cfritze
Established
Asked this question a few times in a few different places, but haven't gotten a real answer, nor a "what an idiot!" response - I'll take either if I can understand the reason why.
Let's leave aside the differences in pixel density, sharpness, noise, etc that may derive from the difference between the technical aspects the sensors themselves, be they full-frame (35mm size) or smaller. What about how much we see of the lens' coverage? That is, whereas on a cropped sensor we only see the central portion of what the lens would see on a 35mm sensor, we get all that view on a full-frame. Does this matter? Does the image drawn by lenses in those outer regions differ in character from the central region, and does this give an image a quality you'd miss in a smaller sensor? Though many lens designs strive for edge-to-edge resolution and lightness/contrast consistency, are there images that appeal to us because light fall off, loss of resolution at the perimeter are features that we can use to construct a likable image?
Ugh, been hard to say that clearly and succinctly. Anyone get what I'm driving at? Or am I missing something fundamental an exposing my inexperience?
thanks for any help
Let's leave aside the differences in pixel density, sharpness, noise, etc that may derive from the difference between the technical aspects the sensors themselves, be they full-frame (35mm size) or smaller. What about how much we see of the lens' coverage? That is, whereas on a cropped sensor we only see the central portion of what the lens would see on a 35mm sensor, we get all that view on a full-frame. Does this matter? Does the image drawn by lenses in those outer regions differ in character from the central region, and does this give an image a quality you'd miss in a smaller sensor? Though many lens designs strive for edge-to-edge resolution and lightness/contrast consistency, are there images that appeal to us because light fall off, loss of resolution at the perimeter are features that we can use to construct a likable image?
Ugh, been hard to say that clearly and succinctly. Anyone get what I'm driving at? Or am I missing something fundamental an exposing my inexperience?
thanks for any help
If you shoot portraits, the crop factor can give new use to an old lens. The Nikkor 55/1.2 comes to mind. Over the center 2/3rds of the image, it is as sharp as the 50/1.4. It sacrifices performance at the corners. As an "80/1.2 equivalent" on a Nikon 1.5x crop factor, it is superb.
Most Leica users want wide-angle fields-of-view. Here is where the FF sensor is much more important.
However, many long-time photographers want their favorite lenses to act like they always have with film. Those users will be the M9's biggest fans.
Most Leica users want wide-angle fields-of-view. Here is where the FF sensor is much more important.
However, many long-time photographers want their favorite lenses to act like they always have with film. Those users will be the M9's biggest fans.
ferider
Veteran
Good question, no, you are not missing anything and yes on all your questions.
Further and from the opposite angle, since (due to medium density) digital is less tolerant to focus inaccuracies, etc.; i.e., the M9 with its FF sensor will expose undesired lens weaknesses much more than the M8 with crop sensor.
When you check Raid's last 50mm lens test, you will see that most lenses, from the 50$ J8 to the 50/1.4 Summilux performed almost identically in picture center (at least for, say, an 8x10 print). Lens "character" is mostly outside of the center, and can distract or enhance, depending on subject and viewer.
Further and from the opposite angle, since (due to medium density) digital is less tolerant to focus inaccuracies, etc.; i.e., the M9 with its FF sensor will expose undesired lens weaknesses much more than the M8 with crop sensor.
When you check Raid's last 50mm lens test, you will see that most lenses, from the 50$ J8 to the 50/1.4 Summilux performed almost identically in picture center (at least for, say, an 8x10 print). Lens "character" is mostly outside of the center, and can distract or enhance, depending on subject and viewer.
Last edited:
jfretless
Established
I don't get the big fuss either.
...isn't a 35mm FF a cropped sensor compared to a MF FF sensor? Why is a 35mm FF sensor the bees' knees?
It's the subject that makes the image!
John
...isn't a 35mm FF a cropped sensor compared to a MF FF sensor? Why is a 35mm FF sensor the bees' knees?
It's the subject that makes the image!
John
count_zero
Established
You do get more of the sweet spot with the cropping, and your depth of view is magnified, which is what makes the 4/3 format so interesting (2x the depth of view). Take for example a Nikon DX lens designed for DX cameras looks like crap through a full frame Nikon D because of strange vingetting you get at the edges, they purposely leave the edges unrefined to save cost on DX lenses. As far as the sensor quality goes, more is usually better, but weird aberations happen at wide angles, which is why Leica's M9 added that fancy micro lens technology.
ferider
Veteran
It's the bees' knees for 35mm lenses.
Disaster_Area
Gadget Monger
I think the big push for full frame is because a big advantage of RF's is low distortion wides. I find the FF debate much less an issue with SLR's. I love the fact that my 85L is now a 135 f1.2 on my 50D. On my SLR I'm almost always reaching for more distance so the crop is a bonus, not a hindrance. Plus I'm getting the sharpest bit of the lens...another plus. But on a RF most people are trying to get wider and sharpness isn't always the be all of RF pics. Why else would people love the Jupiters 
cfritze
Established
Thanks for the responses. The RFF comes through again.
And I'd lost track of the lens tests by Raid with assistance from Ferider - work interrupted me just before the results came out. Having fun going back through those now.
And I'd lost track of the lens tests by Raid with assistance from Ferider - work interrupted me just before the results came out. Having fun going back through those now.
ramosa
B&W
i'm glad this question was asked--as i, too, have wondered about this. i have only used one RF--the M8--so i do not know how it would be to shoot a FF RF. i currently have a 28mm and 50mm. so am i right to think that the benefits of FF would be limited. (i don't mean to hijack this thread. i'm trying to expand upon it.)
Roger Hicks
Veteran
Yes, the benefits are limited if your lenses were bought to suit the M8. The main reason FF matters to so many Leica users is that we have been using Leicas for a very long time; probably use only a few focal lengths (which we already own) for the vast majority of our pictures; and have an almost instinctive understanding of their coverage. For example, your 28+50 equates to 37 and 67 on FF, for which my two main lenses are 35+75; the 35 for decades, the 75 since the Summicron came out, when it replaced the 90 Summicron in my affections.
Another reason is fast wide-angles. A 35 Summilux is expensive; a 24 Summilux is hair-raising; and there is no 28 Summilux.
Cheers,
Roger
sam_m
Well-known
Sorry, don't mean to nit pick, but I don't think this is accurate. The size of the sensor doesn't affect the 'tolerance', the pixel density does. Interestingly, the M8 and M9 have identical pixel densities, 6.8 microns (http://www.dpreview.com/previews/leicam9/), the increase in resolution happens to be offset perfectly by the increase in sensor size in this case - but if the m9 sensor was the same size as the m8, but with 18MP instead of 10MP, it would be less tolerant of lens defects.
ferider
Veteran
Agree, Sam I worded it poorly.
I meant the following: both crop and FF sensor have less tolerance to lens mis-collimation and abberations (for example focal plane curvature and shift) than film. The FF sensor will show it more since abberations often show stronger in the FF corners.
I meant the following: both crop and FF sensor have less tolerance to lens mis-collimation and abberations (for example focal plane curvature and shift) than film. The FF sensor will show it more since abberations often show stronger in the FF corners.
CK Dexter Haven
Well-known
1. Typically, the discussion about Full- versus Cropped- sensors has been amongst SLR users. And, there is a tremendous difference in the viewfinders. Probably moot between rangefinder models.
2. With regard to the character of images, i believe you get less of a dimensional rendering, the smaller the film or sensor. Compare 35mm to 6x6 to 4x5 and then to 8x10. With every step up, when there is defocus within the image, the larger format has more of a dimensional presence. I don't know if/how you can quantify that. If that's true, there should be an incremental increase in that quality when going from a compact digital to micro 4/3 to APS-C to full frame, and then up to the S2's middle-middle format, and then again up to the Hassy H. I'm sure the difference is pretty slight when we're talking about APS-C to full frame, but aren't we all here, talking about slight diffferences? [Which version of the Summicron 50mm should i buy? Is the Canadian 35 Cron better for b&w than the German?] This is why pictures from compact digicams look so flat (even ignoring their lack of DOF control), even when they might still have as much detail as pictures from bigger sensors.
3. I do think there is a difference between using the central part of a lens and using the entire glass. I see a difference when people crop 35mm rectangular images to make them look like square/6x6 images. The cropped pix are 'too perfect' in some way. Almost like you expect unsharpness and/or subtle vignetting in a real 6x6 picture. I remember one image by a guy i found online - shot with a bronica 6x6 - and the corners had a really beautiful distortion effect. They weren't obviously blurred or warped, but there was something subtly different about the edges versus the central, sharp portion. It was completely organic. I wish i could replicate that, and i've even tried with PS and modified/junked-up filters....
2. With regard to the character of images, i believe you get less of a dimensional rendering, the smaller the film or sensor. Compare 35mm to 6x6 to 4x5 and then to 8x10. With every step up, when there is defocus within the image, the larger format has more of a dimensional presence. I don't know if/how you can quantify that. If that's true, there should be an incremental increase in that quality when going from a compact digital to micro 4/3 to APS-C to full frame, and then up to the S2's middle-middle format, and then again up to the Hassy H. I'm sure the difference is pretty slight when we're talking about APS-C to full frame, but aren't we all here, talking about slight diffferences? [Which version of the Summicron 50mm should i buy? Is the Canadian 35 Cron better for b&w than the German?] This is why pictures from compact digicams look so flat (even ignoring their lack of DOF control), even when they might still have as much detail as pictures from bigger sensors.
3. I do think there is a difference between using the central part of a lens and using the entire glass. I see a difference when people crop 35mm rectangular images to make them look like square/6x6 images. The cropped pix are 'too perfect' in some way. Almost like you expect unsharpness and/or subtle vignetting in a real 6x6 picture. I remember one image by a guy i found online - shot with a bronica 6x6 - and the corners had a really beautiful distortion effect. They weren't obviously blurred or warped, but there was something subtly different about the edges versus the central, sharp portion. It was completely organic. I wish i could replicate that, and i've even tried with PS and modified/junked-up filters....
dimitris
Established
Having moved from a Canon 20D to a 1Ds I can honestly say that it makes a big difference. FF gives you much more out of focus areas for the same angle at the aperture. I know its only one thing but that can sometimes make or break a photo. If you are into blurring background and great bokeh then FF is the way to go. I know it sounds like you are paying a high premium for just one thing but such is life. Its the same reason why 85L is so much more expensive than the 85mm f1.8. To some it might be worth it and to others not.
GoneSavage
not actually
From my experience, there is a noticeable difference between images taken with cropped sensor formats vs full frame sensors. I agree with CK Dexter Haven when he says that cropped images can look "too perfect", and have also noted a loss of dimensional rendering when using a cropped sensor. However, I think the field of view of a lens has a huge effect on the images it takes as well, and this is where I find cropped sensor cameras to fall behind. A crop taken from a 50mm lens does not look like it was taken with a 40mm lens, even if the subject is the same size in the final image. Conversely, using a 28mm lens on a full frame camera and cropping the center will not create an image that looks like it was shot with a 35.
Mcary
Well-known
Don't know if this applies to rangefinders but for DSLRS with cropped sensor(s) lens makers have had to come-up with a new lens designed, know as an inverted telephoto for lens that are wider then a certain number of mm(s) due the fact that the rear element of a traditionally designed wide angle lens would protrude so far into the body that the mirror would hit. So far the fastest lens of this design has been F 2.8 Which while fast for a SLR zoom is rather slow compared to a SLR or rangefinder prime.
kuzano
Veteran
Is everybody driving at the same answer as I pose my "what an idiot" question
Is everybody driving at the same answer as I pose my "what an idiot" question
Full Frame sensor will display all the lens weaknesses (not sharp in the corner) that using them on film will.
Crop Sensors will hide the use of these weaker lenses by not utilising the actual weaker portion of the glass as the lens is using wide open, BUT benefit by using the light from wide open.
Hence, can one conclude that Crop sensors are better than Full Frame sensors for the majority of legacy lenses... even some of those from Zeiss and Leica.
Is everybody driving at the same answer as I pose my "what an idiot" question
Full Frame sensor will display all the lens weaknesses (not sharp in the corner) that using them on film will.
Crop Sensors will hide the use of these weaker lenses by not utilising the actual weaker portion of the glass as the lens is using wide open, BUT benefit by using the light from wide open.
Hence, can one conclude that Crop sensors are better than Full Frame sensors for the majority of legacy lenses... even some of those from Zeiss and Leica.
CK Dexter Haven
Well-known
This is only true if "better" is strictly a quantitative assessment. If sharper = better, and uniform sharpness = best, then, okay.
But, i don't think 'photographers' would agree. Technicians, maybe.
This is kinda like saying a Summicron is a better lens than a Sonnar. Maybe, if you have to shoot flat artwork, or need to use it on a copystand. But, 'better,' for photography? I LIKE a lens that has character.
40oz
...
to me, the main issue with using a crop sensor has to do with field of view and "angle of view" (I don't know what else to call it).
For example, we've all seen close-ups shot with a wide angle that show the dog's nose as a huge protrusion in the center of the frame. This just exaggerates the distortion a wide angle lens puts in the image. The same shot with a crop sensor would show the same distortion at the same distance (but you'd have less of the dog in the frame, of course).
I guess what I am saying is that the use of a crop sensor with full-frame lenses means that while the actual coverage of say a 35mm is similar to what you'd get with a 50mm with full-frame, the angle of view is still that of a 35mm, and so you get the same foreshortening effect (or lack thereof).
In addition, while it is true that cropping out the corners removes the weakest areas of a fast lens, you are magnifying the center areas and therefore magnifying any aberrations present. The advantage of a larger format is the reduction in apparent aberrations. The weakness of a smaller format is the magnification of aberrations, making them more apparent. In this case, it doesn't matter if you use a lens designed for a crop body - the aberrations are present and being magnified greatly to result in your standard 8x10. You have the same aberrations present in a lens intended for an 8x10 view camera but you aren't magnifying those aberrations so they are never apparent.
In addition, no matter what lens you are using, the number of pixels available to draw the scene is a limiting factor. A crop sensor might have 10000 pixels going across, but the final image merges 3 or 4 into 1 pixel. So that cable holding up a suspension bridge in the distance disappears as it is too small to show up on more than a single pixel. And you then get a bridge held up by a handful of random cables instead of a steady march the length of the bridge - some cables show up on multiple pixels and make it into the final image, some do not.
And as you resize the image to share on the internet, more cables are lost as the pixels representing them get dropped - you can't make a single pixel smaller, but the data it represents will be removed in the resampling.
However, if the same shot was taken on a larger format, each cable takes up more real estate on the sensor (be it film or digital) and so is maintained as the print is down-sized. This obviously has limits, as a larger digital sensor might just have larger pixels rather than a significantly higher number of them.
That said, an 8 MP full-frame sensor image should appear to have fewer apparent aberrrations than the 8 MP 1.6 crop sensor, even with the same lens. You are magnifying the crop sensor image much more, aberrations included. That's without even getting into the fact that to get the same coverage of the scene with the crop sensor, you need to be farther away, which means the lens simply cannot capture the same level of detail as it could at a closer distance. Also, atmospheric distortions will have a greater effect simply because there is more atmosphere between you and the subject, and you are magnifying them to a greater degree.
For example, we've all seen close-ups shot with a wide angle that show the dog's nose as a huge protrusion in the center of the frame. This just exaggerates the distortion a wide angle lens puts in the image. The same shot with a crop sensor would show the same distortion at the same distance (but you'd have less of the dog in the frame, of course).
I guess what I am saying is that the use of a crop sensor with full-frame lenses means that while the actual coverage of say a 35mm is similar to what you'd get with a 50mm with full-frame, the angle of view is still that of a 35mm, and so you get the same foreshortening effect (or lack thereof).
In addition, while it is true that cropping out the corners removes the weakest areas of a fast lens, you are magnifying the center areas and therefore magnifying any aberrations present. The advantage of a larger format is the reduction in apparent aberrations. The weakness of a smaller format is the magnification of aberrations, making them more apparent. In this case, it doesn't matter if you use a lens designed for a crop body - the aberrations are present and being magnified greatly to result in your standard 8x10. You have the same aberrations present in a lens intended for an 8x10 view camera but you aren't magnifying those aberrations so they are never apparent.
In addition, no matter what lens you are using, the number of pixels available to draw the scene is a limiting factor. A crop sensor might have 10000 pixels going across, but the final image merges 3 or 4 into 1 pixel. So that cable holding up a suspension bridge in the distance disappears as it is too small to show up on more than a single pixel. And you then get a bridge held up by a handful of random cables instead of a steady march the length of the bridge - some cables show up on multiple pixels and make it into the final image, some do not.
And as you resize the image to share on the internet, more cables are lost as the pixels representing them get dropped - you can't make a single pixel smaller, but the data it represents will be removed in the resampling.
However, if the same shot was taken on a larger format, each cable takes up more real estate on the sensor (be it film or digital) and so is maintained as the print is down-sized. This obviously has limits, as a larger digital sensor might just have larger pixels rather than a significantly higher number of them.
That said, an 8 MP full-frame sensor image should appear to have fewer apparent aberrrations than the 8 MP 1.6 crop sensor, even with the same lens. You are magnifying the crop sensor image much more, aberrations included. That's without even getting into the fact that to get the same coverage of the scene with the crop sensor, you need to be farther away, which means the lens simply cannot capture the same level of detail as it could at a closer distance. Also, atmospheric distortions will have a greater effect simply because there is more atmosphere between you and the subject, and you are magnifying them to a greater degree.
Last edited:
cfritze
Established
I like were this discussion has gone. I've learned a bunch from all the posts, even the "what an idiot" response from kuzano that recapitulates an earlier point from ferider. CK, dimitri and savage really hit on what I had in mind when I first asked the question. It's that lack of perfection and perception of dimensionality in an image that perhaps perversely I would look forward to exploring with full frame. Meanwhile, I still have much progress to make not to consistently take "woeful images" as referred to in another thread here on rff... Thanks all.