Yet Another Cycling Forum
Random Musings => Gallery => OT Gallery => Topic started by: andygates on 10 October, 2008, 11:48:12 pm
-
I do like a new trick...
(http://farm4.static.flickr.com/3035/2930411000_6f36393b77_m.jpg) (http://www.flickr.com/photos/andygates/2930411000/)
Tilt-shift lenses are heinously expensive. The effect - which involves getting a weird depth of field by tilting the lens so that the plane of focus isn't parallel with the film/sensor - is kinda dreamy and hallucinatory, and makes everything look like it's a shot from Staggeringly Realistic Model Maker Monthly's "trippy cities" special.
-
Ben Kinetics (http://www.flickr.com/photos/cycleologist/sets/72157600200535618/) has been having a go with tiltshift photography too.
I especially like the one he did of two toy trains (http://www.flickr.com/photos/cycleologist/2340785771/sizes/l/in/set-72157600200535618/).
-
Schweet! :thumbsup:
-
There's a really good cycling one here (http://cache.boston.com/universal/site_graphics/blogs/bigpicture/tdf_07_28/tdf9.jpg).
-
I love those.
-
There's a fantastic movie of Sydney harbour here: http://vimeo.com/1831024?pg=embed&sec=1831024&hd=1 (http://vimeo.com/1831024?pg=embed&sec=1831024&hd=1).
I've never quite understood what the lenses are supposed to be for. Are they actually for doing the opposite of fake miniatures? -- i.e. increasing depth of field.
-
I've never quite understood what the lenses are supposed to be for. Are they actually for doing the opposite of fake miniatures? -- i.e. increasing depth of field.
Tilt changes the plane of the depth of field so that objects at different distances are in focus at the same time, shift changes the plane through the centres of the lenses, whilst keeping them parallel.
When (plate | rollfilm) cameras had bellows, you could sometimes shift the objective lens with respect to the film plane, usually to adjust a wide-angle shot (say, a building) so that the verticals remained straight. 35mm SLRs lacked that capability, and building shots on wideangle lenses invariably ended up with converging verticals due to the camera not being parallel to the building sides. Eventually someone thought to make a lens that allowed you to achieve similar shots to the bellows cameras.
I've never owned one, but I did once run a rollfilm through an old Coronet (http://flickr.com/photos/crayzy_ray/2065927354/) bellows camera, and the effects whilst not the same, are somewhat similar - the lens is of relatively poor quality, and the aberrations mean that the focus falls off towards the edges anyway. Unfortunately, I haven't got those pictures in digital format (have you seen the price of Medium Format scanners? :o )
-
The reason the pictures look like models is that we associate shallow depth of field with close-up shots. These lenses (or a 5 x 4 camera) can affect depth of field, giving unexpected results.
Their most common use is to avoid converging verticals in architecture shots.
-
Tilt shift is standard practice in large format photography. The trick is to get the plane of the film, theplane of the lens and the plane you want in focus all to converge at the same point. That is the tilt part and you cna use it to gain a selective focus (as seen in the images posted) or to increase depth of field.
The shift part affects the perspective. Coming back to the parallel bit, if you place your film parallel to the object being photographed, and the lens at a point where it can cast an image on the film of the thing you want to photograph, then the image will have the proper perspective. The lens is shifted with respect to the film.
Lots of fun, but quite pricey to run a large format camera.
..d
-
Lots of fun, but quite pricey to run a large format camera.
How to use Photoshop's Lens Blur tool for tilt-shift fakery (http://www.tuaw.com/2008/10/07/how-to-use-photoshops-lens-blur-tool-for-tilt-shift-fakery-par/)
-
An informative thread, and some interesting pictures!
Those pictures do look amazingly like the photos my brother used to have in his Model Railway magazine!
(Phew, luckily he got them when I was a kid, so I don't have to admit to having ever bought any! ;D
With modern digital cameras and digital manipulation it's realtively trivial to straighten up the sides of buildings, although it won't solve depth of field issues, so I suppose you have to make sure you use a suitably small aperture and long exposure in the first place.
-
Lots of fun, but quite pricey to run a large format camera.
How to use Photoshop's Lens Blur tool for tilt-shift fakery (http://www.tuaw.com/2008/10/07/how-to-use-photoshops-lens-blur-tool-for-tilt-shift-fakery-par/)
That only works one way. Using tilt shift to get great depth of field in a model pic is not possible with photoshop.
..d
-
Well, yeeees, but that's kinda obvious: you can't make something from nothing. I can no more add depth of field than I can add stuff that was off-shot!
You can make stylish nothing from something, though. :thumbsup:
-
Even The Time is in to it now!
How to make Wembley look like a Subbuteo pitch | International Football - Times Online (http://www.timesonline.co.uk/tol/sport/football/international/article4932164.ece)
The printed version is way better than this online one.
-
Lots of fun, but quite pricey to run a large format camera.
How to use Photoshop's Lens Blur tool for tilt-shift fakery (http://www.tuaw.com/2008/10/07/how-to-use-photoshops-lens-blur-tool-for-tilt-shift-fakery-par/)
That only works one way. Using tilt shift to get great depth of field in a model pic is not possible with photoshop.
..d
It's not anything to do with depth of field, Any sense of perspective we have depends on a natural vanishing point, lenses of a different focal length alter this, but if we have obvious cues such as railway tracks and these behave as if the object is a lot smaller then we will perceive them as such. Stretch the top of the picture and that will move the vanishing point.
Damon.
-
I thought this photo (http://static.guim.co.uk/Guardian/sport/gallery/2008/aug/27/osm.olympics/olympics_100m_jf-5149.jpg) from the Guardian looked like a model I didn't know how the affect had been achieved. I assume it's another photo using a the tilt-shift effect?
-
Yes indeedy. Nice one, too.
-
I thought this photo (http://static.guim.co.uk/Guardian/sport/gallery/2008/aug/27/osm.olympics/olympics_100m_jf-5149.jpg) from the Guardian looked like a model I didn't know how the affect had been achieved. I assume it's another photo using a the tilt-shift effect?
It shows how it was manipulated by the way the white lines between the lanes widen as they recede into the background.
Damon.
-
That's your theory, but the perspective on the "Subbuteo" shot looks pretty normal to me.
Something else these pictures seem to share is pumped-up colour saturation, which increases the unworldly effect.
-
I thought this photo (http://static.guim.co.uk/Guardian/sport/gallery/2008/aug/27/osm.olympics/olympics_100m_jf-5149.jpg) from the Guardian looked like a model I didn't know how the affect had been achieved. I assume it's another photo using a the tilt-shift effect?
It shows how it was manipulated by the way the white lines between the lanes widen as they recede into the background.
Damon.
That could just be the circle of confusion widening.
The reason it looks like a model shot is that we expect a lens with that angle of view to be pretty much at infinity (or hyperfocal anyway) so we expect the whole scene in focus. But we know that close ups of the same perspective have severe depth of field limitations.
So we translate it to being a closeup.
..d
-
Something else these pictures seem to share is pumped-up colour saturation, which increases the unworldly effect.
Indeedy: that and the hard contrast make it look like it's been painted with bright model paints and illuminated with a spotlight.
-
Something else these pictures seem to share is pumped-up colour saturation, which increases the unworldly effect.
Indeedy: that and the hard contrast make it look like it's been painted with bright model paints and illuminated with a spotlight.
Ah. I knew Jordan would appear in this thread at some point.
-
The reason it looks like a model shot is that we expect a lens with that angle of view to be pretty much at infinity (or hyperfocal anyway) so we expect the whole scene in focus. But we know that close ups of the same perspective have severe depth of field limitations.
So we translate it to being a closeup.
..d
That's the intriguing bit really, do we have to know what models look like and is this interpretation learned. It occured to me that the Gerry Anderson puppet series models would probably have been shot with tilt shift lenses as they used 35mm cameras which would have been difficult to get down onto the same plane as the dioramas they used, so they would have shot from above and corrected with tilt shift. The same was prabably true of railway modelling pictures. Is this effect obvious to someone unfamiliar with the technique applied to models.
Damon.
-
The reason it looks like a model shot is that we expect a lens with that angle of view to be pretty much at infinity (or hyperfocal anyway) so we expect the whole scene in focus. But we know that close ups of the same perspective have severe depth of field limitations.
So we translate it to being a closeup.
..d
That's the intriguing bit really, do we have to know what models look like and is this interpretation learned. It occured to me that the Gerry Anderson puppet series models would probably have been shot with tilt shift lenses as they used 35mm cameras which would have been difficult to get down onto the same plane as the dioramas they used, so they would have shot from above and corrected with tilt shift. The same was prabably true of railway modelling pictures. Is this effect obvious to someone unfamiliar with the technique applied to models.
Damon.
It is obvious becasue that is how our eyes work (irrespective of the perspective correction). Look at something close up and the background becomes a blur. Look at something large but far away and the background is in focus. (Father Ted moment).
..d
-
I've been exploring Depth of Field in relation to video cameras recently, The small size of the CCDs means that video cameras have a greater depth of field relative to a 35 mm camera. Various adapters are used Depth-of-field adapter - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Depth-of-field_adapter) to give the impression of 35mm film. Plate cameras have an even poorer depth of field and of course the type of telephoto lens with a large apertures used by sports photographers have a very tight depth of field. The video upthread was probably taken with a depth of field adapter with tilt shift, something which is fairly easy to acheive. The main use of tilt shift is to correct converging verticals in architectural photography, it is also suitable for correcting perspective. I'm wondering where we learned the visual conventions which lead us to perceive these shots and the photoshop manipulations as being similar to models. Is it because this is how models are shot?
Damon.
-
A regular lens has a narrower depth of field close-up, and a macro lens even more so.
I need some brightly-coloured unrealistic people shots. Jordan and Christine Lydon will do... 8)
-
Depth of field is a subject which is less understood by the current generation as they haven't got those lovely engraved lines on the lens like we had in the days of Zorkis and Feds. Compact digitals have small CCDs so their depth of field is pretty deep.
Damon.
-
Tilt and shift are two different things.
Tilt alters the angle of the focal plane, allowing the plane of focus to be along a non-parallel plane relative to the film/sensor. see Scheimpflug principle - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Scheimpflug_principle)
Shift is used to alter perspective, using a different part of the image circle and is limited by the lens. Most 35mm lenses should shift fine (to a reasonable degree) if used on a DX sized sensor. Even better if on a smaller sensor.
..d
-
*missing my Zorki* :(
-
Tilt and shift are two different things.
Tilt alters the angle of the focal plane, allowing the plane of focus to be along a non-parallel plane relative to the film/sensor. see Scheimpflug principle - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Scheimpflug_principle)
..d
That's interesting David, the origins of tilt in correcting perspective in aerial photographs is the part which makes me think that there may be some connection with the general techniques used in model photography. They are usually seen from an aerial perspective and may well be subject to correction. Certainly the examples upthread which work best are seen from above.
*missing my Zorki* :(
I've got two Feds, one with a focus ring marked in feet, the addition of a nurled ring for the light meter almost pushes it into engraved marking overload.
Damon.
-
Out of curiosity I set out to find out how to film models so they look life size, the answer would seem to be the Frazier Lens system.
Frazier Lens (http://media.panavision.com/ReferenceLibrary/html/Frazier.html)
Damon.
-
It's pretty usual to shoot at f/64 on large format for even moderate close-ups (not macro stuff, just dead leaves and stuff). This doesn't work with 35mm because the diffraction with such a tiny aperture is appalling - most 35mm lenses aren't that good at even f/22.
-
Out of curiosity I set out to find out how to film models so they look life size, the answer would seem to be the Frazier Lens system.
Frazier Lens (http://media.panavision.com/ReferenceLibrary/html/Frazier.html)
Damon.
You could take lots of exposures with the focus set at different distances into the scene and stack them. I'm sure I've seen a tutorial how to do it somewhere.
-
Out of curiosity I set out to find out how to film models so they look life size, the answer would seem to be the Frazier Lens system.
Frazier Lens (http://media.panavision.com/ReferenceLibrary/html/Frazier.html)
Damon.
You could take lots of exposures with the focus set at different distances into the scene and stack them. I'm sure I've seen a tutorial how to do it somewhere.
That is an interesting idea. A similar process is used in fluorsecence microscopy where a process called deconvolution is used to remove the blur introduced from out of focus image planes.
It shouldn't be too hard to write a filter that would produce a greyscale image based on contrast between adjacent pixel groups to make relevant areas of each layer transparent. Which is a bit cheaper in terms of computation and doesn't rely on knowing what the other images in teh stack are like.
..d
-
There was some kind of technique reported a year or so ago where a photo on a computer had all the focus plains in it and you could move through them, bringing different parts into focus. It may have been a camera rather than a technique. I'm trying to find it!
-
OK Here's my first go(http://i127.photobucket.com/albums/p148/AdrianC_No1/Misc/MilleniumBridgeTS.jpg)
-
That is a hard one as you have conflicting planes -- you want a soft background to the bridge in sharp focus. As it is, it just looks wrong.
It would be easier when the plane of focus should extend the whole way across the image.
..d
-
Here' another great example of tilt-shift photography to give that 'minature model' appearance
(http://www.bacam.demon.co.uk/ekmrc/Graphics/cbridge1.jpg)
-
::-)
Does Photoshop have a "turn the water into varnish" filter?
;)
-
Not convincing - shouldn't there be a piece of cotton wool coming out of the funnel?
-
Behold! Web toy!
tiltshiftmaker.com - Transform your photos into tilt-shift miniatures (http://tiltshiftmaker.com/)
-
Mini-Rome (http://i46.photobucket.com/albums/f130/PaulRide/DSCN3641-tiltshift1.jpg) and Mini-somewhere else. (http://i46.photobucket.com/albums/f130/PaulRide/tiltyrye.jpg)
Pointless but fun.
-
Mini-Rome (http://i46.photobucket.com/albums/f130/PaulRide/DSCN3641-tiltshift1.jpg)
Pointless but fun.
That just looks wrong. Why would you get a focal point change half way along a surface parallel to the film plane?
this is a case of people finding a nifty tool without realising what a genuine result should look like.
..d
-
Mini-Rome (http://i46.photobucket.com/albums/f130/PaulRide/DSCN3641-tiltshift1.jpg)
Pointless but fun.
That just looks wrong. Why would you get a focal point change half way along a surface parallel to the film plane?
this is a case of people finding a nifty tool without realising what a genuine result should look like.
..d
Sorry :-[
I'll see if I can find a better victim for my next experiment. I fear the main problem is that I've never composed holiday snaps with this trick in mind.
-
Mini-Rome (http://i46.photobucket.com/albums/f130/PaulRide/DSCN3641-tiltshift1.jpg)
Pointless but fun.
That just looks wrong. Why would you get a focal point change half way along a surface parallel to the film plane?
this is a case of people finding a nifty tool without realising what a genuine result should look like.
..d
What you may want to do is to mask out the buildings in photoshop/gimp/epoc, perform the t/s munging, then drop back in the buildings with the appropriate amount of blur. Obviously this is a lot more work than a simple point and click program but would give better results aesthetically.
..d
Sorry :-[
I'll see if I can find a better victim for my next experiment. I fear the main problem is that I've never composed holiday snaps with this trick in mind.
-
I know I'm cheating a bit, but some real miniatures from this morning:
(http://i8.photobucket.com/albums/a19/mikes99mail/88floorhotel.jpg)
from the top if here:
(http://i8.photobucket.com/albums/a19/mikes99mail/bottleopener.jpg)
Mad place. Absolutely barking.
-
Spotted this one:
http://www.boredpanda.com/diy-tilt-shift-photography-broken-lens/
-
I do like a new trick...
(http://farm4.static.flickr.com/3035/2930411000_6f36393b77_m.jpg) (http://www.flickr.com/photos/andygates/2930411000/)
Tilt-shift lenses are heinously expensive. The effect - which involves getting a weird depth of field by tilting the lens so that the plane of focus isn't parallel with the film/sensor - is kinda dreamy and hallucinatory, and makes everything look like it's a shot from Staggeringly Realistic Model Maker Monthly's "trippy cities" special.
You can get a Lensbaby Muse for about £80
-
On this one, some are more successful than others:
http://www.smashingmagazine.com/2008/11/16/beautiful-examples-of-tilt-shift-photography/
-
If anyone is interested in learning more about this subject then Harold Merklinger's work is essential reading. Please do donate to his cause if you find the book linked below useful..
http://www.trenholm.org/hmmerk/FVC161.pdf
Before kids arrived I was a bit of a large format devotee, moderated the LF forum on photo.net and taught on a couple of large format residential courses. I do miss my Ebony but a digital compact is far better at taking pictures of the boys!
-
On this one, some are more successful than others:
http://www.smashingmagazine.com/2008/11/16/beautiful-examples-of-tilt-shift-photography/
Started looking at those and got bored after the second one.
-
I know what you mean. It only seems to work for certain subjects, and more especially trains and environs, where the brain can readily accept the concept of train sets and make that leap.
-
This is a short video I made from a £70 Canon camera.
After several attempts it's clear that the optimal POV is from high above, as if you were looking down on your toy village.
https://www.youtube.com/watch?v=kBUyb6ZsSVY (https://www.youtube.com/watch?v=kBUyb6ZsSVY)
Since the OP it seems a fairly standard feature on most cameras...but with about 10 minutes appeal. Once you've used it once, and had a chuckle, I can't imagine most people using it many times more.
-
I can't imagine most people using it many times more.
Apart from the BBC. ::-)