Telephoto Resolution: Which is better? Film or Digital
Discussion
23 replies
Posted by Pat Warfield on April 29, 2007 at 2:33 PM
Which has more information: 35mm film (100-200 ASA) or Digital 12+ MP.
I am perplexed about quality & snap. Digital Images from long telephoto still seem soft after doing everything right: Heavy rock-solid tripod, gimbal head locked down, mirror lockup, shutter release. Camera & lens further supported by a rail. No wind. There is no movement. Even shot in cool clean air after a rain - no polution or atmospheric haze. Using Canon's 400mm f/5.7L lenses with & without extenders on 5D and 20D. Also used 70-200 f/2.8L IS w/ extenders. Med-range images are tack sharp. Is there a digital problem? I don't remember film from telephoto being so soft.
I'm going to try the same shots with my old Canon FD system.
Posted by Pat Warfield on April 29, 2007 at 2:36 PM
Also may add: All other photography from this equipment is superb - sharp & rich colors.
Pat
Posted by Kent Staubus on April 29, 2007 at 10:43 PM
You didn't mention what kind of sharpening you were using. If your next question is, "What's sharpening?", I'd have to say "BINGO!"
Kent in SD
Posted by Tom Webster on April 30, 2007 at 2:53 PM
Because of an "anti-aliasing" filter over the ccd/cmos sensor, all digital images will be soft to some degree. You have to add sharpening to digital images.
I think Pat is asking about sharpness as a function of focal length when shooting digital, i.e., are telephoto digital images going to be softer than normal/wide angle digital images (assuming that the equipment is functioning as it should)?
Posted by Paul Wegemann on April 30, 2007 at 7:03 PM
Pat, your question about which has more information - 35mm or 12+MP digital had been debated extensively on the net. Even Popular Photography mag had an article awhile back basically stating that the Canon 1DS MkII out-resolved 35mm (I think Velvia). This link has a really interesting chart -
http://www.clarkvision.com/imagedetail/film.vs.digital.summary1.html
Anyway, I'm wondering if you don't have a focus problem with your lens. I've gotten some razor sharp images at 200mm with my 70-200 f4 L, but every once in awhile it just doesn't quite get the focus right at 200mm. Then if I manually focus, or even turn the camera (20D) off and back on, things seem to correct themselves. I should probably get this fixed, but I don't want to be without the lens for weeks.
As far as information, high quality 35 mm film is superior to digital. The question above is about information which equates to resolution.
If memory serves me correctly, Velvia had a resolution of 150 line pairs per mm. It takes 4 RGB dots to resolve one (black and white) line pair. With perfect alignment, two dots would be black while two would be white. At worst, one would be black, two would be 50% gray while the fourth would be white.
A 35mm frame is 24mm by 36mm. The requirement would be a sensor of 14,400 by 21,600 to potentially equal the resolution of Velvia. That equates to a 311 megapixel sensor. And, if one were to consider that the sensor uses a Bayer pattern along with the previously mentioned anti-aliasing filter, the real requirement is even greater.
Every time I have read someone state that their digital camera is so superior to the resolution of film, the only explanation can be that they are comparing a top end DSLR to 800 ISO film in a poor quality body.
T. David, I'll comment, but only because I shoot film. I love film, and have moved to 6x7. I have a Nikon 9000 ED scanner and feel I get better results than drum scans because the PMT picks up crazy amounts of film grain, even at the best imaging places. Even with the sharpness of a drum scan, the very best scan I've seen with Velvia 100, a fixed lens at f/8 is close to 4,000 ppi. IMHO from some experiments with scanning and some limited DSLR shooting, I'd guess the best, upper end, realistic MP equivalent to 35mm is somewhere between 15 to 22. Even then, you need some work with Neat Image to address the film grain problem. Your example may work theoretically, but I've never seen a 48" x 72" print that had a true resolution of 300 dpi (max resolution on a lightjet print) with 35mm, less alone MF or high end digital.
To Pat, I can't think of any reason a longer telephoto is going to change the way the light is focused on the sensor to result in a softer image. I know there can be certain exposures around 1/60 to 1/30 of a sec that can introduce some camera shake when longer or shorter exposures do not exhibit that same problem. I have a Pentax 67ii and a 200mm lens that is hard to get tack sharp with a tripod and mirror lockup at 1/60 of a second.
The guy whose website is linked above states that he got 48 lp/mm resolution from Velvia with a Canon 500/4 IS. So, my question would be - which lens gives you 150 lp/mm, which is the basis for your claim that film is superior to digital (300+ MP seems to be a bit much...)? And why are 4 RGB dots needed to see a line pair (one black, one white), and not 2?
Anyway, I think you always have to look at the whole optical system, not the sensor or film alone. As for the OP - I've never heard of a film vs. digital effect that worsens telephoto shots, but not those with shorter focal lengths. So my guess would be that a) your telephoto setup can't outresolve the digital sensor and you've never looked closely enough on film to see the tiny bit of softness, or b) there is some misfocusing issue that becomes apparent by the shallow DOF usually obtained when shooting telephotos wide open.
The 150 line pairs per mm is a "laboratory" measurement made under perfect conditions by the manufacture, Fuji. (I just looked up Fuji's data and they show 160 line pairs per mm for a 1000:1 contrast, so my estimate is slightly low.) It's like the EPA mileage estimates where actual results may be different in the real world. That is why some images are "tack" sharp which requires superb optics (lens), a rock solid support (tripod) and an excellent frame (camera body). And, lastly, one also needs an equally fine microscope so that the line pairs can be measured.
While the measurement may be a laboratory value, it still means that the film will store the information. That is, the fuzziness on the film is true to the optical system, not an artifact of the data recorder. In the latter case, information is not being recorded rather it has been modified.
My feeling that 4 RGB dots are needed per line pair comes from the following logic. This explanation counts the dots in a line perpendicular to the line pair. If only two RGB dots were used, then one could observe a black dot and a white dot if the registration was perfectly aligned with the pixels. But if the alignment was exactly between the pixels, then one would observe two equally gray dots. If one were to use 4 dots, then when perfectly aligned there would be 2 dots which are black and two dots which are white. At the worst registration, one would have in sequence a black dot, a gray dot, a white dot and finally a gray dot. The two gray dots are caused by a blend of the black and white along the boundary between the white and black of the line pair.
This is really a simplistic estimate. In reality, digital camera sensors use a Bayer pattern, with 1/2 the pixels being green, 1/4th red and 1/4th blue. The pattern is subject to Moire (hairingbone) effects so an anti-aliasing (blurring) filter is installed in front of the sensor. The effect is demonstrated in the article by Clark cited above. His example, found on
shows that the Moire effect for lines on a 100 line pairs per inch pattern are effectively eliminated when a resolution of 600 DPI is used which is a Nyquist 3X sampling. My calculation was based on a 2X Nyquist sampling and could still have ugly artifacts. That sampling would increase my 311 mega-pixel requirement by 225% to 700 mega-pixels.
This problem is eliminated with film as the grains are randomly scattered in the emulsion layers. The blurring of the anti-aliasing filter will significantly increase the pixel requirement over that calculated above.
Everyone! This is a wonderful discussion! Good science! Thanks for all the input. And yes, Kent I do know about several methods of sharpening and usually apply some to each photo as needed. (Not sure what the BINGO is about. Is it to imply I'm not as SMART as you?)
Steve. In regards to possible equipment failure, the same day I had problems with this particular long shot, I also made some 'perfect' shots of other subjects. Some were even hand-held from my vehicle (driving through a preserve) and were tack sharp & color saturated, I couldn't improve on them by sharpening more. (I used to use a 6x7 Pentax and also a Bronica & found the focusing to be off frequently as they aged. I believe the film pressure plates had relaxed.)
I am a visual artist not a scientist nor an engineer, so I have to rely upon my vision & my knowledge of photographic techniques (years of film & large/medium format & darkroom work). When all known techniques seem to fail - I have to look to others for answers. And you have stepped up, we can all learn from each other.
Tom & David, thanks for the info on the anti-aliasing filter. So. . . this would somewhat answer my question of digital characteristics. It seems logical that this filter in front of the sensor could be causing the 'soft' effect, which would be more obvious in cropped subjects (eagles & shore birds) at 100 yards away. I have improved the images with various sharpenings but it is still not a 'great' shot. However, I am only evaluating it on my screen - not as a printed piece. (Printing by a good lab gives amazing results!) But my screen is usually very reliable.
T. David. Can we counter Moire effect with image processing or interpolating to higher DPI? Is this a solution? Or a diagnosis? I don't know how to put the science to work.
Would spending 5 or 6 K for the 500 or 600 F4L IS (for better & stronger glass) solve the problem? What are some of your experiences with these lenses?
I will shoot some film of the subject shortly, then I will write about my results. It won't be scientific, it will be just how it looks to me.
Everyone! This is a wonderful discussion! Good science! Thanks for all the input. And yes, Kent I do know about several methods of sharpening and usually apply some to each photo as needed. (Not sure what the BINGO is about. Is it to imply I'm not as SMART as you?)
Steve. In regards to possible equipment failure, the same day I had problems with this particular long shot, I also made some 'perfect' shots of other subjects. Some were even hand-held from my vehicle (driving through a preserve) and were tack sharp & color saturated, I couldn't improve on them by sharpening more. (I used to use a 6x7 Pentax and also a Bronica & found the focusing to be off frequently as they aged. I believe the film pressure plates had relaxed.)
I am a visual artist not a scientist nor an engineer, so I have to rely upon my vision & my knowledge of photographic techniques (years of film & large/medium format & darkroom work). When all known techniques seem to fail - I have to look to others for answers. And you have stepped up, we can all learn from each other.
Tom & David, thanks for the info on the anti-aliasing filter. So. . . this would somewhat answer my question of digital characteristics. It seems logical that this filter in front of the sensor could be causing the 'soft' effect, which would be more obvious in cropped subjects (eagles & shore birds) at 100 yards away. I have improved the images with various sharpenings but it is still not a 'great' shot. However, I am only evaluating it on my screen - not as a printed piece. (Printing by a good lab gives amazing results!) But my screen is usually very reliable.
T. David. Can we counter Moire effect with image processing or interpolating to higher DPI? Is this a solution? Or a diagnosis? I don't know how to put the science to work.
Would spending 5 or 6 K for the 500 or 600 F4L IS (for better & stronger glass) solve the problem? What are some of your experiences with these lenses?
I will shoot some film of the subject shortly, then I will write about my results. It won't be scientific, it will be just how it looks to me.
Pat, I'll be interested in your opinion after the comparison.
Would spending 5 or 6 K for the 500 or 600 F4L IS (for better & stronger glass) solve the problem? What are some of your experiences with these lenses?
I have a 400mm f5.6 L and the 500mm f4.0 IS L. If you don't like the results from the 400mm, the 500 mm will not make you happy.
Since you are satisfied with the performance at 200mm, I'm wondering if there isn't something about the long lens technique that we are missing - but from your description, nothing comes to mind. You seem to have minimized any shake. What shutter speeds are you using??
Pat, I know you stated you used a rock solid tripod, but maybe you want to increase your ISO more. I've heard the 5D can shoot a pretty fast ISO and not have noise. You may be getting some motion blur for some odd reason due to a 'longer' exposure.
Pat, interpolating the pixels will tend to make moire effects worse. The good news is that most nature photos consist of random elements which do not cause the moire effects. The times which I have encountered a moire effect were with closely stripped prints on clothing. For that matter, the print of the image was fine while only the image on the screen had the interference pattern.
While most moire interference patterns are caused by man made objects, such as picket fences, nature can provide subjects which can cause the effect. Anything that has numerous close parallel lines, i.e. tall grasses, can cause the problem.
If you suspect you have moire interference in your image, you need to view it at 100% so as to eliminate any interference pattern from the monitor.
There is a technique to reduce the moire effect described by Katrin Eisman in Photoshop Restoration and Retouching, ISBN 0789723182. With the image in the L-A-B mode, the clean up uses a Gaussian blur of the A and B color channels and then a sharpening of the lightness channel using the unsharp mask. Blurring the color gets rid of most of the moire effect while not sacrificing significant image quality. Image sharpness is perceived from the lightness information, not colors. Moire patterns which I have seen on the screen usually have colored areas standing out like a sore thumb.
Again, the good news is that most nature photography images do not cause moire patterns.
Bill, I work at f18-22 at about 350-500 when shooting the nest. I am really pleased with the 70-200 2.8 IS, which performs well with 1.4 extender, falls off a little with 2x. I can't say I don't like the 400 f5.6, but for this project it can't quite capture the eagles' nest I'm shooting without an extender. With one, I lose autofocus & have to work at manual focusing which is difficult on Canon's DSLRs. (Canon please do something about this!) Plus the 2x does degrade the image. I would get closer if allowed - I did luck in one day, they flew to a tree close by me & said howdy. Well, they acted like they knew me. Got some nice flight shots.
Tell me how your 400 f5.6 & 500 f/4L IS works for you.
Steve, I often use 400-800 ASA for the depth of field & shutter speed. Then I will try some at 100 ASA for less grain (noise).
T.David, I understand the moire effects, I have seen them in certain patterns, but not with my eagles. I will make note of your corrections if I do have a problems with a pattern.
Well, I am not giving up. I do have some nice work from the shoots - I am competent in Photoshop & have created some very nice 'Art' prints. But, I want to solve the 'distance' problem for pristine shots. Thanks again for being so attentive.
At f18-22 you may be seeing diffraction degrading the image quality ???? I generally shoot wide open, or only slightly stopped down due to the need for shutter speed to stop subject motion. Since the shallow DOF helps isolate the subject, I prefer to optimize shutter speed.
With the Canon 1DsMk2, I still have autofocus when using the TC, and the viewfinder is bright for manual focus, if needed. So, for additional $$, Canon does solve your problem. This last weekend I was at a wildlife refuge shooting birds in flight from a boat with the 400mm + 1.4TC, handheld. Wide open, at ISO 800, the shutter speeds were 1/2500 +, and the eyeballs were sharp.
I take back my earlier comment about the 600mm. For your particular situation, the 5D with the 600mm & 1.4x should get you a noticeable jump in image quality, with hopefully little to no image cropping. Any possibility of renting or borrowing that lens in your area?
Bill, I will try shooting wide-open. I hesitated to do that because thought I needed the DOF since it was hard to tell if the subject was truly in-focus. But, I do prefer out of focus backgrounds. I am considering selling my 400mm because the 70-200 w/converter covers that length & with autofocus & IS.
While I dream of owning the EF 600mm f/4 IS - it will have to wait. I know I can rent one of the 500mm lens, maybe a 600. However for now, I'm considering purchasing a 600 or 800 FD lens for my old A1 or F1 which focus easily. For the long shots on tripods I don't need autofocus & my lab does a good job of scanning film. If I can get one for $1000 or less it would be an economical leap up . . . !??? What do you guys think. We used to do good work in the old days didn't we. My old F1 had been a treasure in cold or inclimate weather when batteries got cold & all the electronics went dead.
Thanks again for your comments, we may solve the problem yet.
Pat Warfield
Cleo Springs, Oklahoma
www.warfieldstudios.com
I`ve seen a test of the 400/5.6L that showed definite loss of quality at apertures smaller than f/11. I`m not sure if it was deemed purely a diffraction issue or something more.
I own this lens and I`ve tested it both on film and digital: it`s very good indeed. I usually shoot wide open or at f/8 and only stop further down for DOF. If your lens behaved significantly better on film than your digital bodies, it`s AF may need to be calibrated by Canon to suit the new bodies.
If you find that the 70-200 plus 2x gives you comparable quality as the bare 400, there`s definitely need for calibration.
Instead of reviving your FD series bodies, you may adapt a manual focus super telephoto to your EOS bodies. Canon FD lenses don`t work (they require a rare 1.26x teleconverter/adapter to retain infinity focus) but almost anything else will: Nikon, Contax/Yashica, M42, Leica R, Olympus OM (and Pentax K with minor modification). Nikon MF 500/4P, 600/5.6 or 400/3.5 can be had for significantly less than their AF counterparts. It is also possible to replace your focus screens with third party alternatives that have split image or other manual focus aids.
Ott,
Thanks for the info on the aperture size relating to loss of quality. I did not research the optimum aperture for this lens. My mistake! Opening up could solve the problem. I hope!
I'm also going to try the lens on a film camera, EOS A2 I sold to a friend - see if there is a difference. Will post my observations.
My idea of using super tele FD 'L' lenses on FD bodies is for the purpose of taking advantage of their split image finders. I never used to have a problem 'seeing' when a subject was in focus. There was some darkening of half of the split but could see enough to line up the halves. (It would be a good backup to have a totally manual system in remote or hostile areas.)
I installed a Haoda Diagonal Split-Image screen in the 20D. It is good for checking focus on general photography, & for macro to fine-tweak a subject, but it's worse than useless in super tele outdoors at 5.6 or more. If anyone has experience with some other brands of screens with this same usage, let me know.
It sound good to use high quality manual lenses on these DSLRs. But in practice it didn't work for me. I used a Tamron 300mm 2.8 manual Adaptal lens on the Canons for a while. Same problem with Canon's EOS focusing system in manual. The image would look focused in the viewfinder, but when downloaded they were not. If someone knows of a really reliable manual focusing aid for EOS, let me know.
So as you can see, this is a long story & a long journey. I am anxious to try the wider aperature concept as soon as the weather clears.
Sorry to come back to this discussion late, Pat. Your image sharpness issues are two-fold, IMO.
First, I, too, believe your image sharpness issue is due to diffraction effects at small apertures. Diffraction effects occur most dramatically using long focal length lenses. I have the Canon EF 500mm f 4L IS lens and it is its sharpest from f 4 to f 8, after which the images lose contrast and sharpness. At the most, even with stacked teleconverters, I only stop the lens down a maximum of 2 f stops.
Focusing screens with split-image centers are nearly useless at small maximum apertures, as you have found. I believe either Canon or Nikon offered a special split image screens for telephoto lenses but they are rare to find on the used market and only fit the older manual focus bodies.
Second, focusing manual focus lenses on autofocus bodies depends a great deal on the quality of the FD/EF lens adapter you place on the lens. The cheaper units sold by third party manufacturers will severely degrade image sharpness. These adapters have to have an optical element in the lens adapter so that the FD lenses will focus at infinity on the Canon EOS bodies. The lens flange to film (cmos) plane is different between a manual body and an autofocus body. The optical element also acts as a slight teleconverter, too.
The best FD/EF adapter was made by Canon but was only produced for a short period of time after Canon switched to EF lenses. They made a few of these adapters for the pro photographers so as to give the pro photographers, who were heavily invested in FD lenses, time to sell their FD lenses and purchase newer EF lenses. These adapters were never meant to be a permanent solution to using FD lenses on EOS autofocus bodies. The optical element in the Canon FD/EF adapters were a higher quality but also degraded image quality when FD lenses were used on EOS bodies. You can remove the optical element from the FD/EF adapters but then you are left with basically an extension tube between the FD lens and EOS body. It will have little effect on image quality but it will severely limit how far away you can focus on more ditant subjects.
So, if you are using FD mount lenses on EOS bodies using an FD/EF lens adapter with the optical element intact, you are degrading contrast and image sharpness. If, on top of that, you are using f stops smaller than 2 f stops past the maximum aperture of the telephoto lens, you are degrading image sharpness and contrast further. Using both techniques together will compound the degradation of image sharpness and contrast. It is no surprise to me that, given the circumstances, that you will get sharper results on film when you mate an FD lens on a manual Canon body. This lens/body combination was made to work together properly.
T. David: the problem with the analysis you're providing is that grain is equivalent to noise. While you might be able to resolve 150lpi with Velvia, the lines aren't clean or unbroken by the grain. This reduces the effective sharpness of the image to a point somewhere below that of the 12+MP digital cameras Pat is asking about. I feel much more comfortable enlarging images from my D2X than I ever did enlarging oil-mounted drum scans of my 35mm transparencies.
Les, yes the lines are blurry. BUT, they can be recognized. And, they are several times the size of the grain.
All one has to do is look at a good transparency through a good loop to see detail that cannot be seen in the current digital cameras, no ifs, no ands and no buts. The arithmetic above is valid. Granted, to see 150+ line pairs per mm in a transparency takes a good eye with an excellent microscope. But, digital sensors currently are only capable of consistantly resolving at most 30 line pairs per mm And that is only possible if no anti-aliasing mask is used..
The information is there. And, the question was which media captures more information. And, no matter what you may want to believe, fine grained film still captures more information.
Discussion
23 repliesI am perplexed about quality & snap. Digital Images from long telephoto still seem soft after doing everything right: Heavy rock-solid tripod, gimbal head locked down, mirror lockup, shutter release. Camera & lens further supported by a rail. No wind. There is no movement. Even shot in cool clean air after a rain - no polution or atmospheric haze. Using Canon's 400mm f/5.7L lenses with & without extenders on 5D and 20D. Also used 70-200 f/2.8L IS w/ extenders. Med-range images are tack sharp. Is there a digital problem? I don't remember film from telephoto being so soft.
I'm going to try the same shots with my old Canon FD system.
Pat
Kent in SD
http://www.clarkvision.com/imagedetail/film.vs.digital.summary1.html
Anyway, I'm wondering if you don't have a focus problem with your lens. I've gotten some razor sharp images at 200mm with my 70-200 f4 L, but every once in awhile it just doesn't quite get the focus right at 200mm. Then if I manually focus, or even turn the camera (20D) off and back on, things seem to correct themselves. I should probably get this fixed, but I don't want to be without the lens for weeks.
If memory serves me correctly, Velvia had a resolution of 150 line pairs per mm. It takes 4 RGB dots to resolve one (black and white) line pair. With perfect alignment, two dots would be black while two would be white. At worst, one would be black, two would be 50% gray while the fourth would be white.
A 35mm frame is 24mm by 36mm. The requirement would be a sensor of 14,400 by 21,600 to potentially equal the resolution of Velvia. That equates to a 311 megapixel sensor. And, if one were to consider that the sensor uses a Bayer pattern along with the previously mentioned anti-aliasing filter, the real requirement is even greater.
Every time I have read someone state that their digital camera is so superior to the resolution of film, the only explanation can be that they are comparing a top end DSLR to 800 ISO film in a poor quality body.
To Pat, I can't think of any reason a longer telephoto is going to change the way the light is focused on the sensor to result in a softer image. I know there can be certain exposures around 1/60 to 1/30 of a sec that can introduce some camera shake when longer or shorter exposures do not exhibit that same problem. I have a Pentax 67ii and a 200mm lens that is hard to get tack sharp with a tripod and mirror lockup at 1/60 of a second.
Anyway, I think you always have to look at the whole optical system, not the sensor or film alone. As for the OP - I've never heard of a film vs. digital effect that worsens telephoto shots, but not those with shorter focal lengths. So my guess would be that a) your telephoto setup can't outresolve the digital sensor and you've never looked closely enough on film to see the tiny bit of softness, or b) there is some misfocusing issue that becomes apparent by the shallow DOF usually obtained when shooting telephotos wide open.
Basil
While the measurement may be a laboratory value, it still means that the film will store the information. That is, the fuzziness on the film is true to the optical system, not an artifact of the data recorder. In the latter case, information is not being recorded rather it has been modified.
My feeling that 4 RGB dots are needed per line pair comes from the following logic. This explanation counts the dots in a line perpendicular to the line pair. If only two RGB dots were used, then one could observe a black dot and a white dot if the registration was perfectly aligned with the pixels. But if the alignment was exactly between the pixels, then one would observe two equally gray dots. If one were to use 4 dots, then when perfectly aligned there would be 2 dots which are black and two dots which are white. At the worst registration, one would have in sequence a black dot, a gray dot, a white dot and finally a gray dot. The two gray dots are caused by a blend of the black and white along the boundary between the white and black of the line pair.
This is really a simplistic estimate. In reality, digital camera sensors use a Bayer pattern, with 1/2 the pixels being green, 1/4th red and 1/4th blue. The pattern is subject to Moire (hairingbone) effects so an anti-aliasing (blurring) filter is installed in front of the sensor. The effect is demonstrated in the article by Clark cited above. His example, found on
http://www.clarkvision.com/imagedetail/sampling1.html
shows that the Moire effect for lines on a 100 line pairs per inch pattern are effectively eliminated when a resolution of 600 DPI is used which is a Nyquist 3X sampling. My calculation was based on a 2X Nyquist sampling and could still have ugly artifacts. That sampling would increase my 311 mega-pixel requirement by 225% to 700 mega-pixels.
This problem is eliminated with film as the grains are randomly scattered in the emulsion layers. The blurring of the anti-aliasing filter will significantly increase the pixel requirement over that calculated above.
Steve. In regards to possible equipment failure, the same day I had problems with this particular long shot, I also made some 'perfect' shots of other subjects. Some were even hand-held from my vehicle (driving through a preserve) and were tack sharp & color saturated, I couldn't improve on them by sharpening more. (I used to use a 6x7 Pentax and also a Bronica & found the focusing to be off frequently as they aged. I believe the film pressure plates had relaxed.)
I am a visual artist not a scientist nor an engineer, so I have to rely upon my vision & my knowledge of photographic techniques (years of film & large/medium format & darkroom work). When all known techniques seem to fail - I have to look to others for answers. And you have stepped up, we can all learn from each other.
Tom & David, thanks for the info on the anti-aliasing filter. So. . . this would somewhat answer my question of digital characteristics. It seems logical that this filter in front of the sensor could be causing the 'soft' effect, which would be more obvious in cropped subjects (eagles & shore birds) at 100 yards away. I have improved the images with various sharpenings but it is still not a 'great' shot. However, I am only evaluating it on my screen - not as a printed piece. (Printing by a good lab gives amazing results!) But my screen is usually very reliable.
T. David. Can we counter Moire effect with image processing or interpolating to higher DPI? Is this a solution? Or a diagnosis? I don't know how to put the science to work.
Would spending 5 or 6 K for the 500 or 600 F4L IS (for better & stronger glass) solve the problem? What are some of your experiences with these lenses?
I will shoot some film of the subject shortly, then I will write about my results. It won't be scientific, it will be just how it looks to me.
Regards to you all,
Pat Warfield
Steve. In regards to possible equipment failure, the same day I had problems with this particular long shot, I also made some 'perfect' shots of other subjects. Some were even hand-held from my vehicle (driving through a preserve) and were tack sharp & color saturated, I couldn't improve on them by sharpening more. (I used to use a 6x7 Pentax and also a Bronica & found the focusing to be off frequently as they aged. I believe the film pressure plates had relaxed.)
I am a visual artist not a scientist nor an engineer, so I have to rely upon my vision & my knowledge of photographic techniques (years of film & large/medium format & darkroom work). When all known techniques seem to fail - I have to look to others for answers. And you have stepped up, we can all learn from each other.
Tom & David, thanks for the info on the anti-aliasing filter. So. . . this would somewhat answer my question of digital characteristics. It seems logical that this filter in front of the sensor could be causing the 'soft' effect, which would be more obvious in cropped subjects (eagles & shore birds) at 100 yards away. I have improved the images with various sharpenings but it is still not a 'great' shot. However, I am only evaluating it on my screen - not as a printed piece. (Printing by a good lab gives amazing results!) But my screen is usually very reliable.
T. David. Can we counter Moire effect with image processing or interpolating to higher DPI? Is this a solution? Or a diagnosis? I don't know how to put the science to work.
Would spending 5 or 6 K for the 500 or 600 F4L IS (for better & stronger glass) solve the problem? What are some of your experiences with these lenses?
I will shoot some film of the subject shortly, then I will write about my results. It won't be scientific, it will be just how it looks to me.
Regards to you all,
Pat Warfield
I have a 400mm f5.6 L and the 500mm f4.0 IS L. If you don't like the results from the 400mm, the 500 mm will not make you happy.
Since you are satisfied with the performance at 200mm, I'm wondering if there isn't something about the long lens technique that we are missing - but from your description, nothing comes to mind. You seem to have minimized any shake. What shutter speeds are you using??
While most moire interference patterns are caused by man made objects, such as picket fences, nature can provide subjects which can cause the effect. Anything that has numerous close parallel lines, i.e. tall grasses, can cause the problem.
If you suspect you have moire interference in your image, you need to view it at 100% so as to eliminate any interference pattern from the monitor.
There is a technique to reduce the moire effect described by Katrin Eisman in Photoshop Restoration and Retouching, ISBN 0789723182. With the image in the L-A-B mode, the clean up uses a Gaussian blur of the A and B color channels and then a sharpening of the lightness channel using the unsharp mask. Blurring the color gets rid of most of the moire effect while not sacrificing significant image quality. Image sharpness is perceived from the lightness information, not colors. Moire patterns which I have seen on the screen usually have colored areas standing out like a sore thumb.
Again, the good news is that most nature photography images do not cause moire patterns.
Tell me how your 400 f5.6 & 500 f/4L IS works for you.
Steve, I often use 400-800 ASA for the depth of field & shutter speed. Then I will try some at 100 ASA for less grain (noise).
T.David, I understand the moire effects, I have seen them in certain patterns, but not with my eagles. I will make note of your corrections if I do have a problems with a pattern.
Well, I am not giving up. I do have some nice work from the shoots - I am competent in Photoshop & have created some very nice 'Art' prints. But, I want to solve the 'distance' problem for pristine shots. Thanks again for being so attentive.
Pat Warfield
patwarfield.com
With the Canon 1DsMk2, I still have autofocus when using the TC, and the viewfinder is bright for manual focus, if needed. So, for additional $$, Canon does solve your problem.
I take back my earlier comment about the 600mm. For your particular situation, the 5D with the 600mm & 1.4x should get you a noticeable jump in image quality, with hopefully little to no image cropping. Any possibility of renting or borrowing that lens in your area?
While I dream of owning the EF 600mm f/4 IS - it will have to wait. I know I can rent one of the 500mm lens, maybe a 600. However for now, I'm considering purchasing a 600 or 800 FD lens for my old A1 or F1 which focus easily. For the long shots on tripods I don't need autofocus & my lab does a good job of scanning film. If I can get one for $1000 or less it would be an economical leap up . . . !??? What do you guys think. We used to do good work in the old days didn't we. My old F1 had been a treasure in cold or inclimate weather when batteries got cold & all the electronics went dead.
Thanks again for your comments, we may solve the problem yet.
Pat Warfield
Cleo Springs, Oklahoma
www.warfieldstudios.com
I own this lens and I`ve tested it both on film and digital: it`s very good indeed. I usually shoot wide open or at f/8 and only stop further down for DOF. If your lens behaved significantly better on film than your digital bodies, it`s AF may need to be calibrated by Canon to suit the new bodies.
If you find that the 70-200 plus 2x gives you comparable quality as the bare 400, there`s definitely need for calibration.
Instead of reviving your FD series bodies, you may adapt a manual focus super telephoto to your EOS bodies. Canon FD lenses don`t work (they require a rare 1.26x teleconverter/adapter to retain infinity focus) but almost anything else will: Nikon, Contax/Yashica, M42, Leica R, Olympus OM (and Pentax K with minor modification). Nikon MF 500/4P, 600/5.6 or 400/3.5 can be had for significantly less than their AF counterparts. It is also possible to replace your focus screens with third party alternatives that have split image or other manual focus aids.
Thanks for the info on the aperture size relating to loss of quality. I did not research the optimum aperture for this lens. My mistake! Opening up could solve the problem. I hope!
I'm also going to try the lens on a film camera, EOS A2 I sold to a friend - see if there is a difference. Will post my observations.
My idea of using super tele FD 'L' lenses on FD bodies is for the purpose of taking advantage of their split image finders. I never used to have a problem 'seeing' when a subject was in focus. There was some darkening of half of the split but could see enough to line up the halves. (It would be a good backup to have a totally manual system in remote or hostile areas.)
I installed a Haoda Diagonal Split-Image screen in the 20D. It is good for checking focus on general photography, & for macro to fine-tweak a subject, but it's worse than useless in super tele outdoors at 5.6 or more. If anyone has experience with some other brands of screens with this same usage, let me know.
It sound good to use high quality manual lenses on these DSLRs. But in practice it didn't work for me. I used a Tamron 300mm 2.8 manual Adaptal lens on the Canons for a while. Same problem with Canon's EOS focusing system in manual. The image would look focused in the viewfinder, but when downloaded they were not. If someone knows of a really reliable manual focusing aid for EOS, let me know.
So as you can see, this is a long story & a long journey. I am anxious to try the wider aperature concept as soon as the weather clears.
Thanks again
Pat Warfield
Cleo Springs, OK USA
http://diglloyd.com/diglloyd/infos/TheSharpestImage/TheSharpestImage-info.html
First, I, too, believe your image sharpness issue is due to diffraction effects at small apertures. Diffraction effects occur most dramatically using long focal length lenses. I have the Canon EF 500mm f 4L IS lens and it is its sharpest from f 4 to f 8, after which the images lose contrast and sharpness. At the most, even with stacked teleconverters, I only stop the lens down a maximum of 2 f stops.
Focusing screens with split-image centers are nearly useless at small maximum apertures, as you have found. I believe either Canon or Nikon offered a special split image screens for telephoto lenses but they are rare to find on the used market and only fit the older manual focus bodies.
Second, focusing manual focus lenses on autofocus bodies depends a great deal on the quality of the FD/EF lens adapter you place on the lens. The cheaper units sold by third party manufacturers will severely degrade image sharpness. These adapters have to have an optical element in the lens adapter so that the FD lenses will focus at infinity on the Canon EOS bodies. The lens flange to film (cmos) plane is different between a manual body and an autofocus body. The optical element also acts as a slight teleconverter, too.
The best FD/EF adapter was made by Canon but was only produced for a short period of time after Canon switched to EF lenses. They made a few of these adapters for the pro photographers so as to give the pro photographers, who were heavily invested in FD lenses, time to sell their FD lenses and purchase newer EF lenses. These adapters were never meant to be a permanent solution to using FD lenses on EOS autofocus bodies. The optical element in the Canon FD/EF adapters were a higher quality but also degraded image quality when FD lenses were used on EOS bodies. You can remove the optical element from the FD/EF adapters but then you are left with basically an extension tube between the FD lens and EOS body. It will have little effect on image quality but it will severely limit how far away you can focus on more ditant subjects.
So, if you are using FD mount lenses on EOS bodies using an FD/EF lens adapter with the optical element intact, you are degrading contrast and image sharpness. If, on top of that, you are using f stops smaller than 2 f stops past the maximum aperture of the telephoto lens, you are degrading image sharpness and contrast further. Using both techniques together will compound the degradation of image sharpness and contrast. It is no surprise to me that, given the circumstances, that you will get sharper results on film when you mate an FD lens on a manual Canon body. This lens/body combination was made to work together properly.
All one has to do is look at a good transparency through a good loop to see detail that cannot be seen in the current digital cameras, no ifs, no ands and no buts. The arithmetic above is valid. Granted, to see 150+ line pairs per mm in a transparency takes a good eye with an excellent microscope. But, digital sensors currently are only capable of consistantly resolving at most 30 line pairs per mm And that is only possible if no anti-aliasing mask is used..
The information is there. And, the question was which media captures more information. And, no matter what you may want to believe, fine grained film still captures more information.
End of argument.