SOLVED Propping up my floors

[Guest]

OK - my brain works quite slowly, I originally posted this problem as http://e-group.uk.net/forum/showthread.php?t=6593 - My Floors Sag.

I think I've now got a method which works quite well (though I've been looking at these pictures for so long my judgement may well be awry).

The problem is that perspective correction works on a matrix which has its centre at the exact middle of the image. This is where the software imagines the vanishing point to be. Unfortunately if you shoot from an angle the vanishing point is displaced to one side ( and up/down if you are not level). It is about this point that transformation matrix should work.

So take this image (hdr lens corrected through PT lens - but otherwise as shot)



Now make a second layer and draw lines to work out the actual vanishing point



This shows the locus as being just below the lh of the window - but ill defined as a result of perspective issues.

Now copy the original image (without lines)to a new image which is more than twice the size of the original, and anchor it so that so that the centre of the new image is just below the lh of the window.

Apply perspective correction to the new image - this now has its transformation matrix centred close to the actual vanishing point so introduces less distortion.

This gives



which if you draw on the lines again shows a nicely defined vanishing point



But

• Have I fooled myself - it looks right but am I just blinded by looking too long
  
• Does this make sense

And why didn't I think of it ages ago?

Nick

[Ian]

Re: Propping up my floors

I think that works well, Nick - I'll be borrowing your method in future!

Ian

PS My main problem with perspectiev correction (in PS at any rate) is that sometimes the height goes out of whack.

[Guest]

Re: Propping up my floors

I think that works well, Nick - I'll be borrowing your method in future!

Ian

PS My main problem with perspectiev correction (in PS at any rate) is that sometimes the height goes out of whack.

Thank you Ian - I hope that also means the explanation makes enough sense so that the method can be used (in whatever tool supports layers).

As to the height - well I've taken to applying half of each correction as stretch (pulling out of the image) and the other half as sqeeze (pushing in). But after a couple of minutes it's hard to remember what the actual proportions were (maybe a 2'nd screen with the original would help)

Nick

[Ellie]

Re: Propping up my floors

... and anchor it so that so that the centre of the new image is just below the lh of the window.

How do you do that?

[Guest]

Re: Propping up my floors

How do you do that?

In The GIMP (if you are still using that) - the layers dialogue and workbench have (respectively) the word 'anchor' or a little anchor sign.

So after you have opened a new file of say 7000/5000 pixels. You copy and paste from the original - which becomes a 'floating layer' centred on the middle of the new file. The tools workbench gives you a little cross with arrows on the arms - this enables you to move the 'floating layer' till you've aligned it where you think the vanishing point is over the centre of the new image.

In the words of Soap - 'Confused - you will be'

Hope this helps

Nick

[Ellie]

Re: Propping up my floors

'Confused - you will be'

Errm, if you anchor it, how do you move it afterwards?

Sorry, I'm probably being very dim. Blame it on the weather!

[Guest]

Re: Propping up my floors

Errm, if you anchor it, how do you move it afterwards?

Sorry, I'm probably being very dim. Blame it on the weather!

You move it first - then anchor it - so you read my comments backwards (it's all about perspective)

Sorry

Nick

[Zuiko]

Re: Propping up my floors

Errm, if you anchor it, how do you move it afterwards?

Sorry, I'm probably being very dim. Blame it on the weather!

Dont worry, Ellie, it's gone right over my head too, but I don't even know how to make another layer! Think I'll stick to landscapes, where perspective doesn't need to be quite so accurate.

[Guest]

Re: Propping up my floors

Dont worry, Ellie, it's gone right over my head too, but I don't even know how to make another layer! Think I'll stick to landscapes, where perspective doesn't need to be quite so accurate.

It's Layer - new, on The GIMP menu system from the screen where the picture is shown. Of course it might be a bit more complicated on paid for software.

Nick

[snaarman]

Re: Propping up my floors

I probably don't understand the method, but I should get inside my copy of Gimp and learn more..

Mainly though I don't understand the result.. Were we trying to make the handrail horizontal as well as retaining a rectilinier look to the rest of the scene?? If not, then where was the offending sag?




Pete

[Guest]

Re: Propping up my floors

I probably don't understand the method, but I should get inside my copy of Gimp and learn more..

Mainly though I don't understand the result.. Were we trying to make the handrail horizontal as well as retaining a rectilinier look to the rest of the scene?? If not, then where was the offending sag?




Pete

OK - invoking a Sondheim presentation



The disparity in angle between handrail and the alter/east window is obviously wrong. The floor on the lhs of the original slopes downwards towards the viewer, similarly the wall is too tall. The roof line around the windows is out, with the roof on the rhs far too low.

Now I may not have completely solved each of these problems (indeed the eye seems to want some perspective distortion in a photograph) but I think they are reduced.

But perhaps I'm fooling myself - or perhaps there is a better representation.

Nick

[Zuiko]

Re: Propping up my floors

It's definitely a big improvement, Nick, albeit at a slightly reduced angle of view.

[Guest]

Re: Propping up my floors

It's definitely a big improvement, Nick, albeit at a slightly reduced angle of view.

Thank you John.

Actually I'm not sure about the angle of view - effectively we originally had two - one portrayed by the elements that went across the scene at an angle (and were therefore most subject to distortion) and one by the walls that went into the scene. The walls seem fairly close in both - it's the cross elements that have changed.

I think I need a tame maths teacher (or an old style physicist), some graph paper and a pair of compasses.

Nick

[snaarman]

Re: Propping up my floors

OK, we are barking up the same tree.

If you had stood in the centre of the chancel facing the altar, then the walls left and right will recede towards a common vanishing point dead centre, straight ahead. If the lens is a good 'un (and they used good builders) then you can draw straight lines towards this vanishing point and all is well.

The laterals (handrail etc) goleft to right parallel to the image frame.

However... If you stand to one side (lets say to the left side as per the pictures), then you invoke two sets of vanishing points: The side walls will converge on the one ahead and now to the left a bit. The laterals will converge on a second vanishing point well out of shot to the right. I think that is what you are seeing in the first shot..

I shall now sit down with a nice cup of tea.

Pete

edit..

Rather like this fine example from the world of crosstich (!) http://www.needlepoint.org/Archives/...tive/index.php

[Guest]

Re: Propping up my floors

Well yes - but added in we have the distortion caused by the fact the lens is not parallel with those elements going across the screen - they are in the same state that if we partially emulate the diggers and turn the world on its side - a tower would be if we were to lean back to include both the top and the bottom.

So in the original the 2'nd vanishing point is ill defined to one side (and in reality far too close). In the modified image well..........



we have 2 effects - first I left a little roofline distortion - because it looked closer to right to me (at the time) - secondly the actual vanishing point moves further from the image. So we have the crossing lines caused by the 'left in' roofline distortion and the converging lines for the lower half tending towards the vanishing point.

Now it should be possible to calculate (given the distance of the nearest/furthest object and the angle of the viewer compared to the rail/far wall) where the vanishishing point is exactly using simple trig functions, but I'm getting old and am no longer quite the keen schola buriensis that Edward the VI intended me to be.

Interestingly the angle of the lens to the far wall accounts for Johns comments that the angle of view has altered - because the same effect that makes towers lean has moved the corner of the walls (LH rear) beyond 90%, so if it's reconciled back to 90% it appears a flatter presentation.

Nick