when does newer DX outperform older FX?

[pixmedic]

snapsort, DPreview, and other review sites aside...

I was just curious....at what point do NEWER DX cameras start to outperform OLDER FX cameras? (mostly in terms of IQ, low light, etc etc)
for instance...Nikons D7100. great DX camera. love it. obviously however, not as good as nikons entry level FX D600 in terms if IQ and low light performance. BUT, how far back do you have to go in FX DSLR history before the D7100 will give you a better image and/or low light performance than the FX camera?

I have heard on several occasions people suggest getting an older FX camera over a newer DX camera. so, how much older on the FX DSLR end before you lose the FX advantages to newer DX tech?

 

[cgipson1]

D7100 beats the D3 in everything except low light capability... and totally whups the D2 series... is this sort of what you are after?

DxOMark - Compare cameras side by side

 

[Derrel]

I think the D7100 easily beats the Canon 5D Classic in both resolution and color depth and dynamic range. Heck, a LOT of newer cameras have better dynamic range than the 5D Classic. The 5D is about the same in image quality as the original D3. The D3s sensor has better low light performance than the D3 or 5D Classic. Scott Kelby said the D7100, after three days of shooting indoor hockey with the new 80-400 G-series lens,that the D7100 is about equal to the D4 up to around 4,000 ISO, if I recall correctly, but that the camera is a little bit sluggish compared to a D4.

I shot the D2x and the 5D Classic simultaneously for about seven and five years, respectively; the D3x whips both those cameras in every measure possible. The sensor in the D600 is a bit better than the one in the D3x. I do not think there is yet any DX-sensor camera that can outperform any of the 24-MP NIkons or the 36-MP D800, in any category. Sony's new 24-MP FX camera has a great sensor too..it's right up there with the D600.

DxO Mark has probably the largest and most rigorous comparison database. Their database ought to be able to show numerical values for the metrics people wonder about. Some people worry more about Low Light/High ISO performance, while others worry more about dynamic range, and some about color depth and richness.

 

[Derrel]

I JUST STOPPED BY DXO Mark and compared the D7100 versus the D3 versus the Canon 5D.

The D7100 has wayyyyy better DR than the Canon 5D--about 2 and 3/4 EV more DR for the D7100. it's about 1.5 EV better than the D3 is.

The D3 is still much better than either at High ISO noise performance.

DxOMark - Compare cameras side by side

 

[pixmedic]

It is pretty interesting. Guess the D3 is still a formidable camera.

 

[SCraig]

I JUST STOPPED BY DXO Mark and compared the D7100 versus the D3 versus the Canon 5D.

The D7100 has wayyyyy better DR than the Canon 5D--about 2 and 3/4 EV more DR for the D7100. it's about 1.5 EV better than the D3 is.

The D3 is still much better than either at High ISO noise performance.

DxOMark - Compare cameras side by side

Did you by chance compare the D7000, D7100, and D600? Other than low-light the numbers are somewhat interesting. The low-light numbers are pretty much what I expected, but the dynamic range and color depth weren't.

 

[pixmedic]

I JUST STOPPED BY DXO Mark and compared the D7100 versus the D3 versus the Canon 5D.

The D7100 has wayyyyy better DR than the Canon 5D--about 2 and 3/4 EV more DR for the D7100. it's about 1.5 EV better than the D3 is.

The D3 is still much better than either at High ISO noise performance.

DxOMark - Compare cameras side by side

Did you by chance compare the D7000, D7100, and D600? Other than low-light the numbers are somewhat interesting. The low-light numbers are pretty much what I expected, but the dynamic range and color depth weren't.

I did. low light performance aside, (FX obviously having the advantage) the rest of the numbers were closer than i was expecting.
DxOMark - Compare cameras side by side

I realize of course, that this is only comparing the sensor. a lot of other factors can influence someones decision for buying a camera.
still, for what I do, I am happy with the decision to get the D7100. maybe next year i will make the jump to FX and grab a used D600.

 

[Derrel]

The "new-sensor" Nikons have set the benchmark for performance. The once-dominant, utterly superior performance of FX over DX (APS-C) is now no longer the way things "used to be". Nikon has moved to newer, better sensor technology across its entire line, and has moved DX camera sensor performance much higher than it used to be. Canon on the other hand, has been stuck with the same 18 MP APS-C sensor since 2009, made using outdated .5 micron fabrication technology.

FX still has an advantage in High ISO performance, but now the dynamic range and color depth of the smaller-sensor Nikon cameras is simply amazing. People shooting D90's and older cameras of other makes probably have no idea about the actual highlight recovery and "malleability"of Nikon RAW files coming out of these new-sensor Nikon cameras.

 

[Antithesis]

The 5D and D3/D700 are not even in the same league in terms of IQ, as the 5D is a much older camera. Any new DX or MFT camera is going to trounce the old 5D in the IQ dept. I owned and shot on one for a couple years, and newer cameras are substantially better. It was a fine camera as long as the scene isn't overly contrasty or you didn't need to shoot at ISOs higher than 1250 or so.

The X-factor on full frame cameras is not image quality alone, as there are a lot of options these days, but more control over DoF. The DoF at f1.8 on a DX sensor will look like F2.8+ on a FX sensor. Getting an image that is really sharp but maintains the sort of separation that can really make an image pop is tough on a DX sensor. Your options are limited to the really fast L primes, and they are often not terribly sharp at F1.2 (not to mention heavy as hell and slow to focus in the case of the 85L). If you need to stop down to get sharpness out of your lens, you lose a lot of your subject isolation. For DoF alone, I'd probably take a 5D over any DX camera.

Also, in staring at an enormous amount of images, it seems like the FX cameras (even the old 5D) have a lot nicer looking color than their counterparts. Adjusting the WB on a 5Dmk2 or a D700 is going to give better results than doing the same on a DX camera. I don't know why it happens, but the DX images always look just a touch weird (light color casts in the shadows, etc.). This effect is multiplied at higher ISOs.

On a side note, I think the D3X, D600 and Sony's 24mp FX camera all share the same sensor but are just processed in slightly different ways. In camera processing has a pretty enormous effect on image quality. I don't think the D3S had a radically different sensor than the D3 (if it was different at all), but gained a full couple stops in low light.

 

[Antithesis]

Oh, and on another side note, DXOMark seems a little more than biased towards Nikon. I shoot Nikon, I love Nikon, but I have a hard time believing that the IQ on the D600 scores a 15%+ better "score" than the 5Dmk3 (a fantastic camera by all accounts), and that a 24mp DX camera does better as well. The scoring system just seems like an arbitrary way to say "X is better than Y, just let us handle the science".

 

[Derrel]

Oh, and on another side note, DXOMark seems a little more than biased towards Nikon. I shoot Nikon, I love Nikon, but I have a hard time believing that the IQ on the D600 scores a 15%+ better "score" than the 5Dmk3 (a fantastic camera by all accounts), and that a 24mp DX camera does better as well. The scoring system just seems like an arbitrary way to say "X is better than Y, just let us handle the science".

Again...you're apparently just looking at "megapixel" counts and ignoring sensor fabrication technology and image processing differences between Canon and Nikon cameras. You've seen the shadow noise problems the 5D Mark III has when "lifting the shadows"--even at base ISO, right? Chroma noise galore from the 5D-III sensor...

Part II - Controlled tests

Look at how BAD the 5D-III sensor performs. And this is at BASE ISO...

Did you see that DxO Mark compared the same 72 lenses on the D600 and D800, and noted only a 12% increase in resolution for the D800, despite the ~30% larger file size?

 

[Garbz]

Just to throw some additional spanners into the it must be the sensor only thought: When I started astrophotography I used a D200. Then I upgraded to a D800 and started using the D800, got a massive improvement in noise. Then I upgraded my D800 to a QHY10 (astronomy sensor) and that was a quantum leap improvement in noise over the D800. The QHY10 and the Nikon D200 have the same Kodak sensor. Yet for long exposure noise wise one way outperforms the D800.

It's not just about the sensor.

...

Though these days with CMOS being digital devices it's mostly about the sensor, but a 6+ year old CCD using in an appropriate way with modern ADCs can still have better noise performance in certain scenarios than a modern top of the line CMOS sensor.

 

[KmH]

It's not just about the sensor.

+1

The high ISO gains in recent years have primarily been accomplished with camera image processor software.
When more pixels are packed onto a same size image sensor each pixel is smaller, and each pixel's SNR (signal to noise ratio) gets worse.

DXO Mark uses industry standard testing protocols.

Sony makes Nikon's image sensors.

DxOMark - Pushed ISO: Let's make it clear
DxOMark - SNR evolution over time

 

[Tinderbox (UK)]

Is back-light illuminated, and micro lenses part of the reason the new dx sensors beating the older fx ones.

John.

 

[Garbz]

No. Well not totally anyway. Those technologies have been around for 5-ish years and definitely can't be attributed to the sudden amazing advances we've seen in the past year or so. The 50D had microlenses, as did the 5D mkII, both cameras have been left in the dust in terms of modern quality metrics.