andy_mclean Posted September 17, 2006 Share #1 Posted September 17, 2006 Advertisement (gone after registration) I think I know the answer to the question I'm about to pose, but I'm going to ask it anyway.. Will the DoF scales of new and 'old' M series lenses work for two different sensor (therefore image) sizes? I thought you got greater DoF for a smaller image size (on the sensor) than a larger one for a given focal length and f-stop... in which case the DoF with the M8 will not be the same as for say an M6 for a given focal length and f-stop. Or is the DoF only dependent on focal length, focus distance and f-stop? Best wishes- Andy Link to post Share on other sites More sharing options...
Advertisement Posted September 17, 2006 Posted September 17, 2006 Hi andy_mclean, Take a look here M8 and DoF. I'm sure you'll find what you were looking for!
rosuna Posted September 17, 2006 Share #2 Posted September 17, 2006 Read this: http://photo.net/learn/optics/dofdigital/ ...or... http://www.bobatkins.com/photography/technical/digitaldof.html In short: DoF scales change if the format size changes, due to the modification of the circle of confusion... Link to post Share on other sites More sharing options...
andy_mclean Posted September 17, 2006 Author Share #3 Posted September 17, 2006 Thanks for your reply... but you haven't really answered my question...and the links don't really either. If the smaller sensor is located the same distance away from the lense as the film plane, then you can consider the digital image a crop of the larger film image... and the DoF will be the same for a given fstop and focus distance, the FoV will be different. If the lense to film/sensor distance is altered, then the DoF markings should be different for the two sensor sizes...no? There is a format change between film and the M8, so you suggest that the DoF is altered and therefore different lense markings are needed, no? This depends on the whole image being compressed to a smaller frame.... but that isn't happening in the M8 is it? Does the digital sensor lie in the same plane in the M8 as film does in the other Ms? Andy Link to post Share on other sites More sharing options...
adan Posted September 17, 2006 Share #4 Posted September 17, 2006 The very short answer - assume about 1 stop LESS DoF using any given lens (NOT field of view - LENS) on the M8 over the M7. I.E. shooting at f/11, use the DOF scale marks for f/8. The longer answer: On a smaller sensor you get more DoF for THE SAME FIELD OF VIEW - NOT for the same focal length. Shoot with a 21 on the M8, and a 28 on the M7 - and both images will have the same framing and FoV, and the M8 shot will have MORE DoF - but they were NOT "shot with the same lens", since 21mm does not equal 28mm. DoF is dependent on final enlargement - the more enlargement, the more slightly blurred things become obviously blurred. to get (for the sake of simple math) a 24 x 36 cm (10 x 15") print from an M7 image shot with a 28mm lens, one would enlarge it 10x. To get the same final print with the same field of view from an M8, one would shoot with a 21, and enlarge the 27x18mm image from the sensor by 13.33x. The extra 33% enlargement will make additional marginal blur visible, so there will be less DoF from the M8 than one would expect from a 21mm lens (thus the scales no longer apply perfectly) - but more DoF than one would expect (given the framing) from a 28mm lens. In other words, ther are two things workin in opposition to each other - more enlargement reduces apparent DoF, while shooting with a shorter focal length to maintain similar framing increases DoF. The net result means a shot made with a 21 on the M8 will have DoF somewhere in between shots from an M7 using a 21 or a 28. I don't expect the DoF scale to change from 'old' to 'new coded' lenses - the new lenses are digital-capable, not digital-only. Film users will still expect DOF scales applying to 24 x 36 format. Leica will likely include DoF tables in the M8 instruction book (or on-line). Or they may just say the same thing I said in the short answer at the beginning. - or one can calculate one's own tables using the kind of info Ruben provided. Link to post Share on other sites More sharing options...
grober Posted September 17, 2006 Share #5 Posted September 17, 2006 Interesting discussion, but to remind us all of the obvioius: the M8 is a digital camera. (Checking the DOF on the screen after the exposure sure beats waiting for the photos to come back from the lab.) If y'all don't like the focus in a particular photo, change it and take another shot. Link to post Share on other sites More sharing options...
TimF Posted September 17, 2006 Share #6 Posted September 17, 2006 The depth of field will be as for the true focal length of the lens, as that doesn't change when mounted on a cropped sensor digital camera. All that does is the angle of view you are seeing. This is the reason digicams have almost unlimited DoF even at apertures of around f/2 or f/2.8, because the true focal length of the lens ranges from approx 7mm to about 28mm or so (depending on zoom ratio of course). As for the circle of confusion, I wonder whether that actually applies in the same way as it does with film. The CoC is one factor, along with aperture setting, print size and viewing distance that governs depth of field in an enlargement. A digital camera however, has a native maximum print size that is dependant on the ppi setting decided by the photographer in post-processing. In which case, would it be true to say that CoC etc only comes into play when enlarging beyond that, ie, ressing-up, or at the least when making a print at a resolution below that considered the norm (240-300pppi)? Link to post Share on other sites More sharing options...
andy_mclean Posted September 17, 2006 Author Share #7 Posted September 17, 2006 Advertisement (gone after registration) Thanks for the replies guys. What led to this is that I've been shooting with an E-1 these last couple of years and I am getting more and more irritated by the extended DoF for a given FoV particularly in the 28 to 50mm equivalent range. I miss the shallow DoF I was getting at f1.4 and f2 with my M and also missed the RF overall approach. The M8 looks very, very attractive... yet this topic hadn't cropped up as far as I could see. I thought of it along the lines of.... a 28mm lense is a 28mm lense is a 28mm lense... ie that all that was happening was that a crop of the larger film image was being made... like drawing aorund a part of a larger projection on a wall. Therefore, the DoF scale would be just as appropriate for one sensor as the other. Hence no mention of this by Leica. I hadn't taken into account the enlargement that Andy talks about. Quite what this means given print resolution and so on I'm not too sure either, although I'm not too worried about it since I accept that DoF scales are indicative rather than absolute. And of course, for non-dynamic sitautions you can re-shoot based on reviewing the image in camera. I'm looking forward to getting back into RF photography... how long before the pre-orders clear? Many thanks for the replies, especially their clarity given the usual circular and confused discussions which ensue under this topic. Best wishes- Andy Link to post Share on other sites More sharing options...
frc Posted September 17, 2006 Share #8 Posted September 17, 2006 This topic seems to be creating a constant circle of confunsion around the photographers discussing it. DOF stays DOF for a given lens, chopping of the edge of the receptionfield doesn't disturb the lens. It is not even aware of it. Lenses have les brains than a cat. Or a flea! What does change is a shift from 50 to 35mm for the standard angle of view. Yes this lens gives more DOF at the same aperture than a 50 mm does. Myth busted? Link to post Share on other sites More sharing options...
lct Posted September 17, 2006 Share #9 Posted September 17, 2006 The very short answer - assume about 1 stop LESS DoF using any given lens (NOT field of view - LENS) on the M8 over the M7. I.E. shooting at f/11, use the DOF scale marks for f/8.... Agree. Works fine on the R-D1 with any lens. Link to post Share on other sites More sharing options...
ho_co Posted September 18, 2006 Share #10 Posted September 18, 2006 DOF stays DOF for a given lens, chopping of the edge of the receptionfield doesn't disturb the lens. Not quite. You're right that the light rays still come the same from the lens--perfect analysis. But because the sensor is smaller, the image must be blown up more to get the same final size print. For that reason, the crop factor must be applied also to the circle of confusion, which is used to calculate DoF, hyperfocal distance, etc. Look up those formulas; check LCT's computations on this forum while awaiting the M8 announcement for practical application. See also the article on effective f-stop in LFI a few months back: Depth of field from a given focal length is different depending on format IF you're talking about the same finished image size. --HC Link to post Share on other sites More sharing options...
marknorton Posted September 18, 2006 Share #11 Posted September 18, 2006 All of which points to the need for faster wide-angles... Link to post Share on other sites More sharing options...
TimF Posted September 18, 2006 Share #12 Posted September 18, 2006 But because the sensor is smaller, the image must be blown up more to get the same final size print. No it doesn't. The maximum print size of any digital camera is determined by the number of pixels on the sensor, and the output resolution chosen during post-processing. The Canon 5D and the Nikon D2X both have 12.xx million pixels (the exact number might be slightly different, but I can't be bothered to look up the Nikon details). Following your argument about smaller chip, it would be neccessary to enlarge the Nikon shots more to get the same sized print as from the Canon, but we know that's not so. Enlargement only comes into play when ressing-up beyond what the camera makes available. Link to post Share on other sites More sharing options...
gareth_c Posted September 18, 2006 Share #13 Posted September 18, 2006 Hmm, I can't answer the question and won't cloud the issure any more by trying to but I do have another question! AM I right in thnking that the distance from the lens to film in an M7 is the same as teh distance from lens to sensor in the M8? I am sure I read that the lens mount protrudes slightly on the M8 to ensure that the distance is the same...am I wrong? Link to post Share on other sites More sharing options...
marknorton Posted September 18, 2006 Share #14 Posted September 18, 2006 The lens mount is set slightly proud of the front of the camera on the M8 whereas it is flush on the M7. The distance to the focal plane in both cameras is the same at 27.8mm but of course the focal plane of the sensor is behind the cover glass. The reason for moving the mount forward by this couple of mm is because the package of the sensor, its mount and the LCD display can only be made so thin and it would otherwise protrude from the back of the camera as it does on the DMR. Space is needed in front of the sensor as well because the focal plane is not the front of the sensor and there in the shutter in front of it, which is presumably the reason why collapsible lenses are bad news - an imprint of the shutter blades on the front of the sensor would not make your day. Link to post Share on other sites More sharing options...
adan Posted September 18, 2006 Share #15 Posted September 18, 2006 Tim, that just is not how it works. If a sensor is 1/2" on the short side, and you make a print from that sensor that is 10" on the short side - that is a 20x enlargement. Any blurs projected by the lens will be 20x larger in the print than they were on the sensor. If the sensor is 1" on the short side, and you make a print from that sensor that is 10" on the short side - that is a 10x enlargement. Any blurs projected by the lens will be 10x larger than they were on the sensor. Absolutely no different than making traditional darkroom enlargements from a 35mm negative vs. a Minox negative in terms of the calculation. Original dimensions x enlargement = final dimensions. In terms of pixels per inch - prints from the Canon and Nikon of equal megapixels (like you, I'll assume they are close enough) will have the same resolution. If they were both shot with "equal field of view" lenses at the same aperture - a 28 on the Canon and an 18mm on the Nikon, both @ f/5.6 - the Nikon picture will have MORE DoF than the Canon. If you shoot with a 28mm lens on both cameras at a moderate aperture, so that something in the background is blurred to a diameter of 1mm on the sensor - then in equal-sized prints (let's say 10 x 15) the blur diameter will be 10mm in the Canon print and 15mm in the Nikon print - more obvious blur, and thus less DoF for the Nikon using the same lens (not the same field of view). Link to post Share on other sites More sharing options...
frc Posted September 18, 2006 Share #16 Posted September 18, 2006 With a given angle of view it will in my opinion not change dramaticaly. 50mm has a less deep DOF at f2 than a 35mm. 35 1.33 crop becomes more or less the 50 AOV on film. 35 gives more depth, crop needs a smaller CoC due to the increased enlargement needed. DOF deminishes, and comes closer to the 50. How close it will be I don't know. I must say I don't care either, I'll find out when the cam arrives. The DOF behaviour of my 35 lux asph is dramaticaly different compaired to the cron 50. A factor 0.75 reduction on the size of the CoC on a handheld camera already overengineered shoold not make a big difference. On a tripod or the testbench it'll become a bit more scientific, that's where the formulas come in. I'm more concerned about the lenses subjective image ( fingerprint ) Besides, I wonder when i'll have time to study the DOF-scales on my lenses. Just to busy taking pictures. Link to post Share on other sites More sharing options...
sean_reid Posted September 18, 2006 Share #17 Posted September 18, 2006 Ah yes, one of our more famous regular topics. Since there has already been much discussion here, I'll add a simple point which is that "Circle of Confusion" is a theoretical construct - it doesn't exist in any firm state. The degree to which a given area of a picture will appear sharp vs. unsharp will be influenced by many different variables, not the least of which are print size, dot spread on the printer, etc. My suggestion is that these calculations are of limited value as are the DOF markings on any lens (because they are all set in relationship to a set of assumptions about factors that can vary tremendously). So, DOF scales never were absolute with film cameras and they aren't absolute with digital cameras. The idea of allowing for about one stop less DOF than expected could be useful for a camera with a 1.3 - 1.5 crop. But the very best method is to ignore all theoretical calculations and instead do some individual experiments with a given camera and a given set of lenses at various distances and apertures. An afternoon spent doing this will tell one more than all the COF charts in the world. Perhaps done with a notebook and some nice wine and a good lunch... I'm with the poet William Carlos Williams in this respect: "No ideas but in things" Above all, trust your own eyeballs over a formula. Or, as Mark Twain once said, there are three kinds of liars: liars, damn liars and statistics. Cheers, Sean Link to post Share on other sites More sharing options...
johnbuckley Posted September 18, 2006 Share #18 Posted September 18, 2006 Sean, but the question I would have is -- separate and apart from the technical specs, and without violating your embargo agreement with Leica -- is it reasonable to assume that the glorious bokeh one gets using fast Leica lenses will make the transition from the analog Ms to the M8? If an advantage of rangefinders over SLRs is, as you have stated, that you have to imagine the blur behind what is viewed in the viewfinder as being in focus, does it in fact blur with the M8? I love my D2 as having many of the best attributes of a rangefinder, but bokeh it does not produce, and the reasoning -- back to the tech specs -- was always the sensor size. Please tell us that we can still take a portrait with our 50 Summilux and have a blurred background! Link to post Share on other sites More sharing options...
ho_co Posted September 18, 2006 Share #19 Posted September 18, 2006 Even Leica has long made known that their DoF markings are based on a traditional measurement for circle of confusion. With modern lenses, the CoC should be noticeably smaller. Compare the DoF engravings for Zeiss Contax lenses to those of Leica. The prewar and postwar Zeiss lenses show twice the DoF of their Leica counterparts. Zeiss apparently based their calculations on a final print of ca 4" x 6", while Leica's calculations were for 8" x 10". So if you want a technical background, I haven't seen a more comprehensive one than the ones suggested by Rubén above at http://www.leica-camera-user.com/digital-m-m8-release/5541-m8-dof.html?posted=1#post52591. And if you want to take pictures, use Sean's rule above at http://www.leica-camera-user.com/digital-m-m8-release/5541-m8-dof.html?posted=1#post53375. --HC Link to post Share on other sites More sharing options...
TimF Posted September 18, 2006 Share #20 Posted September 18, 2006 If you shoot with a 28mm lens on both cameras at a moderate aperture, so that something in the background is blurred to a diameter of 1mm on the sensor - then in equal-sized prints (let's say 10 x 15) the blur diameter will be 10mm in the Canon print and 15mm in the Nikon print - more obvious blur, and thus less DoF for the Nikon using the same lens (not the same field of view). I'm not sure I fully get what you're saying here Andy. From your last sentence (the bit in brackets/parentheses/whatever) you seem to imply that both cameras in this example are set up at the same distance from the subject, so the captured angle of view on the 5D will be for a 28mm lens. On the D2X you will only be getting the angle of view of a 43mm (ie x1.54, which Thom Hogan says is the more exact Nikon crop factor). If the full captured frames are then printed at the same size, then yes, you are quite right. However, if the Canon frame is cropped to the same angle of view as the Nikon shot then both are printed at the same size things are different. It all depends on the output resolution; with a full frame print the resolution is equal, and the Canon will appear to be superior in terms of depth of field. Crop the Canon shot to the AoV of the Nikon and print to the same size - you will then need to either ress the Canon shot up, or print the Nikon at a higher resolution. The pixel size of the 5D is 8.2 microns, that of the D2X is 5.5. I'll leave the maths to you. Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.