[ElJamoquio]

The last time someone challenged me with two terrible photos, they were both FF cameras.

I won't answer the SNR, because you've represented my position incorrectly. At the same DOF, with the same sensor technology and the same shutter speed, the SNR is the same.

My point is that you would NOT be able to tell the difference.

Extrapolating that posit to mean that, with slightly different equivalent iso's, someone can tell the difference between 99.5 and 99.8% fidelity (with different sensor technologies, different lenses, different AA filters, etc) is unrepresentative to the point of dishonesty.

---

Last edited by ElJamoquio; 05-08-2014 at 04:05 PM.

[bossa]

The last time someone challenged me with two terrible photos, they were both FF cameras.

I won't answer the SNR, because you've represented my position incorrectly. At the same DOF, with the same sensor technology and the same shutter speed, the SNR is the same.

My point is that you would NOT be able to tell the difference.

Extrapolating that posit to mean that, with slightly different equivalent iso's, someone can tell the difference between 99.5 and 99.8% fidelity (with different sensor technologies, different lenses, different AA filters, etc) is unrepresentative to the point of dishonesty.

Surely the magnification factor of the APS-C output makes noise more obvious? Yesterday I did an experiment with my D800e and my Tamron 24-70 and found that unless you're looking pretty close there's not a lot in it. Scaling the FF image down to APS-C would be an advantage over the full APS-C sensor I believe but things would only become really obvious at higher ISO's and also the quality of the lens. Output size would have a major influence on the results and I feel that not outputting to the same format size would definitely show some respect to the sensor size which many overlook.

[ElJamoquio]

Surely the magnification factor of the APS-C output makes noise more obvious?

If you're doing the test the way I suggested, there would be no noise difference. You're magnifying smaller pixels from a smaller sensor but taken at 100 ISO vs slightly larger pixels from a larger sensor at 220 ISO.

There is no difference in this case, if all else is held constant about the picture.

Yesterday I did an experiment with my D800e and my Tamron 24-70 and found that unless you're looking pretty close there's not a lot in it. Scaling the FF image down to APS-C would be an advantage over the full APS-C sensor I believe but things would only become really obvious at higher ISO's and also the quality of the lens. Output size would have a major influence on the results and I feel that not outputting to the same format size would definitely show some respect to the sensor size which many overlook.

I agree...

1) ISO 400, Tamron at 60mm, F/11
2) ISO 100, Tamron at 30mm, F/5.6, cropped to same fov.

1) will be sharper than 2), but noise will be the same and difficult to see in both images.


On the other hand...

3) ISO 12800, Tamron at 60mm, F/11
4) ISO 3200, Tamron at 30mm, F/5.6, cropped to same fov.

3) will be sharper than 4), but noise will be the same and easy to see in both images.

[Mock]

The last time someone challenged me with two terrible photos, they were both FF cameras.

I won't answer the SNR, because you've represented my position incorrectly. At the same DOF, with the same sensor technology and the same shutter speed, the SNR is the same.

My point is that you would NOT be able to tell the difference.

Extrapolating that posit to mean that, with slightly different equivalent iso's, someone can tell the difference between 99.5 and 99.8% fidelity (with different sensor technologies, different lenses, different AA filters, etc) is unrepresentative to the point of dishonesty.

Sorry, I did not completely understand your point. From my perspective when someone says f/2.8 is equivalent to f/1.8 (paraphrasing) it drives me a little mental because the increased noise on the APS-C chip is negligible to the point of insignificance nowadays. Allow me a rant, it is no longer directed at you individually...

I sell cameras in a camera shop, been doing it for 9 years now, so I tend to think things through to an end: How does this benefit the photographer?
Explaining to someone purchasing equipment that "x lens" on "x camera" will offer them a 1 + 1/3 stop advantage in shutter speed or ISO value (given comparable systems) generally goes over well. People understand how that will affect their photography.

Never in 1000 years would I attempt to talk someone out of a Sigma 18-35 f/1.8 on the basis of: "Well this lens isn't so special really, it's just like any 24-70 f/2.8 on a full frame camera. The DOF is the same because of the difference in focal length/aperture, and f/1.8 is not actually faster because of (insert paragraph here)..." ... "You say you like your crop camera? You don't plan on laying out $4k for the creme de la creme? Well I guess this lens will do for you."
The snark and condescension finds it's way in there naturally for the type of salesmen who would say such a thing, but worse yet, this salesman has likely gone way over the customer's head talking a bunch of techno-babble. No sale + customer will not come back!

Comparing anything to full frame and then crapping on it because it doesn't meet that special benchmark is something I wish would stop. There are comments all over photography websites stating things like: "The 02 Standard Zoom (5-15mm f/2.8-4.5) for the Q is only equivalent to an f/13-20 on full frame." SO WHAT!? This information is useful to if you a) care about how much DOF/bokeh you will have, AND b) have experience shooting with a full frame (or 35mm) camera to make use of that information properly.
It seriously confuses the casual observer, then they come into the shop and I've gotta pick up the broken pieces and mend their understanding.

I blame dpreview partially, ever since they started reporting aperture-equivalence in previews and reviews, the amount of misinformation and confusion has increased tenfold. There needs to be a better way to communicate the concepts we are discussing here.
If I might offer a solution from a marketing standpoint: each lens field of view in degrees should be displayed much more prominently (instead of using focal length and crop factors to eke out equivalent FL on FF - which is somehow supposed to communicate a FOV) , and, a standard measure of "blur units" (or something) is used to more concisely convey the DOF of any given lens on it's camera when wide open. A Blur Unit of 0 = a hyperfocal lens, and there would be no upper limit, but it could possibly be based around the amount of centimetres in acceptable focus with the lens set to 1 metre. Benchmark numbers would be set in the public eye by typically creamy lenses such as 85mm f/1.2s and 300mm f/2.8s.
Aperture is readily understood as a measure of light gathering (in terms of setting exposure values), so it stays. Focal length ÷ aperture = DOF is a difficult concept to grasp. Even for me, I tend to gloss over whenever someone mentions circle of confusion.

I'm not trying to say these discussions shouldn't happen, I just don't think it does much good to have them in the great wide open. This forum = fine, dpreview articles = please no.
You may now understand why I offered to leave the discussion earlier. The topic is a wee bit personal for me.

-----------------------------------------------------
Finally, the cameras in the test are as follows...

Upper Left - Pentax K-5 IIs @ 100 ISO
Upper Right - Nikon D4 @ 100 ISO
Lower Left - Olympus OMD E-M5 @ 200 ISO (base for that camera)
Lower Right - Sony NEX-6 @ 100 ISO

---

Last edited by Mock; 05-08-2014 at 07:32 PM.

[Edgar_in_Indy]

I feel like I need to do more to take this thread even further off topic while we wait for the next tidbit of information on the 18-35mm to dribble out. So in that spirit I'll ask, does dpreview use the same lens on all the DSLR's for shooting the still life? Because it seems like if they were using all different lenses, then a lot of the difference you will see will come down to differences in the optics. Not so much in noise, but in other aspects such as sharpness, colors, etc.

Ideally, they would use the same exact lens, and adapt it to different mounts whenever possible. Something like a nice Zeiss, or some other prime lens that is universally regarded as being excellent.

Or if that is not possible, then at least use the same model of lens, but in various mounts, with each lens being hand selected as an excellent specimen.

[Mock]

Here is a brief article that explains their newest set up a little bit. Their staff offer more insight into lenses and other topics in the comments.

[ElJamoquio]

The small size of the scene meant that some cameras, particularly those with fixed normal and wide-angle lenses, needed to be shot very close to the scene (often requiring a mini tripod to be set up inside the box it's housed in). Additionally, some modern compacts with large sensors and complex lenses looked terrible when shot at these close-quarters - meaning the scene didn't reflect their real-world behavior.
The three-dimensional nature of the old scene also meant that the scene had to be shot at very small (diffraction-limited) apertures to ensure all the key targets were acceptably sharp, yet still didn't have any corner detail within depth-of-field. These drawbacks, plus the need to revise the lighting of the scene, prompted us to move to a new scene around 7 times larger than the existing one.

...completely expected... with today's camera's it's not easy to test them to the limits with small targets. Imatest recommends over 40"x60" targets for something small like 24 MP or so.

[bossa]

If you're doing the test the way I suggested, there would be no noise difference. You're magnifying smaller pixels from a smaller sensor but taken at 100 ISO vs slightly larger pixels from a larger sensor at 220 ISO.

There is no difference in this case, if all else is held constant about the picture.



I agree...

1) ISO 400, Tamron at 60mm, F/11
2) ISO 100, Tamron at 30mm, F/5.6, cropped to same fov.

1) will be sharper than 2), but noise will be the same and difficult to see in both images.


On the other hand...

3) ISO 12800, Tamron at 60mm, F/11
4) ISO 3200, Tamron at 30mm, F/5.6, cropped to same fov.

3) will be sharper than 4), but noise will be the same and easy to see in both images.

The pictures I shot were:

D800e & Tamron 24-70 VC in FF and 1.5x crop modes.

FF....... 1/640 sec @ 35mm f/4 ISO 200
APS-C 1/640 sec @ 52mm f/2.8 ISO 100 (crop mode on the same camera with the same lens)

I was unable to really see much difference in this scenario, especially when the FF image was scaled down. I suspect the higher ISO's would be much more revealing. So I'll try it again with a more considered scene and lighting and with higher ISO's when I get a chance.

[ElJamoquio]

For the D800, you're comparing ISO 144 (indicated 200), which should give you 99.28% fidelity.
For the D800 cropped, you're using ISO 73 (indicated 100), which should give you 99.50% fidelity, but cropped by 1.5x, multiplying the 0.5% by 1.5x, giving you 99.25% fidelity.

Ignoring lens variation and measurement error (#'s taken from DxOMark) of course. Actually with a completely controlled experiment those numbers would be exactly the same, but even f/4 and f/2.8 aren't exactly the same; which is why I suggested a larger crop (2x) because it's easier to execute the experiment.

If you can tell the difference between 99.28% and 99.25%...

[LFLee]

I keep checking this thread to see if anyone have news about the availability of this lens....

but all I see is "I can't tell the difference of XXX, why bother to have XXX".
Strange, you see, I am a bit color blind (red & green), so I can't separate red & green very well, but I didn't request (or whine) camera company to make camera that don't record red and green, since "I can't tell them apart".

[ElJamoquio]

but all I see is "I can't tell the difference of XXX, why bother to have XXX".

Where do you see that? I see it nowhere.

[dbh]

I keep checking this thread to see if anyone have news about the availability of this lens....

but all I see is "I can't tell the difference of XXX, why bother to have XXX".
Strange, you see, I am a bit color blind (red & green), so I can't separate red & green very well, but I didn't request (or whine) camera company to make camera that don't record red and green, since "I can't tell them apart".

I keep checking this thread for the same reason. I suspect when there is some news on the availability of this lens, or it actually starts showing up somewhere, the thread may then get back on topic.

[Kozlok]

Since we're so off subject, maybe we should talk about turtles. Their eyes have very limited depth perception, but a really wide FOV. They have a FF equivalence of 12mm at f1.2. Except none of that is true. But it is every bit as relevant as the rest of this thread. Or maybe not.

I for one am excited about this lens, and if someone wants a FF 24-70, go for it. I don't want a FF camera. I actually prefer an APS-C, but really want a wide, fast lens. None of that FF equivalence stuff means any more to me than my turtle comments.

[dadipentak]

Since we're so off subject, maybe we should talk about turtles. Their eyes have very limited depth perception, but a really wide FOV. They have a FF equivalence of 12mm at f1.2. Except none of that is true. But it is every bit as relevant as the rest of this thread. Or maybe not.

I for one am excited about this lens, and if someone wants a FF 24-70, go for it. I don't want a FF camera. I actually prefer an APS-C, but really want a wide, fast lens. None of that FF equivalence stuff means any more to me than my turtle comments.

Being a turtle, I couldn't agree more!

On a side note, we're getting close to the first anniversary of the start of this thread (and the first anniversary of my order is just over the horizon) but, then, since I'm a turtle, speed is not my top priority.

[ElJamoquio]

Since we're so off subject, maybe we should talk about turtles. Their eyes have very limited depth perception, but a really wide FOV. They have a FF equivalence of 12mm at f1.2. Except none of that is true. But it is every bit as relevant as the rest of this thread. Or maybe not.

I for one am excited about this lens, and if someone wants a FF 24-70, go for it. I don't want a FF camera. I actually prefer an APS-C, but really want a wide, fast lens. None of that FF equivalence stuff means any more to me than my turtle comments.

I want the best camera that suits my needs. To compare two cameras I have to trust the following sentence: "Multiply advertised aperture by crop factor, mulitply advertised FOV by crop factor, multiply advertised ISO by crop factor".


That didn't take long at all.