New Mandate: 46.8 Degrees; due Januray 17th, 2010
in [Digital Cameras]
|
Prev: [SI] New Mandate: 46.8 Degrees; due Januray 17th, 2010
Next: Lens with Depth of Field indicator. (For full frame [36x24mm] digital camera.)
From: Ralph Tolden on 14 Dec 2009 17:06 On Tue, 15 Dec 2009 06:17:54 +1000, Troy Piggins <usenet-0912(a)piggo.com> wrote: >* tony cooper wrote : >> On Mon, 14 Dec 2009 08:02:08 -0500, "Bowser" <its(a)bowzah.ukme> wrote: >> >> [---=| Quote block shrinked by t-prot: 11 lines snipped |=---] >>> >>>BS. You're just pissed because Tiger wasn't sharing with you, after all >>>you've done for him. Lock that zoom near 50, shoot something, and send it >>>in. No excuses. Golf is done forever. Elin isn't wearing her wedding ring >>>any more, and the universe is upside down. So go shoot. >> >> Tiger, as you might know, lives in this area. Not my neighborhood, >> but in the area. I'd need a wide angle lens, or maybe have to do a >> panorama, to take a shot of the media trucks parked outside of the >> gated community in which he lives. > >If you get a shot of Tiger with one of his mistresses using a >50mm lens, you win the mandate. It's not a "Tiger", but how about this 15-footer with a 50mm from about 2-ft away? (full frame, no crop) http://farm3.static.flickr.com/2611/4186064984_1971e17ae5_m.jpg (No, you don't get to see a larger one, copyright protection.)
From: Paul Furman on 14 Dec 2009 22:51 Paul Furman wrote: > 28mm could be cropped. Anything wider than 31mm or so could be cropped and if significantly wider, the mandate could change to a perspective correction/shift lens challenge. For a typical scenario tilted up to view a tall building, shoot wider and tilt down so that verticals are vertical. Now crop off the bottom of the shot and you have what a tilt lens provides. It's simple, just keep the verticals vertical in the viewfinder with a wide lens and crop the un-wanted part out later. The result can look very wide with lots of wide angle distortion (or not) and still meet the 46.8 degree mandate requirements. -- Paul Furman www.edgehill.net www.baynatives.com all google groups messages filtered due to spam
From: Calvin Sambrook on 15 Dec 2009 12:31 "JimKramer" <newsreader1NOFSPAM(a)jlkramer.net> wrote in message news:hg8bg0$87i$1(a)news.eternal-september.org... > "Calvin Sambrook" <csambrook(a)bigfoot.com> wrote in message > news:hg8238$gvr$1(a)news.eternal-september.org... >> Sorry to keep banging on about this but the more I look at it the more I >> think I'm correct. I dusted off some textbooks and reminded myself how >> painful the equations look then realised things have moved on and >> someone, somewhere has probably done the work for me by now. Sure enough >> http://www.dofmaster.com/dofjs.html is a handy (!) DoF calculator and >> dropping some real-world numbers in gives: >> >> Sony a200, 18mm lens at F5.6, focused at 3m gives a DoF from 1.47m to >> infinity. That is, a subject at say 3m away will be in focus as will >> everything from 1.47m in front of the camera to the horizon. >> >> Sony a200, 300mm lens at F5.6, focused at 3m gives a DoF from 2.99m to >> 3.01m. That is, a subject at say 3m away will be in focus but almost >> nothing else will. The horizon will be well out of focus. >> >> That feels like it fits with experience too and it says that for this >> mandate cropping to an equivalent angle of view of a 50mm lens is not >> correct and will produce visually different results. >> > > You are still mistaken: > > Now do the math to figure out what an equivalent FoV to crop the 18mm down > to 300mm. Now go back to Dofmaster and recalculate it with the new CoC > for the same print size, you may assume that we have infinite resolution > for this exercise. Do the numbers suddenly look the same, or very nearly > so? > > The problem with the cute little online calculators is that there are a > number of assumptions that are taken for granted by you, because the math > going on behind the calculations is not transparent to the casual end > user. > > Go out with a zoom lens on a tripod, focus on a single subject and prove > it to yourself, then come back and tell us what happened. > I don't follow your logic there. The Circle of Confusion is fixed at the point you capture the image. The rays are recorded at that point as confused, that is to say it is no longer possible to tell where they came from, you only have the recording of them. [1]. Resolution doesn't matter, it can be as infinite as you like :-) but the confused rays will still be confused, you'll just be able to see them in lots of detail. You can't just re-scale the CoC for convenience. Imagine the CoC of a particular lens was 1 unit wide and the whole image was 1,000,000 units wide. If we now crop and enlarge a portion of the image, let's say we zoom 1,000 times then we'd be looking at an image formed from a part just 1,000 of the original units wide and the confused circle, originally 1 unit wide and just 1 millionth of the image, would now be taking up 1,000th of the image. It doesn't get any sharper (less confused), just bigger. I agree that you need to be careful with online calculators but I was just using one to avoid having to do the sums by hand. I can't see anything in the original equations which would alter the thrust of the argument. [1] OK, there are techniques which sharpen images based on knowledge of the likely cause of the confusion (ie. the lens characteristics) but that's cheating.
From: JimKramer on 15 Dec 2009 13:48 "Calvin Sambrook" <csambrook(a)bigfoot.com> wrote in message news:hg8h3h$tk4$1(a)news.eternal-september.org... > "JimKramer" <newsreader1NOFSPAM(a)jlkramer.net> wrote in message > news:hg8bg0$87i$1(a)news.eternal-september.org... >> "Calvin Sambrook" <csambrook(a)bigfoot.com> wrote in message >> news:hg8238$gvr$1(a)news.eternal-september.org... >>> Sorry to keep banging on about this but the more I look at it the more I >>> think I'm correct. I dusted off some textbooks and reminded myself how >>> painful the equations look then realised things have moved on and >>> someone, somewhere has probably done the work for me by now. Sure >>> enough http://www.dofmaster.com/dofjs.html is a handy (!) DoF calculator >>> and dropping some real-world numbers in gives: >>> >>> Sony a200, 18mm lens at F5.6, focused at 3m gives a DoF from 1.47m to >>> infinity. That is, a subject at say 3m away will be in focus as will >>> everything from 1.47m in front of the camera to the horizon. >>> >>> Sony a200, 300mm lens at F5.6, focused at 3m gives a DoF from 2.99m to >>> 3.01m. That is, a subject at say 3m away will be in focus but almost >>> nothing else will. The horizon will be well out of focus. >>> >>> That feels like it fits with experience too and it says that for this >>> mandate cropping to an equivalent angle of view of a 50mm lens is not >>> correct and will produce visually different results. >>> >> >> You are still mistaken: >> >> Now do the math to figure out what an equivalent FoV to crop the 18mm >> down to 300mm. Now go back to Dofmaster and recalculate it with the new >> CoC for the same print size, you may assume that we have infinite >> resolution for this exercise. Do the numbers suddenly look the same, or >> very nearly so? >> >> The problem with the cute little online calculators is that there are a >> number of assumptions that are taken for granted by you, because the math >> going on behind the calculations is not transparent to the casual end >> user. >> >> Go out with a zoom lens on a tripod, focus on a single subject and prove >> it to yourself, then come back and tell us what happened. >> > > I don't follow your logic there. The Circle of Confusion is fixed at the > point you capture the image. The rays are recorded at that point as > confused, that is to say it is no longer possible to tell where they came > from, you only have the recording of them. [1]. Resolution doesn't > matter, it can be as infinite as you like :-) but the confused rays will > still be confused, you'll just be able to see them in lots of detail. > > You can't just re-scale the CoC for convenience. Imagine the CoC of a > particular lens was 1 unit wide and the whole image was 1,000,000 units > wide. If we now crop and enlarge a portion of the image, let's say we > zoom 1,000 times then we'd be looking at an image formed from a part just > 1,000 of the original units wide and the confused circle, originally 1 > unit wide and just 1 millionth of the image, would now be taking up > 1,000th of the image. It doesn't get any sharper (less confused), just > bigger. > > I agree that you need to be careful with online calculators but I was just > using one to avoid having to do the sums by hand. I can't see anything in > the original equations which would alter the thrust of the argument. > > [1] OK, there are techniques which sharpen images based on knowledge of > the likely cause of the confusion (ie. the lens characteristics) but > that's cheating. > CoC is for determining acceptable DoF for a print not for the sensor. You're going at backwards again. We're not measuring the overall effectiveness of the lens or sensor; thus infinite resolution for the sensor, and a perfect set of lenses. We are trying to make it clear that for a given position, focus point and f stop that the image formed does not change other than the over all FoV; i.e. you can crop a wide angle image to the same FoV as a longer focal length lens image and it will be identical to the image taken with the longer focal length lens. You are having a difficult time grasping the math and are making it overly complicated, so again go out and take some pictures to prove it to yourself. Any variations from this are caused by real world defects in the lens or the image capture device. -Jim
From: Troy Piggins on 15 Dec 2009 15:10
* JimKramer wrote : > "Calvin Sambrook" <csambrook(a)bigfoot.com> wrote in message > news:hg8h3h$tk4$1(a)news.eternal-september.org... > [---=| Quote block shrinked by t-prot: 60 lines snipped |=---] >> >> [1] OK, there are techniques which sharpen images based on knowledge of >> the likely cause of the confusion (ie. the lens characteristics) but >> that's cheating. > > CoC is for determining acceptable DoF for a print not for the sensor. You're > going at backwards again. We're not measuring the overall effectiveness of > the lens or sensor; thus infinite resolution for the sensor, and a perfect > set of lenses. > > We are trying to make it clear that for a given position, focus point and f > stop that the image formed does not change other than the over all FoV; i.e. > you can crop a wide angle image to the same FoV as a longer focal length > lens image and it will be identical to the image taken with the longer focal > length lens. You are having a difficult time grasping the math and are > making it overly complicated, so again go out and take some pictures to > prove it to yourself. Any variations from this are caused by real world > defects in the lens or the image capture device. Sorry Jim. Standing in the same spot, focusing on the same object, using the same aperture and same camera, with 2 different focal length lenses will yield 2 different depths of field. The shot taken with the shorter focal length lens will have more stuff in focus than the longer focal length lens shot. -- Troy Piggins |