Scott Murray
Senior Member
Last edited:
The D7100 has better resolving power than the D800
- Thread starter Geoffc
- Start date
Wayne
I normally respect your posts, but on crop factor we must disagree. I do agree that DX crops in the sense that a big chunk of sensor area is missing so you have a narrower field of view. What it doesn't do is magnify the lens and bring things closer. The image projected onto a DX sensor by a given lens is the same physical size on an FX sensor. It just happens that the FX sensor also displays potentially superfluous scenery surrounding the subject. Effectively the DX sensor saves you some cropping in Lightroom.
Now if you get a D300 and a D700, both of which are around 12 mp you get more resolution on the D300 because it's not throwing as much away.
On that basis it's not the sensor size providing greater resolution and detail, it's the relative amount of pixels per area.
I think you need to see my very first post on the subject regarding the bear image to understand the points bring made and how they are relevant. Everybody has tried to introduce all sorts of fanciful D800 stuff to win the battle, which is interesting as I never criticised it in the first place. Sadly male pride gets dented at the mere suggestion that their expensive 36mp camera isn't the perfect tool for every job.
OK this thread has probably gone on long enough and I'm certainly getting bored with it so I can't imagine what everybody else thinks.
On that basis:
1. The D800 out resolves everything no matter what the circumstances
2. The D7100 IQ is so far short of the D800 it shouldn't even be discussed in the same post
3. Did I mention the D800 has better noise control
4. and dynamic range
5. not to mention the colour depth
Now that I've cleared that up I don't feel the need to defend my original post about the amount of detail/resolution/pixels that can be cropped from a far off image.
If anybody's interested my wife will be selling her D7100 later tonight for a couple of pounds and we will share my D800 in future
On that basis:
1. The D800 out resolves everything no matter what the circumstances
2. The D7100 IQ is so far short of the D800 it shouldn't even be discussed in the same post
3. Did I mention the D800 has better noise control
4. and dynamic range
5. not to mention the colour depth
Now that I've cleared that up I don't feel the need to defend my original post about the amount of detail/resolution/pixels that can be cropped from a far off image.
If anybody's interested my wife will be selling her D7100 later tonight for a couple of pounds and we will share my D800 in future
I fully agree, to extent that there is no change whatsoever in the lens. DX simply crops away some of it, to be a more narrow view. But to view that image, we can only show it much larger, enlarged on a video screen or a paper print. Assuming we framed the two images the same (to be able to compare them meaningfully), the smaller cropped DX image must be enlarged 1.5x larger than the FX image (just to be the same viewed size). This enlargement of a narrow view is perceived as a telephoto effect, and we compare the DX view as 1.5x greater EFFECTIVE FX focal length (meaning DX must stand back 50% farther to see the same view). That's pretty much the entire idea of it.
Pixels per sensor area is only important to noise. Pixels per viewed area is another thing, and it is very important, and it happens that DX always has to be enlarged 50% more than FX to see it anywhere. So.... any (dimensional) number you computed on the DX sensor must be divided by 50% before it is acceptable to compare to FX. That is simply how life is.
Last edited:
Pixels per sensor area is only important to noise. Pixels per viewed area is another thing, and it is very important, and it happens that DX always has to be enlarged 50% more than FX to see it anywhere. So.... any (dimensional) number you computed on the DX sensor must be divided by 50% before it is acceptable to compare to FX. That is simply how life is.
Image size when viewed is based on pixels wide by pixels high. So a D300 DX and D700 FX would display the same size as they have the same pixel dimensions (give or take a few). A 24mp D7100 would be bigger than a D700 even though it's DX.
Here is a test I did some time ago.
http://www.flickr.com/photos/teerecks/sets/72157629889340922/
Read the description to get the idea properly.
Your point almost sounds like the frequently quoted images should be uploaded at 1000 pixels and 72dpi. It's 1000 pixels wide end of story, the DPI does not affect it.
Now look what you've done. I wasn't going to respond to any more comments as its run its course and now another can of worms has been opened
Last edited:
OK, mp is about resolution, but a D800 image is 36 mp, and which automatically wins, regardless of what this threads OP said about it.
But if you are going to compute pixels per inch of sensor dimension, it is not meaningful until you also compute in the greater enlargement required of DX before they can even be viewed and the numbers compared to FX.
We can agree on that. Any mention of 72 dpi for video is mistaken and misleading, confuses the newbies. Hearing it is like stepping back in time 10 or 15 years. But my saying this probably means the powers that be will delete this message again.
Last edited:
And that is the argument for better IQ of Medium Format sensors. With larger (fatter) pixels the colour depth will increase, so will the Dynamic Range, hence you can extract more information via post processing.
By better image I assume you mean better technically and not aesthetically.:tennis:
More pixels will give you more details. If the picture is exposed perfectly and the DR of the scene reasonable, then even a tiny cell phone sensor will have excellent quality. After all while seeing on monitor or printing we rarely exceed 8 bits of resolution, while the DSLR cameras have any where between 12 and 16 bits of data. What it means is that though there can be upto 2^48 colours (16 bit image, 3 colours), our reproduction limits it to 2^24 only.
What you can do with higher number of bits is to recover data from shadows or tone down the high lights, that is remap 12-16 bits to 8 bits per colour for visualization on screen or printer. Of course there are some devices which support 10 bits, but you are still below the 14 bit (Nikon) or 16 bit (MF) data in the image.
Another thing to consider is that the denser the sensor (higher the MP), the finer the details. This means that the lens has to resolve better. Some older lenses are still upto it, but a lot of older designs start showing their short coming with higher density sensors.
BackdoorArts
Senior Member
This is making my head spin because you guys (Geoff & Wayne) are talking in circles around one another. I'm not wading back in here as I've been through this a couple too many times already, but I'm an FX guy and I 100% agree with Geoff's original post, which if read carefully speaks from an "all other things being equal" standpoint - meaning lens, subject and original shooting location. If those things are not changed and only the body is swapped, a D7100 will offer you more pixels of your subject than the D800 (i.e. greater resolution). Period. Change any of those things or start talking about pixel size and light factors and it's a new post and a different point, many of which are valid, but none of them refute Geoff's original premise.
Horoscope Fish
Senior Member
I've had some similar conversations here and at some point you just have to let it go...
But "all other things being equal" is precisely what Geoff did NOT do, and which caused my objection.
His two exposures were some different too, which tended to hide detail in the smaller FX image (metering of a FX wider area of dark background could do that, but he says these were manual, so maybe the flash varied a bit? Or maybe auto something in post processing, which was said to be considerable.)
Standing in the same place is not enough, all other is NOT equal. The scene being compared was not equal (caused by standing in the same place). If standing in same place, with same camera body, and taking a picture with a 100mm lens, and then another with 150mm (effective focal length due to crop factor), and then being surprised that we imagine we can see more in the larger image, is (sorry) nonsense logic. We can only make valid comparisons of the same image scene, which others have also pointed out here. Zoomed in or not is at least apples and oranges.
Sure, yes, we certainly can appreciate DX for its telephoto effect, and DX is prized for sports and wildlife pictures, but which is about crop factor, NOT about resolution. But then DX makes wide angle suffer in the opposite effect however. All things are not equal.
Tiny compact cameras are have passed 12 mp now, and as compared to say 12 mp FX, we could make arguments about their resolution. Resolution certainly is about mp (the fineness of pixels in the final enlarged view), but enlarging them 6x more does not help them. Their tiny sensor dimensions are maybe 1/6 of FX (crop factor 6x). In bright sun, they do reasonably well, but when ISO gets slightly high, say above ISO 200, they start becoming a poor picture. All things are not equal.
Last edited:
Nathan Lanni
Senior Member
Yeah! I love it - right in the thick of it spike the punch!!!!! ROTFLMAO
Wayne your now entering the realms of BS. To make things equal I metered the subject with a Sekonic 478DR meter and used studio strobes for light. I'm sure I mentioned this in a previous post. I then set both cameras in manual mode for shutter, aperture and ISO. Same settings for both. Don't assume I would use your methods of testing rather than the correct ones.
This has been done to death now so live with whatever you believe, I don't much care as i have enough info to make my own informed decisions. I never set out to bias a result it just came out that way.
I have no interest in continuing this debate. A couple of contributors who I respect fully understood the point being made while others have gone on a fanciful journey that bore little relevance to the original theme.
Nobody who has contested my original statement has gone beyond spouting un-corroborated here say, rather than doing a controlled test of their own that would make all those detailed hairs in my 7100 image blur like the D800 image.
I am just saying that trying to compare your two images is very frustrating to me (as impossible), because the larger D800 image is obviously shown smaller (harder to see any detail), and also shown brighter, missing some of the shadow that would have helped to define the small fur strands. Compare the noticeable greater brightness of the yellow background for example. Simply not the same exposure. I don't know why not, but there it is, and I cannot see it as a valid comparison of anything, esp not resolution.
I'm not sure what you did to confuse it so much. We know the cropped DX frame is obviously smaller than an uncropped FX frame. We also know the same lens standing at same place will show a wider scene view in the larger DX frame (1.5x more angular width shown in 1.5x larger frame). But using the same lens from same location, the bears head obviously ought to be projected on the sensor at the exact same size though - the lens is what it is. But we do not look at the image on the sensor. We must enlarge both images to the same screen or paper print size to compare them, the DX must be enlarged 1.5x more to be same size, so it then appears enlarged, or telephoto.
You are arguing in favor of this telephoto effect. Many people do like it. The DX image does appear with this telephoto effect. But this is hardly a resolution test.
Last edited:
There is the mistake you made Geoffc. You did not use the same focal length in your tests. You used the same lens, the same subject, the same ISO, the same lighting but you did not achieve the same focal length. The D7100 sensor does not see the same size of image as the D800 so in reality you magnified the D7100 image just by the crop factor alone of using an FX lens on a DX body. Besides the power to resolve does not lie in pixel size or pixel pitch but in EVERYTHING the sensor is SEEING and RESOLVING from the data it gathered. I'd accept you argument if you said that to you the D7100 images cropped looked better then the D800's. That's just opinion and your so very welcome to as many as you want. Like the opinion you stated in this thread that "peoples manhoods appeared to be questioned". Maybe, maybe not. I can't speak for everyones manhood or even if you were being sarcastic or if it was stereotypical dry British sense of humour but I for one was not offended nor had my feathers ruffled. I just truly want to understand what makes you insist that the D7100 resolves better then the D800 as if was an established fact when all he articles I've found suggest the opposite. Just my respectful 2 bits.
Ha Ha love it!!
Oh Ya?!?! Wel......well......mydadcan'tbeatupyourDadeither!! So there!!
The point being made is that the D7100 has higher density of pixels. So as the image (in this case the bear) will be of same size on any sensor (if you are using the same lens at the same FL), the sensor with higher density will have more pixels of the image.
That said, the magnification for either display (100 dpi?) or print (300 dpi?) required will be less for a higher density image than for a lower density image.
Alternatively think of it as
. If original image of bear on sensor is 5mm wide
. at 5 microns the image will be 1000 pixels wide
. at 4 microns will be 1250 pixels wide
. to print it at 200mm you need 40x magnification
. at 5 micron the 100mm print will have a density of 25/mm
. at 4 microns the 100mm print will have a density of 31.25/mm
So an image taken with higher density sensor can be larger in print for same DPI or denser for same size