Is DX superior to FX?

GreenFlash

The discussion has gotten a little distance from the original question. Meantime I found a site called Cambridge in Colour, where there are some awesome tutorials on sensors and how they work. The material is extremely useful for this discussion, so I am presenting what I have gotten so far from this site,about the technical reasons fx is still better:
1. In low light, what damages images the most is noise. Noise is not from extraneous ambient light but rather is the result of false signals generated by the electronics of camera sensors, and also variations in how individual pixel sensors operate, that can create "hot spots". In general, noise is consistent across a camera's sensor.
2. What is important is the level of noise compared to the strength of the light the camera is receiving which is called the "signal". The stronger the signal is, compared to the noise, the less the noise impacts the quality of the picture.
3. Now here's the big one: the larger individual pixel sensors are, the more light they receive, but the level of noise does not increase with sensor size. Thus, the signal to noise ratio is much better in a 24 mp image from an fx sensor, than from a 24 mp image from a dx sensor, because the individual pixel sensors are getting more light. The worse the light conditions are, the worse will be the signal to noise ratio.
4. The way ISO works is essentially to put an amplifier on a light signal from individual pixels, when the light signal is too weak. Sad to say, the amplifier amplifies noise just as well as it amplifies the light signal, so that is why noise is so much worse at high ISO's, Further, noise is worse, the longer the shutter has to be open to allow in more light. These two basic facts explain why larger sensors with larger pixels have a much higher signal to noise ratio in very low light situations, and therefore have better, lower noise images.
So here are a couple of strategies that work: Dark images have more visible noise than light ones. So in very low light situations, it's better to shoot lighter images and darken in post processing, than the other way around; and (2) it's better to shoot more open with faster shutter speeds to reduce noise.
There's a lot more on the Cambridge site than just these things, that support dx in some circumstances and fx in others. But this alone seems to answer why it is, even with equally clear, equally fast dx lenses, in low light, fx will still be much better.
I hope I got this right.

Msmoto

@GreenFlash

I think you have this right… IMO. Especially in low light, any underexposure, even a half stop will kill shadows and increase noise, in my experience. I regularly shoot up to 12,800 ISO, no worries unless I miss the exposure (which I have been known to do.. LOL)

DenverShooter

This discussion reminds me of what I was told about 50 years ago by Gerhard Bakker…."They tried so hard to be different they forgot to be good." And, in some ways our preoccupation with the gear gets us sidetracked.

An example… this weekend I was shooting BIF, and in a venue I had no idea how it would be set up until I was there…. Well, all my fancy gear got me very few images as I could not get the camera set to grab the rapidly moving, variable subject….. I finally tried AF-C, with 3-D tracking, and maybe this is what works. I am going to see how this goes with BIF near the ocean in a week or so.

I think I like the full sensor as I am a bit rusty and not able to track the subject as well as I used to. Also, my objective on this first shoot at the Carolina Raptor Center was to learn about the venue and when I return in the fall, I think I might get some good shots.

If I had shot this weekend on my D90, I am of the opinion the difference in the results would not be noticeable in the sizes on Flickr. Maybe if I printed to about 24" x 36" the different size would mean something….but the overall image requires a lot more than simply using a pro level camera, crop sensor or full.

The Nikon tech folks told me to not use 3D for BIF on my D4 or D800E...

Denver Shooter

SquamishPhoto

While shooting in AF-C its also important to chose the appropriate focus tracking setting for the type of flight path you are attempting to shoot. It also helps to reduce the number of AF points down to as few as possible so that accuracy is increased. The 52 point system is not designed for capturing fast movement.

sevencrossing

The discussion has gotten a little distance from the original question. .

+!

snakebunk

Well, the answer to the original question is no.

Ironheart

I thought it was yes!. Actually the answer is "it depends".

GreenFlash

@ironheart
Amen. Clearly for low light, the answer is fx. But personally I have found that in general use, as long as there is moderate light or more, the quality difference between my Nikon D600 and D 5300 is indistinguishable.
One of my questions was whether using lower resolution on the same camera, would give better low light results. I did some admittedly limited tests shooting the same indoor scene with the same light with my D5300 at its highest, mid and low resolutions. The result was that in each case the camera used the exact same ISO, f setting, and shutter speed. The higher resolution images were noticeably clearer and less noisy. That was consistent with one of the other Cambridge comments, that higher resolution on the same senor, will tend to reduce the appearance of noise because the noise appears in pixel size, so the noise appears finer and more evenly spread in higher resolution images on the same sensor. So on the same camera, shooting in low light, higher resolution is better.
But if you read through all of the comments, and thank you everyone, there are situations in which the smaller dx sensors are better.
In one particular use, I need as much depth of field as I can get, and my D-5300 is significantly better than my D600 for that. In uses where pixel density is needed for example macro shots of small objects, or telephoto shots, dx can be clearly superior as long as the lens can deliver the detail necessary to use the pixels.
There are of course the practical size and weight issues, and economic issues, but these do not bear directly on iq.
What I appreciated about the discussion is that we created almost a catalog of the various situations in which there are advantages to one or the other. It certainly has clarified my thinking as to which camera and lens I should use, in which situation, and why. Thanks everyone.

Msmoto

Thanks, guys, the BIF maybe needs a special thread….

PB_PM

I thought it was yes!. Actually the answer is "it depends".

Heavily weighted on the "it depends." For several reasons, besides image quality. If all we are talking about is pure image quality the two systems are very close at the end of the day, no questions asked. The thing is, there are other factors.

Nikon, in it's great wisdom (not), seems to keep the best stuff for the top tear cameras (D4/D4s/D800(e) ). For example, even though the D7100 has the 51 point AF system, it just doesn't seem as good as the module in the D800/D4. For most people the difference isn't noticeable, but it is there and it affects focus performance whether it be a mid-day or at dusk. For still life and landscape type work this is totally irrelevant, but if you shoot any kind of action it shows up. Of course the modules could actually be the same, but because more light comes through the FX circle, more light reaches the AF sensor, enabling it to focus fast/more accurately (purely speculation here).

Viewfinders. For me this is a big deal, and I find DX viewfinders painfully small. I get eye fatigue much faster when I use a D300 (yes I still have one, it's a nice paper weight) than with the D700 or D800. I cannot think of any technical reason for this, because DX cameras tend to have better magnification, which one would think would lead to less eye fatigue.

manhattanboy

The higher resolution images were noticeably clearer and less noisy. That was consistent with one of the other Cambridge comments, that higher resolution on the same senor, will tend to reduce the appearance of noise because the noise appears in pixel size, so the noise appears finer and more evenly spread in higher resolution images on the same sensor. So on the same camera, shooting in low light, higher resolution is better.

This is only true if you size the images the same (i.e. take a 36mp FX and a 24mp FX and force both of them into the same X inches by Y inches print).

If you look at the image at full 100% magnification (which unfortunately I do), then the increased resolution leads to noticeably more noise because there are less photons hitting the smaller pixels. Ironically, there is also the potential for more blur at 100% magnification as the resolution of the high MP sensor needs to be matched by the resolution of the lens, which does not happen for cheaper lenses (or probably even the high priced, just released 18-300 DX superzoom). Hope that helps!

PB_PM

This is only true if you size the images the same (i.e. take a 36mp FX and a 24mp FX and force both of them into the same X inches by Y inches print).

Not to mention the difference in generations of cameras. The D3x is noticeably noisier than the D600/D610, even though they are both 24MP cameras.

GreenFlash

@PB_PM and Manhattanboy:
Actually, Cambridge's observation only related to images from the very same sensor, not to two separate sensors, because as you have accurately noted: some sensors have more noise (read electrical interference) than others. What Cambridge was saying is that on the same sensor, noise does not occur in every pixel. but only a certain amount of it occurs on any one sensor total. So, if you have smaller pixels, the appearance of the noise will be smaller because the pixels dominated by noise will be smaller.
In this, I am not saying that what they are saying is true, just trying to accurately report what they say.

sevencrossing

If the OP question was

" In good conditions is Dx nearly indistinguishable. from FX?" The answer might be yes

If the OP question was

"Are there occasions, on a limited budget, when you might be better off with Dx rather than FX?" The answer might be yes

but the answer to the OP question "Is DX superior to FX?" is No

manhattanboy

@PB_PM and Manhattanboy:
Actually, Cambridge's observation only related to images from the very same sensor, not to two separate sensors, because as you have accurately noted: some sensors have more noise (read electrical interference) than others. What Cambridge was saying is that on the same sensor, noise does not occur in every pixel. but only a certain amount of it occurs on any one sensor total. So, if you have smaller pixels, the appearance of the noise will be smaller because the pixels dominated by noise will be smaller.

In quoting the Cambridge site:

You can also see that increasing ISO speed always produces higher noise for a given camera, however noise variation between cameras is more complex. The greater the area of a pixel in the camera sensor, the more light gathering ability it will have — thus producing a stronger signal. As a result, cameras with physically larger pixels will generally appear less noisy since the signal is larger relative to the noise. This is why cameras with more megapixels packed into the same sized camera sensor will not necessarily produce a better looking image. On the other hand, a stronger signal does not necessarily lead to lower noise since it is the relative amounts of signal and noise that determine how noisy an image will appear. Even though the Epson PhotoPC 800 has much larger pixels than the Canon PowerShot A80, it has visibly more noise — especially at ISO 400. This is because the much older Epson camera had much higher internal noise levels caused by less sophisticated electronics.

This is basically more eloquently saying what I and PB_PM were saying.
There are many types of noise, so your reference could be talking about a specific subset that is easier to correct with more pixels such as fixed pattern noise.

TaoTeJared

But if you read through all of the comments, and thank you everyone, there are situations in which the smaller dx sensors are better.
In one particular use, I need as much depth of field as I can get, and my D-5300 is significantly better than my D600 for that. In uses where pixel density is needed for example macro shots of small objects, or telephoto shots, dx can be clearly superior as long as the lens can deliver the detail necessary to use the pixels.

It makes me very uncomfortable when you use the phrase "clearly superior" and then state that DX bodies are better in some situations. To be clear, DX is never actually superior to FX in anything I have ever seen or experienced and the best IQ will come from FX sensors in every situation making everything equal.(equal lens length, iso, settings, etc.) The trade off of DX in macro, extreme telephoto shots is minimal.

Where the Trade offs and tipping points exist:

-ISO: This is the #1 consideration of using DX vs FX. DX Vs FX trade off is minimal until you move to ISO 400 and above. Then FX consistently outpaces DX and even FX crops can look better than un-cropped DX images at higher ISOs. This is the major disadvantage of all smaller sensors. The smaller you go, the IQ drops dramatically as the ISOs move off of Native settings. (Native setting is usually 100 or 200 with DX & FX. Smaller sensors are sometimes as low as 50.)

-Detail: Honestly FX always seems to have more detail to my eye. The argument of DX "density" is better, needs a major qualifier (that is almost always not included) - it assumes FX is cropped to the same field of view. A DX body with a 200mm lens and a FX body with a 300mm lens essentially have the same density. If you are shooting 24mp bodies with both, the density is the same. In this instance though, I generally can see more detail in FX. I did a ton of comparisons with a D300 and D700 with equivalent lenses (length and build quality), and the D700 always pulled more detail out of an image. I do believe it is due to the larger pixels of FX gathering more accurate color (hence detail) for the image.

-Cropping: DX (at native ISO) will perform better than FX that is cropped to the same field of view.

-DOF - Macro: DX has more DOF than FX, but not much and in practice (because of available focal lengths one may own) basically has no advantage and could be a disadvantage.
Example:
DX with 40mm macro @F/4 @3ft has a DOF of 3.12 inches.
FX with 60mm @F/4 @3ft has a DOF of 2.04 inches.
For macro work if you want a FOV of a 60mm lens, DX has more DOF.
But if you just own a 60mm...
DX with 60mm @F/4 @3ft has a DOF of 1.32 inches.
Clearly if you own only a 60mm, the DOF is almost half of a FX body. Also you have changed the FOV and look of the image from a "normal" (close to what the eye sees) to a telephoto that compresses (enlarges) the image. That is personal taste of what you like but it is a consideration.

spraynpray

Sorry Tao, but there are two sides to every coin - you have simply stated those that support your gear. When is too much DoF ever a problem in macro? If FoV is a concern, move closer/further away and if you factor how working distance changes between DX/FX into your macro example, it changes things even more.

Macro and long lens work is DX territory IMHO.

donaldejose

I have 40mm, 60mm and 100mm macro lenses. In my experience DX beats FX in macro when shooting insects because it gives you a greater working distance. I also "top out" telephoto at 300mm f4 to keep costs reasonable: no $5,000 plus super telephotos for me (If I felt I could spent that much money on one piece of gear I would buy a D4 or D4s which would be used far more than a super telephoto). For birds DX beats FX because it allows me to use my 300mm as a 450mm. Thus, I agree with spraynpray: DX beats FX for macro and long telephoto work. When Nikon finally produces a D300s replacement which puts more pixels on the subject than does the DX mode of the D800 it will be interesting to see how many wildlife shooters use it. Will Moose Peterson add one or will he keep shooting his D4s with his super telephotos?

MikeGunter

Hi all,

The original post was, in a word, odd. "Superior"? On the pocketbook, yes. Stand off, yes?

One can cherry pick their own reasons to fit a 'superior' champion for DX - I'm not playing that game - it's plain silly. I have several DX cameras and use the format every day.

FX offers more IQ for bunches of reasons and hits the pocket book for all those reasons.

My best,

Mike

sevencrossing

OED Superior Greater in size or power:

manhattanboy

OED Superior Greater in size or power:

Sorry for the stupidity but am having a senior moment and can't recall what the camera "OED" is.
Even the Oxford English Dictionary was of no help. :-O

heartyfisher

OED Superior Greater in size or power:

Sorry for the stupidity but am having a senior moment and can't recall what the camera "OED" is.
Even the Oxford English Dictionary was of no help. :-O

LOL !

TaoTeJared

Sorry Tao, but there are two sides to every coin - you have simply stated those that support your gear. When is too much DoF ever a problem in macro?

I'm afraid you read an incorrect perception into my post and I didn't mention anything about my gear (which is both DX and FX) nor anything about too much DOF. I actually prefer DX for macro work and love that DX gives the extra reach at a lower price. Those are preferences though, not performance. Performance wise, FX is just better. Medium format sensors are better than FX.
As MikeGunter said very well, "FX offers more IQ for bunches of reasons and hits the pocket book for all those reasons."

If FoV is a concern, move closer/further away and if you factor how working distance changes between DX/FX into your macro example, it changes things even more.

This actually does not work for macro as you loose the 1:1 (or whatever the goal is) when you move further away to get more background in.

manhattanboy

OED Superior Greater in size or power:

Sorry for the stupidity but am having a senior moment and can't recall what the camera "OED" is.
Even the Oxford English Dictionary was of no help. :-O

LOL !

At least someone got it

TaoTeJared

So what the hell is it?

The humor is lost if no one knows what it is