the time....

Nice work, Marc.
100+ degree lenses (like the Super-Angulon 47 XL and Apo-digitar 47XL) help, but, if you can shoot from the building opposite, (or a cherry picker, mast or whatever) my
 
[a href=\'index.php?showtopic=43338\']Virtual Viewpoint technique[/a]

can be very useful, and allow you to get the foreground in, and view the building from ground eye-level, without stretching.
Logged

Hasselblad H4, Sinar P3 monorail view camera, Schneider Apo-digitar lenses
JonathanBenoit
Sr. Member
****
Offline Offline

Posts: 414


WWW
« Reply #23 on: May 16, 2010, 03:57:09 PM »
ReplyReply

Quote from: Dick Roadnight
100+ degree lenses (like the Super-Angulon 47 XL and Apo-digitar 47XL) help, but, if you can shoot from the building opposite, (or a cherry picker, mast or whatever) my
 
[a href=\'index.php?showtopic=43338\']Virtual Viewpoint technique[/a]

can be very useful, and allow you to get the foreground in, and view the building from ground eye-level, without stretching.


Maybe in theory. Has anyone other than yourself seen an example of this?
Logged

Dick Roadnight
Sr. Member
****
Offline Offline

Posts: 1730


« Reply #24 on: May 16, 2010, 04:20:19 PM »
ReplyReply

Quote from: JonathanBenoit
Maybe in theory. Has anyone other than yourself seen an example of this?
I have not done it myself yet, but I cannot see why it would not work.
Logged

Hasselblad H4, Sinar P3 monorail view camera, Schneider Apo-digitar lenses
marc gerritsen
Sr. Member
****
Offline Offline

Posts: 299


WWW
« Reply #25 on: May 16, 2010, 06:13:57 PM »
ReplyReply

Quote from: Dick Roadnight
100+ degree lenses (like the Super-Angulon 47 XL and Apo-digitar 47XL) help, but, if you can shoot from the building opposite, (or a cherry picker, mast or whatever) my
 
[a href=\'index.php?showtopic=43338\']Virtual Viewpoint technique[/a]

can be very useful, and allow you to get the foreground in, and view the building from ground eye-level, without stretching.


with all due respect, i really don't understand your theory
but it is probably me; i am not into theories, i always want to see examples as well
Logged

adammork
Full Member
***
Offline Offline

Posts: 156


« Reply #26 on: May 16, 2010, 07:50:53 PM »
ReplyReply

Quote from: Dick Roadnight
if you can shoot from the building opposite, (or a cherry picker, mast or whatever) my
 
[a href=\'index.php?showtopic=43338\']Virtual Viewpoint technique[/a]

can be very useful, and allow you to get the foreground in, and view the building from ground eye-level, without stretching.

I look really forward to see an image done with this technique.... really   I think, it will be some kind of a challenge to mix, let's say 10 images, where everything in the background, and the building for that matter, have a different perspective, due to different hight of viewpoint of the time of capture.

I know from real life shooting, that if I move my viewpoint 10 millimeters when doing a stitch, for example with a dslr and a shiftlens, then nothing lines up perfect anymore....( thats why a tech camera, where the lens can be fixed when stitching, is so handy) so aligning 10 images, where you move your viewpoint 3 meters between each capture, is a thing I really look forward to see the outcome from.....

At best, I think you will see an image with 10 different perspective's - one for each viewpoint - how this will give you a feeling of viewing a building AND surroundings, from ground level is beyond me.... but please, prove me wrong with an example.

sorry for the rant, but you are presenting here an unproven technique as "very useful"

Marc here in the opposite, presents in a perfect way, a technique with a lot of samples and how to's - so everyone can enjoy it and use it if they want - thanks a lot Marc!  

/adam
« Last Edit: May 16, 2010, 09:51:01 PM by adammork » Logged

marc gerritsen
Sr. Member
****
Offline Offline

Posts: 299


WWW
« Reply #27 on: May 16, 2010, 09:11:31 PM »
ReplyReply

you are welcome adam!

you do some great work there in denmark!
also love the navigation of the projects on your website.

cheers
m


Logged

haefnerphoto
Sr. Member
****
Offline Offline

Posts: 613


WWW
« Reply #28 on: May 16, 2010, 09:29:42 PM »
ReplyReply

Marc, I think your approach works very well.  I've had this problem more often with interiors than exteriors and hope you can be a little more precise in your explanation of the process.  Perhaps a screen grab of the action would suffice.  Thanks for posting your progress and while the interpolation is apparent, I agree with you that it's commercially acceptable.  Jim
« Last Edit: May 16, 2010, 09:30:25 PM by haefnerphoto » Logged

adammork
Full Member
***
Offline Offline

Posts: 156


« Reply #29 on: May 16, 2010, 09:31:51 PM »
ReplyReply

Quote from: marc gerritsen
you are welcome adam!

you do some great work there in denmark!
also love the navigation of the projects on your website.

cheers
m

thank's a lot,

And it goes right back to you as well - your type of subjects is, IMO, often overdone with a lot of additional light on the scenes - your work seems to communicate the feeling of the space and object in a good way!

very best,
adam
Logged

JeffKohn
Sr. Member
****
Offline Offline

Posts: 1671



WWW
« Reply #30 on: May 16, 2010, 10:17:08 PM »
ReplyReply

Quote
I look really forward to see an image done with this technique.... really wink.gif I think, it will be some kind of a challenge to mix, let's say 10 images, where everything in the background, and the building for that matter, have a different perspective, due to different hight of viewpoint of the time of capture.
Yes, you will most definitely have major parallax issues with this approach. There's a guy who used this method to photograph redwoods, although instead of a cherry-picker or crane he gained height by climbing neighboring trees. The images are interesting because the tree is shown in correct perspective from a fairly close-up viewpoint, which makes them all the more impressive. The background actually repeats in layers though, because of the way perspective changes. It it's an interesting effect; and since the tree is the real subject, I don't really see it as a flaw in his photos. But somehow I doubt commercial architecture clients would find the result acceptable.

Here's a link to the book.
Logged

marc gerritsen
Sr. Member
****
Offline Offline

Posts: 299


WWW
« Reply #31 on: May 16, 2010, 11:33:17 PM »
ReplyReply

Quote from: haefnerphoto
Marc, I think your approach works very well.  I've had this problem more often with interiors than exteriors and hope you can be a little more precise in your explanation of the process.  Perhaps a screen grab of the action would suffice.  Thanks for posting your progress and while the interpolation is apparent, I agree with you that it's commercially acceptable.  Jim


hi james

very happy to do that, but i have posted 3 samples and don't know which one you refer to
A the interior
B the building night shot
C the building day shot

cheers
m
Logged

BartvanderWolf
Sr. Member
****
Online Online

Posts: 3012


« Reply #32 on: May 17, 2010, 03:40:14 AM »
ReplyReply

Quote from: adammork
I look really forward to see an image done with this technique.... really   I think, it will be some kind of a challenge to mix, let's say 10 images, where everything in the background, and the building for that matter, have a different perspective, due to different hight of viewpoint of the time of capture.

Hi Adam,

This is not new, it is only restricted to aligning a single flat plane, e.g. facade, storefront, or mural, without any depth/foreground/background. As long as there is no depth in the image things can work just fine, PTAssembler even does it automatically with its Camera Position Optimizer.

PTAssembler also offers adjustable projection methods that allow progressive compression towards edges/corners, much easier than the proposed stretching method (and with a better resampling method). As far as I know PTGUI also has a compression feature, but not as flexible as PTAssembler's choice of projections to mitigate the anamorphic distortion effect, including a hybrid method that can mix several of four projection methods in 4 quadrants of the image.

BTW, the whole visual 'stretching' issue is caused by viewing the output image from the wrong position/distance (too far away), hence the anamorphic distortion. When the focal length is relatively short, we tend to view the image from too far away. When we view the rectilinear output image from a distance that corresponds to the focal length versus the sensor array size ratio, we would see the correct perspective.

Cheers,
Bart
« Last Edit: May 17, 2010, 09:00:09 AM by BartvanderWolf » Logged
fredjeang
Guest
« Reply #33 on: May 17, 2010, 07:19:22 AM »
ReplyReply

I am more and more interested by Marc's approach I must say.
Logged
haefnerphoto
Sr. Member
****
Offline Offline

Posts: 613


WWW
« Reply #34 on: May 17, 2010, 08:20:48 AM »
ReplyReply

Quote from: haefnerphoto
Marc, I think your approach works very well.  I've had this problem more often with interiors than exteriors and hope you can be a little more precise in your explanation of the process.  Perhaps a screen grab of the action would suffice.  Thanks for posting your progress and while the interpolation is apparent, I agree with you that it's commercially acceptable.  Jim

Marc, How about the first two examples.  Thanks, Jim
Logged

marc gerritsen
Sr. Member
****
Offline Offline

Posts: 299


WWW
« Reply #35 on: May 17, 2010, 04:41:09 PM »
ReplyReply

Quote from: haefnerphoto
Marc, How about the first two examples.  Thanks, Jim

hi jim
i uploaded a rudementary iphone movie of a photo i worked on
chose a different photo then the other examples as this one can really demonstrate that the "progressive stretch" actually works
i realized i can not use it very much in architecture but for interiors it works wonders
again this is only to describe the concept and anyone can apply their own ways to it
cheers
marc

https://download.yousendit.com/bFFNbGtESEJreEJFQlE9PQ


Logged

haefnerphoto
Sr. Member
****
Offline Offline

Posts: 613


WWW
« Reply #36 on: May 17, 2010, 07:07:31 PM »
ReplyReply

Thank you Marc, I got it now!  I appreciate you taking the time to illustrate the process.  Jim
Logged

rainer_v
Sr. Member
****
Offline Offline

Posts: 1120


WWW
« Reply #37 on: May 17, 2010, 11:28:43 PM »
ReplyReply

i am sorry, but i think this is not true.
as adamoerk already explained geometric proportions become completely false if this "progressive" stretch is applied. this does not look better , its just looking wrong and is not thought till the end.

of course windows at the top of a high building are more far away, and look smaller for this in "reality", but respecting this you should not apply any perspective correction. in any case you cant apply corrections to the vertical lines and not to the horizontals as well,- they have to be stretched if you want to maintain the same geometric forms in a building. and so function optical shift lenses  as well ( they dont squeeze the image at the top "progressive" together, in fact they enlarge the top for the more angular entry of the light rays ). if you correct electronically you have to do simply the same.
if you want to respect the smaller scale of the upper end of the building you have to use the uncorrected image, because optically this is the correct one.
our brain is a clever animal, and so it corrects perspectives subjectively because we know that buildings stand straight up and so the building looks for us more natural in an image if it is again standing straight in an image, therefor we use shift lenses ( or correct it in ps ).  but our bran does not "correct" vertical  lines and horizontal lines not, quadrats  still look as quadrats and dont transform to squares if they are more far away.
« Last Edit: May 18, 2010, 08:02:56 AM by rainer_v » Logged

rainer viertlböck
architecture photographer
munich / germany

www.tangential.de
marc gerritsen
Sr. Member
****
Offline Offline

Posts: 299


WWW
« Reply #38 on: May 17, 2010, 11:53:41 PM »
ReplyReply

Quote from: rainer_v
i am sorry, but i think this is complete bs. as adamoerk already explained geometric proportions become completely false if this "progressive" stretch is applied. this does not look better , its just looking wrong and is not thought till the end.

of course windows at the top of a high building are more far away, and look smaller for this in "reality", but respecting this you should not apply any perspective correction. in any case you cant apply corrections to the vertical lines and not to the horizontals as well,- they have to be stretched if you want to maintain the same geometric forms in a building. and so function optical shift lenses  as well ( they dont squeeze the image at the top "progressive" together, in fact they enlarge the top for the more angular entry of the light rays ). if you correct electronically you have to do simply the same.
if you want to respect the smaller scale of the upper end of the building you have to use the uncorrected image, because optically this is the correct one.
our brain is a clever animal, and so it corrects perspectives subjectively because we know that buildings stand straight up and so the building looks for us more natural in an image if it is again standing straight in an image, therefor we use shift lenses ( or correct it in ps ).  but our bran does not "correct" vertical  lines and horizontal lines not, quadrats  still look as quadrats and dont transform to squares if they are more far away. sorry that i cant find a better word than bs. for this theory. better you forget it.

sorry, but i don't like your tone!

i am just exploring different options and within an already busy schedule i take time out to share!
if you read the whole list of posts you will see that i have already agreed that this might not work very well for tall architectural buildings
and if you would have gone through all the posts then you might see that it actually works very well for interiors
if you would be so kind to show me with photos what you mean, i think that might be helpfull for all

cheers
m







Logged

rainer_v
Sr. Member
****
Offline Offline

Posts: 1120


WWW
« Reply #39 on: May 18, 2010, 08:01:58 AM »
ReplyReply

Quote from: marc gerritsen
sorry, but i don't like your tone!

i am just exploring different options and within an already busy schedule i take time out to share!
if you read the whole list of posts you will see that i have already agreed that this might not work very well for tall architectural buildings
and if you would have gone through all the posts then you might see that it actually works very well for interiors
if you would be so kind to show me with photos what you mean, i think that might be helpfull for all

cheers
m
you are certainly right marc, i wrote it with too fast mbp key pad .... so i unsharpened a bit my last comment because the post  was not intended to harm anyone here and i appreciate the effort to try out things and to share them too ... thats always great.

in interiors it may happen that with wideangles you get stretched edges, esp. if there are objects closer to the camera as chairs at the side or - more so - round objects which arent centered. in that case i personally copy a layer of the selected zone and try to make it with the ps distort function looking more homogeneously, but i think this should be done motif depending and individually for each motif. there cant be a rule which you apply to all photography, because- interior or exterior -  the interpretation of a static applied "progressive correction" results in wrong geometric renderings. there may be applications where this does not matter, but in architecture photography i believe it does.


« Last Edit: May 18, 2010, 08:16:01 AM by rainer_v » Logged

rainer viertlböck
architecture photographer
munich / germany

www.tangential.de
Pages: « 1 [2] 3 4 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad