full frame vs cropped sensor

So if a full frame sensor has pixels 8.2 x 8.2 microns (The 5D) and a D2x has pixels 5.5 microns then the area difference is 67.2 square microns vs 30.25.. That means that according to the above, the 5D pixel will get 670 photons and have a S/N ratio of 134:1

<{POST_SNAPBACK}>

Well, that scaling works only if we hold everything else constant -- S/N depends on sensor details and camera electronics, right?

Otherwise, we would have to conclude that because the D70 (23.7 by 15.6 with 6.1Mp for about 61 sq mics) would have better S/N than the 20D (22.5 by 15 with 8.2 Mp, 41 sq mics) and the D2X (30), and virtually the same as the 5D (67 vs 61 for the D70).

As a D70 shooter, I am painfully aware that it simply ain't so. Canon gets a better S/N out of a smaller sensor with a higher MP count, comparing the 20D to the D70. Superior engineering can break this scaling law -- and a truly superb small sensor can have better S/N than a larger sensor (again, with different materials and electronics).

Calculation is like masturbation. If you do it long enough, you start to think it's the real thing. Or so I hear.

I think I will go take pictures of ibises with my teeny sensor but oh-so-big 80-200 f2.8!

Cheers,

Chris

Alright, I know nothing about photons and how they influence resolution (my PhD is in biology!). So, let's make this really useful, here is what I gather from all of the above:

1) The main disadvantage of a 24mm wide (a.k.a. crippled) sensor is diffraction limited resolution. In other words, in small apertures a full frame sensor has theoretically higher resolution.

2) The main disadvantage of a 35mm wide (a.k.a. full frame) sensor is vignetting/falloff. The edges of the lenses produce lower quality image, which are only apparent on 35mm sensors because they are cropped out on 24mm sensors.

My request is, give me one REAL WORLD example of one of each of the above. One example in which you had to increase you aperture in a 24mm sensor camera because of diffraction and one example of light falloff that really bothers you. Any lens, any camera. But it has to be real world, better if it is underwater.

Let the games begin...

Luiz

I think the general point still stands Chris - the area of the sensor is very important and bigger is better when it comes to S/N.

Cheers

James

Here's an interesting and very applicable quote by Thom Hogan...

<{POST_SNAPBACK}>

Wow, that's fascinating stuff, James! I had never read/heard anything that specific before - thanks for posting it!

eh I got lost somewhere at the beginning???

Q. Can you still take photos? with "full frame sensor has pixels 8.2 x 8.2 microns (The 5D) and a D2x has pixels 5.5 microns then the area difference is 67.2 square microns vs 30.25.. That means that according to the above, the 5D pixel will get 670 photons and have a S/N ratio of 134:1"

ps. I found this thread in the Beginners Forum

Alright, I know nothing about photons and how they influence resolution (my PhD is in biology!). So, let's make this really useful, here is what I gather from all of the above:

 

1) The main disadvantage of a 24mm wide (a.k.a. crippled) sensor is diffraction limited resolution. In other words, in small apertures a full frame sensor has theoretically higher resolution.

 

2) The main disadvantage of a 35mm wide (a.k.a. full frame) sensor is vignetting/falloff. The edges of the lenses produce lower quality image, which are only apparent on 35mm sensors because they are cropped out on 24mm sensors.

 

My request is, give me one REAL WORLD example of one of each of the above. One example in which you had to increase you aperture in a 24mm sensor camera because of diffraction and one example of light falloff that really bothers you. Any lens, any camera. But it has to be real world, better if it is underwater.

 

Let the games begin...

 

Luiz

<{POST_SNAPBACK}>

Re: #1 : Here's the thread where I tried to show Rand that his shots were diffraction limited.

http://wetpixel.com/forums/index.php?showtopic=9568

I don't know of any carefully done side-by-side comparisons. Diffraction has been well understood for a very long time. I would be surprised if people still spend much time testing its effects.

Just back from a month's trip. We (my wife and I) shot around 5~6000 images on FF Canons. My wife was using her camera for her second trip with it and has found the crossover from a 35mm film Nikon to be very straightforward. This is a useful difference in itself!

You can argue the detail as much as you want to, but my fundamental impressions of the differences are/is that the FF Canon's offer more and faster wide-angles than the cropped sensor cameras!

Other than that, when you get to 10MPixel plus, the results that I have seen from both sensor sizes have been excellent. In the real world there is not that much difference between final images from such sensors, and what differences (and there are a few) there are requires a 'trained' eye to appreciate. No doubt image quality will continue to improve but in all honesty we now have very effective versatile tools for creating excellent images with.

Paul is 100% correct. The more I shoot my D2x, the more I'm convinced this may be the last camera I ever "need." I tend to shoot it at wider than expected apertures, even for macro, in order to get extreme isolation of my subjects. The diffraction argument is tenuous IMO because smaller sensors give a larger DOF anyway, so tiny apertures are not necessary as they are in film/FF shooting.

Like all aspects of photography, there is a balancing act between DOF and diffraction. Interestingly, FF advocates criticize the larger DOF of cropped sensors when discussing portraiture. But in the case of macro, FF advocate criticize that the cameras become diffraction limited. If I get the DOF of f/22 at f/16 (and I am totally making an example because I don't know the actual equivalents), why would I shoot at the smaller aperture?

Besides, prior to the dSLR days I asked numerous prominent uw photographers, many on them teh gurus, whether they worried about diffraction when shooting macro. They unanimously replied no - that for extreme macro they maximized the DOF by shooting at the smallest aperture their stobes would allow. If it works for them it is good enough for me.

To quote Rodney.....

"Can't we all just get along?"

Hi Kasey,

RE:

Besides, prior to the dSLR days I asked numerous prominent uw photographers, many on them teh gurus, whether they worried about diffraction when shooting macro. They unanimously replied no - that for extreme macro they maximized the DOF by shooting at the smallest aperture their stobes would allow. If it works for them it is good enough for me.

That's because it wasn't really an issue in the film days. The problem is that pixels are split up into the bayer pattern on a digital sensor. It's not a good thing when light diffracts at the aperture and light from the subject is spread out and the red hits where it's supposed to (on the sensor) but the blue hits an adjacent pixel instead. This is more pronounced the smaller the pixels get.

Cheers

James

If you think of FF as replacing 645 and the reduced format equating to 35mm then this is a very rough guide to how the two systems work in film terms (~ish!!!).

BUT and it really is a big BUT, you won't really see much difference unless you are producing significantly large final output. I wonder how often this really does become an issue? The higher MPixel cameras produce superbly smooth tonality, excellent colour rendition and nuances of differences which a good observer will see. The real world dictates that many other factors play a larger role - subject matter, point of focus, etc, etc. I happen to like and prefer FF but can see advantages to the reduced format too. Diffraction MAY be an issue, but its not one that I've heard anyone complain about on grounds other than theoretical.

...except that you can't have it both ways. You can't justify needing a 10+MP camera and then say that in practice diffraction doesn't matter because, with a 1.5x crop sensor, sharpness is so bad at f/22 that you may as well have a 5MP camera. If you think you need a 10+MP camera, then for the same reason you are limited to shooting at f/11 or wider (on 1.5x crop cameras).

If anyone is interested in seeing examples, there's a good discussion of diffraction with examples here: http://bobatkins.photo.net/photography/tec...iffraction.html.

Lanier

To quote Rodney.....

 

"Can't we all just get along?"

<{POST_SNAPBACK}>

We have these types of discussions all the time: e.g. bigger dome vs. smaller dome, longer strobe arms vs. shorter strobe arms, split fins vs jet fins, even mac vs. pc.

I don't see a problem.

Like all aspects of photography, there is a balancing act between DOF and diffraction.  Interestingly, FF advocates criticize the larger DOF of cropped sensors when discussing portraiture.  But in the case of macro, FF advocate criticize that the cameras become diffraction limited.  If I get the DOF of f/22 at f/16 (and I am totally making an example because I don't know the actual equivalents), why would I shoot at the smaller aperture?

 

<{POST_SNAPBACK}>

As I showed in the link I gave above, F/22 at FF has slightly better diffraction limited resolution than F/16 on a 1.5x cropped sensor and about the same DOF.

Besides, prior to the dSLR days I asked numerous prominent uw photographers, many on them teh gurus, whether they worried about diffraction when shooting macro.  They unanimously replied no - that for extreme macro they maximized the DOF by shooting at the smallest aperture their stobes would allow.  If it works for them it is good enough for me.

<{POST_SNAPBACK}>

If you think that's the correct approach then you should get the camera with the best high ISO performance so that you can shoot at the smallest aperture for the strobes you are willing to use for macro.

I agree with Herb, these types of exchanges sometimes are biased toward one side or the other, but the information is enlightening and I enjoy seeing the point counter point. I'm not as well versed as some so it's educational.

Although I'm not ready to agree with Lanier's comment about cropped sensor sharpness. Not that he may not be correct, but until someone posts a FF sensor macro shot of a similar subject to the ones I've been shooting that shows the difference.

Rand

Hi Rand,

The only people I know well who have shot FF underwater are Eric, Bruce, and Paul (I don't really know them THAT well though).

Now that there are some 5D's hitting the water, we should be seeing some super macro posted here soon. Of course my first FF trip will be all wideangle...:-)

Cheers

James

Hmmm.

Problem is that my 1DS is perfectly capable of producing A0 prints (I've just sold some to a very happy client). At this size focus (and so DoF) on the 1DS is absolutely critical and shooting for example a lobster means that even at f/11~14 you can focus between the eyes but not have both perfectly sharp on a 100mm lens, but few people will notice this I assure you.. Post processing has a lot to do with the final image quality and I'm rapidly coming to the conclusion that the debates are interesting but reality is far more so. I've said this before and I'll no doubt say it again, at 10MPixel+ we have better than 35mm film cameras. Discussing the nuances of diffraction limitation and other technical aspects is fine but in underwater there are other limiting factors such as shooting through thick glass, conventional domes, and finally water less clear than ity could be. I would say that if you have a 10MPixel+ camera then the more important aspects of underwater photography are those involving lighting and quality of lighting, ergonomics of the housing and viewfinder systems (Okay I may be biased here) and lastly but by far the most important bit, being underwater and shooting pix!!!

You can't justify needing a 10+MP camera and then say that in practice diffraction doesn't matter because, with a 1.5x crop sensor, sharpness is so bad at f/22 that you may as well have a 5MP camera. 

 

Lanier

<{POST_SNAPBACK}>

Ok, this is the type of argument that I don't understand. According to the above, I should just get a D50 and give my worthless D2x away. Yet, here is what I get with it:

Full frame:

100% crop:

Lame photo, I know, but it was the only one I found that I photographed at f32. Yes, the above was photographed at f32 with the 105 (not my sharpest lens). Now, I don't see any diffraction, and, as far as I know, no 5mp camera can produce pictures like that.

Also, correct me if I am wrong, but to get the same photo with the same lens in a full frame camera I would have to be closer to the subject, and would have to close the lens more to get the same DOF, thus increasing diffraction, right?

Luiz

P.S.: No PP except for auto-levels.

I don't have shots at F/32. Here's one with similar features shot at F/22 with the Canon 100 and a 1.4x teleconverter with a 6M pixel DRebel. The teleconverter makes the actual aperture F/32 even though the camera was set at F/22.

Shot in raw format. Convertered in C1 with all setting at zero except for a slight exposure compensation. No sharpening.

Here's the scaled down uncropped shot:

100% crop

100% crop scaled up 1.4x for resolution comparison with a 12M pixel output. Not sharpend which put it at a slight disadvantage since the interpolation makes it look less sharp.

Here's the same crop sharpend: radius 2, 100% , threshold 0

You be the judge.