The D7100 has better resolving power than the D800

Well I thought that would get your attention Following some recent discussions regarding the D800s’ amazing resolution and ability to crop a good image from the far distance I thought I would do a test to demonstrate how a DX body would out resolve it in like for like circumstances. I would have liked to use a D7000 for the test but I don’t have one as it’s much closer to the pixel density of the D800. I believe it would have been just slightly ahead of the D800 unlike the D7100 which is in a different league. The reason that this test is valid is because most of the “amazing” crops that have been posted on Nikonites have used much less than the area of a DX sensor to capture the subject. So before we start the D800 is operating at something below 16mp. This test does not take away the fact that all of our modern DSLRs have fantastic resolving power, it’s just that things are not always what they seem.

I would also like to say that this is not aimed at knocking the D800. I own one and contrary to some views I’m not completely stupid!!

For the test I setup a tripod in my studio and mounted my wife’s D7100 on it with a 50mm 1.8G lens attached. I then composed and took the un-cropped D7100 image shown here. I then did the same thing with the D800 so optically everything is the same (See previous image). At this point I would request that nobody makes reference to the DX crop factor as it doesn’t exist, but that’s another topic. I took the shots at F9, 1/100 sec, raw and ISO 200. I used a studio flash head to light it. I used mirror up and a cable/IR release for each shot.

Having taken the shots I brought them into Lightroom, increased the vibrance to +40 and updated the lens corrections. I turned the sharpening to 0 and then used the Nik output sharpener with the same settings on both images. I then took each image into Photoshop, magnified to 100% and cropped the same part of the image to create to identical looking results, albeit of different resolution.

The results of this test are shown here. By my eye the D7100 is resolving far more detail (The image is about 50% bigger). Whilst the test is not intended to compare things like colour depth, dynamic range or noise handling, in my view the D7100 image in this test looks just as good as the D800 as far as I can see. I know that this would not be the case in certain challenging scenarios.

My conclusion (which is no revelation) with FX cameras vs. DX is that if you want to take shots that fill a similar % of the frame you will need to invest in some big expensive glass with FX. If you do this you can then take advantage of all those pixels. This is not an issue for landscapes and portraits as it’s normally quite easy to fill the frame with standard lenses.

As a closing thought. If we did this test using the D600 for FX, the D200 for DX would probably offer similar resolution and the D300 would stomp all over it!! Again I’m not talking colour, DR or noise, just resolving power.

What you are talking about here is that is referred to among bird photographers as the PPB factor, or "Pixels Per Bird". It's definitely an "all other things being equal" measure, and when you're talking sensor technology they never are, but when it comes to extreme cropping and lens reach it's definitely a factor that's far more important to some types of photographers than others, as you've mentioned. If your subject is always going to occupy only a portion of your viewfinder then I agree that a DX body is your best option.

Hi Geoff, I am not sure I understand what this is about, and let me say that I also am more interested in learning facts than putting up an argument which is the better camera. But there are a couple of issues with your comparison that seem to make it not meaningful to me.

For one, AFAICS the D800's image comprises a much larger area, a much larger portion of the scene, while the 7100 shows a smaller area, which IMHO renders the two images uncomparable if you want to compare the resolving power of either the sensor or the camera.

This to me seems like comparing apples and oranges, because in order to really compare the resolving power of two cameras, it would be necessary to make sure that either the two images show exactly the same content, and then crop from it, or to make sure that the cropped area occupied an identically large area on both sensors, and then crop from it (I am not sure exactly what you are trying to show - that the one camera shows more detail than the other, or that the one sensor shows finer detail than the other).

Apologies if this comes down to crop factor, I'm not yet knowledgeable enough about these terms to use them proficiently, but purely logically, it seems that the comparison as you have uploaded it may compare differently large sensor areas, as you have not taken into account the physical dimension of the sensor area actually occupied by the cropped portion.

Also, the D800 has more dynamic range than the 7100, so its unprocessed images look flatter and seem to have less contrast when cropped like this than the images of a camera with less DR. You can see this by looking at any website comparing for example almost equally good M4/3 or APS-C sensor cameras to FF sensor cameras (dpreview has such reviews, another one is at The 2013 Olympus OM-D E-M1 review, part two: some comparisons (between the two test images in the 'Important testing notes' section) including the D600, where the detail images of the M4/3 look loads better than that of the D600 for this aforementioned reason).

So the fact it may be that the 7100's crop looks more colourful and detailed just because it has less DR, and that it is not an indication of the better resolving power of this camera. AAMOF I had before purchasing the D600 and then the D800 always wondered why their images in those detail comparisons look relatively flat, although they are much more expensive and allegedly better, until I recently found the explanation in this comparison above.

BackdoorHippie said:

What you are talking about here is that is referred to among bird photographers as the PPB factor, or "Pixels Per Bird". It's definitely an "all other things being equal" measure, and when you're talking sensor technology they never are, but when it comes to extreme cropping and lens reach it's definitely a factor that's far more important to some types of photographers than others, as you've mentioned. If your subject is always going to occupy only a portion of your viewfinder then I agree that a DX body is your best option.

Click to expand...

Jake,

Thanks for the reply, as you seem to be one of the few people on this site that understands what I'm trying to say. I see all these crops of images and people saying wow and I just think "so what", my DX body would resolve more detail. Then the D800 owners whose manhood I've obviously called into question insist on throwing dynamic range and noise control into the equation when the point was all about the ability to resolve detail.

One other thing. At no point have I said "best" or "better" others are inferring it from my observation. As you say many factors go towards making an image.

papa2jaja said:

Hi Geoff, I am not sure I understand what this is about, and let me say that I also am more interested in learning facts than putting up an argument which is the better camera. But there are a couple of issues with your comparison that seem to make it not meaningful to me.

For one, AFAICS the D800's image comprises a much larger area, a much larger portion of the scene, while the 7100 shows a smaller area, which IMHO renders the two images uncomparable if you want to compare the resolving power of either the sensor or the camera.

This to me seems like comparing apples and oranges, because in order to really compare the resolving power of two cameras, it would be necessary to make sure that either the two images show exactly the same content, and then crop from it, or to make sure that the cropped area occupied an identically large area on both sensors, and then crop from it (I am not sure exactly what you are trying to show - that the one camera shows more detail than the other, or that the one sensor shows finer detail than the other).

Apologies if this comes down to crop factor, I'm not yet knowledgeable enough about these terms to use them proficiently, but purely logically, it seems that the comparison as you have uploaded it may compare differently large sensor areas, as you have not taken into account the physical dimension of the sensor area actually occupied by the cropped portion.

Also, the D800 has more dynamic range than the 7100, so its unprocessed images look flatter and seem to have less contrast when cropped like this than the images of a camera with less DR. You can see this by looking at any website comparing for example almost equally good M4/3 or APS-C sensor cameras to FF sensor cameras (dpreview has such reviews, another one is at The 2013 Olympus OM-D E-M1 review, part two: some comparisons (between the two test images in the 'Important testing notes' section) including the D600, where the detail images of the M4/3 look loads better than that of the D600 for this aforementioned reason).

So the fact it may be that the 7100's crop looks more colourful and detailed just because it has less DR, and that it is not an indication of the better resolving power of this camera. AAMOF I had before purchasing the D600 and then the D800 always wondered why their images in those detail comparisons look relatively flat, although they are much more expensive and allegedly better, until I recently found the explanation in this comparison above.

Click to expand...

This is nothing to do with which is the better camera. It's a bit like asking is a truck better than a car. It depends what you're trying to achieve. As a D800 owner I'm hardly likely to go to these lengths to rubbish the product.

When cropping like this (Using a very small part of the overall image), as all of the recent posts have been this test works. Those two cropped images represent 0.2mp for the D800 and 0.3mp for the 7100. In my book more pixels equals more ability to resolve detail hence the title. I specifically stated that this was not a DR, colour depth or noise test but you insisted on coming back to it. In reality a modern APSC is pretty good in all those respects and the D800 only blows them away when down sampling. The simple fact is that the hair detail is better on the 7100 image. Both images are raw and processed the same way. No matter what I did in post the resolution of the D800 would still be less when performing extreme crops. If I had filled the original frames with the final image composition the D800 would have displayed much more detail, but that is nothing to do with extreme cropping.

You say that both images should contain the same content in order to compare. Both were offered the same content by use of the same lens and distance to subject. I then cropped to provide the same content. If you are unsure why the D800 shows much more of the scene in the uncropped image I suspect some fundamental knowledge is missing as that is real FX/DX basics when using the same focal length.

The irony of all this is that I was just trying to help those who did not understand the pros and cons of different pixel density cameras especially between FX and DX. All I've had in most responses is people defending the honour of their D800 super cameras. I use my D800 when it's the best tool for the job and my D300s when it is. I think I'll leave the whole thing now. For those whoare able to understand it that's great and those who don't just carry on enjoying all that resolution.

If anyone would like to spend the time showing how the D800 has more resolution using such a small area of the sensor I would welcome it. That's resolution folks, not DR, Noise or colour depth.

I find the results a little surprising, while I know how good the D7100 is because I crop a lot of my photos , so I see it all the time. I had just assumed the D800 would have the upper edge here.

What I dont really understand here, is I thought you were going to use the D800 in DX mode. The photo shows a lot more area in the frame than the D7100 photo. Am I missing something here ?

​I guess it wont matter in the end because your crop will be using the same portion of the sensor as it would have anyway.

As referenced in another thread Nikon D7100 vs. Nikon D800E - Sensor Comparison :
. pixel size of D7100 is 3.9 micron
. pixel size of D800 is 4.87 micron

Thus you will get more pixels per image size in the D7100

The other parameters as measured by DXO DxOMark - Compare cameras side by side are
Colour Depth :- D7100 = 24.2, D800 = 25.3
Dynamic Range :- D7100 = 13.7, D800 = 14.4

So for Shooting objects which can at the most fill a DX sensor, D7100 wins as it will have more pixels, while for portraits and Landscapes D800 wins due to better DR and colour.

aroy said:

As referenced in another thread Nikon D7100 vs. Nikon D800E - Sensor Comparison :
. pixel size of D7100 is 3.9 micron
. pixel size of D800 is 4.87 micron

Thus you will get more pixels per image size in the D7100

The other parameters as measured by DXO DxOMark - Compare cameras side by side are
Colour Depth :- D7100 = 24.2, D800 = 25.3
Dynamic Range :- D7100 = 13.7, D800 = 14.4

So for Shooting objects which can at the most fill a DX sensor, D7100 wins as it will have more pixels, while for portraits and Landscapes D800 wins due to better DR and colour.

Click to expand...

Aroy

Thanks for your replies to a number of posts. I no longer feel alone in this line of thinking. I also fully accept your DR, colour etc points and I never questioned them in my observations.

Ironwood said:

I find the results a little surprising, while I know how good the D7100 is because I crop a lot of my photos , so I see it all the time. I had just assumed the D800 would have the upper edge here.

What I dont really understand here, is I thought you were going to use the D800 in DX mode. The photo shows a lot more area in the frame than the D7100 photo. Am I missing something here ?

​I guess it wont matter in the end because your crop will be using the same portion of the sensor as it would have anyway.

Click to expand...

Brad,

When the D800 is in DX mode it only use the 23*15mm area of the sensor. When you look through the viewfinder you see a smaller image if you set it up too blank off the unused part. If I had done the tet with the D800 in DX mode the uncropped images would have looked the same unless you pixel peeped.

Another interesting point in the DR and colour depth of the D7100. It is actually better than the legendary D700 according to DXO, so it's hardly in the point and shoot category.

Ironwood said:

What I dont really understand here, is I thought you were going to use the D800 in DX mode. The photo shows a lot more area in the frame than the D7100 photo. Am I missing something here ?

​I guess it wont matter in the end because your crop will be using the same portion of the sensor as it would have anyway.

Click to expand...

Exactly. Putting an FX camera in DX mode is essentially a 1.5X "pre-crop" of the image. You no longer get the full resolution of the sensor (36MP's) you only get the part of the sensor that equates to the DX crop (16MP's in this case). If a picture's worth 1000 words, here's Geoff's original photo with the part of the D800's sensor that would be ignored in DX mode inverted (approx measure).



Geoff's point is purely about the number of available pixels in a crop and has nothing to do with the quality of the light available in those pixels. I'm sure given enough time people can come up with examples where each of the cameras out performs the other in a crop test, because all things aren't always equal. That's why knowing how you're going to use a camera is of paramount importance when you go out an buy one. I knew the crop numbers going in and decided that I wanted the D800's sensor even if it sacrificed pixels in a heavy crop as more times than not I was not looking to pull details out of a small area. If I was, I may have chosen differently.

Thanks for enhancing the example Jake. The point about understanding how a camera behaves is the key point when trying to achieve a particular result.

Hi Geoff, no offence meant. I'm really only trying to understand.

The other day I was trying to get a handle on the differences between the FX and DX cameras d800, d600, d7100 and d7000, and I developed this little chart in Excel. This is the first time I've attempted this type of analysis so if I missed something please let me know.

If my analysis is correct the d7100 pixel density is impressive. If the d7100 pixel density were extraploated to FX format I believe it would equal to 53.23 mp.

.

papa2jaja said:

Hi Geoff, no offence meant. I'm really only trying to understand.

Click to expand...

No problem no offence taken.

teagueAMX said:

The other day I was trying to get a handle on the differences between the FX and DX cameras d800, d600, d7100 and d7000, and I developed this little chart in Excel. This is the first time I've attempted this type of analysis so if I missed something please let me know.

If my analysis is correct the d7100 pixel density is impressive. If the d7100 pixel density were extraploated to FX format I believe it would equal to 53.23 mp.

.

Click to expand...

That's exactly right. Given the greater pixel density, the size of the individual pixels must be smaller, which means that they cannot capture the same amount of light as the larger pixels on the less dense sensors, all other things being equal. So while you have more pixels, it's likely that the information contained in them is not as rich as those in the pixels on the less dense sensor. So while you may be able to extract more detail on a crop, the information regarding the light that makes up those details may be such that you can't "do as much" with the information in post processing. On a well exposed, balanced image that likely won't matter. But on an image where you not only need to crop excessively but also do a lot of work with the exposure to see what you want to see, you may be able to extract a "better" result from the less dense sensor than you could from the one that gives you more MP's. There's always a trade-off. Thankfully not only do you have choices, but there are no real "losers" in these arguments.

Jake,

Thanks for that explanation. Basically, if I understand you correctly, more pixels may or may not make a better image.

So I think you may be getting to somethings I read which is some people aren't necessarily excited about continuously racking up a higher pixel count, correct?

teagueAMX said:

So I think you may be getting to somethings I read which is some people aren't necessarily excited about continuously racking up a higher pixel count, correct?

Click to expand...

I'm not on any particular side of the pixel count wars. I'm just saying that pixel count is only one measure of a cameras potential image quality. More does not always equal better when it comes to IQ. If you can't retain or boost the level of light information captured by each pixel as you increase pixel density then you may, in fact, be hurting IQ more than helping it. It's a careful balance.

Geoff started this thread to talk about only one aspect of this, which he called "resolving power", which I interpret as resolution. Meaning, all other things being equal (i.e. light and optics are optimized/not troublesome in any way) the more pixels you have the more detail you can extract from a small portion of the frame. I absolutely agree with this. However, things aren't always equal, and lighting/metering is not always perfect, and sometimes you need to fight a bit with the light in your photo and this is where a smaller pixel can be problematic, even though you have more of them.

Geoffc said:

At this point I would request that nobody makes reference to the DX crop factor as it doesn’t exist, but that’s another topic.

Click to expand...

You are kidding yourself. The crop factor does exist, and is an overwhelming major factor, regarding exactly your point. You can shut your eyes, but it does not go away.
If you cannot understand why it is called "crop factor", see FX - DX Lens Crop Factor

The results of this test are shown here. By my eye the D7100 is resolving far more detail (The image is about 50% bigger).

Click to expand...

Which is precisely the 1.5x crop factor. It is hardly rocket science, the details are this:

D7100 - 6000x4000 pixels (24 mp) on a 23.5x15.6mm sensor (prints have to be enlarged 1.5x more).

D800 - 7360x4912 pixels (36 mp) 50% more pixels, which is resolution. 35.9x24 mm sensor (1.5x larger). (this 50% more pixels is NOT due to the same 50% crop factor, it is just a coincidence of design parameters).

D800 will shoot DX too, 4800x3200 pixels (15 mp). If compared this limited way (to see the same view as the cropped FX), yes, then, D7100 is about 60% more pixels. But compared this way, the DX has to stand back 1.5x farther (which you did not do, you did not compare the same image view). The crop is a plus if seeking telephoto, but it is the pits if you want wide angle. It is really hard to accomplish wide angle from the DX crop factor. The lens may see wide angle, but the DX sensor crops it away.

It seems quite reasonable if to compare images, that it is only meaningful to demand comparing the same actual view, which means the DX stands back 1.5x farther, and the D800 FX still has 50% more pixels (in the same view). Also the uncropped FX frame needs less enlargement to print same size.

D7100 certainly has many fine features, but 24 mp being more resolution than 36 megapixels is not one of them.

teagueAMX said:

Jake,
Thanks for that explanation. Basically, if I understand you correctly, more pixels may or may not make a better image.

Click to expand...

As Jake explained, larger pixels typically have less noise problems, and smaller pixels typically have more noise (more noise means higher ISO is less acceptable). Not exactly a factor of resolution, but it certainly can affect the suitability of the image.

Your Excel chart seems right, no question. However, it discusses resolution on the sensor, which size we never examine. Instead we must enlarge it, to video screen size, or to print size, DX enlarged 1.5x more than FX, but vastly larger than thumbnail size. Which then, to provide this larger view, the COUNT of pixels becomes the important factor (regarding their spacing in that larger view).

Said another way:

D7100 - 6000x4000 pixels (24 mp)
D800 - 7360x4912 pixels (36 mp)

For both printed at 300 dpi, (and assuming we have framed the same image):
the D7100 image prints 20x13.3 inches