Comparing camera IQ

Joined
Jan 15, 2010
Messages
4,001
Location
UK
Recent discussions here about camera IQ comparisons and the fact that I have been asked for my opinions by an old colleague has caused me to look at some independent samples side by side and also show and explain my thoughts on IQ.

This is not offered as a highly exact scientific analysis as it relies on others testing and of course my own opinions about what I am observing. Accepting of course that others MMV

While I have omitted quite a lot of specific questions I hope that this may be of interest to other forum members.

From the outset it should be made clear that this is really a comparison of IQ arrived at by up/downsampling and as presented here comparing an older generation Nikon D700 to a D800E and AFAIK applicable to the latest D810 that while improved may not be a huge IQ leap (I do not have access to similar images therefore this is my best guess)

This is not a feature comparison such as better AF, faster FPS, improved handling that may combine to make one camera superior to another for the individual.

There seems to be a lot of people trying to compare which cameras are better than others at certain ISO and often the replies are, I have owned them both and can tell you that camera A is better than camera B at this or that ISO. Or any other comments the author tries to make usually related to the superiority of his current love - not suggesting that this may not be justified.

Quite often the comments are made after one camera upgraded to another and the owner may feel he should try to justify to the world why he has spent the entire budget of a third world country on his new acquisition and just look at the huge benefits which will make me a better photographer.

Rarely do you see anything other than subjective opinion. Although I am sure they must exist I cannot recall seeing any objective evidence put forward to support an authors view of why the Nikanon Super XXL MkII is so much better than the older Nikanon not SoSuper XXL MK I. That is to say those making the claims not putting forward any photographic evidence to prove the point - say two images taken at the same time with the same lens from the same position of the same subject etc.

Therefore many claims seem to be very subjective with no basis in reality as objective comparisons have never been made much less shown.

The truth is that even with the presentation of 'objective' evidence it is often hard for us to accept this evidence if contrary to our belief.

While many I know pooh pooh the ideas of any testing let alone trying to go for objective testing it is nevertheless a necessity for many of us and particularly for those that have worked in a highly competitive commercial field.

Apart from undertaking carefully thought out side by side testing of one camera vs another the only other sources of objective testing is to be found on the net - although you do have to be a little selective on which sites. Hopefully the testing should be rigorously controlled and consistent and repeatable.

In any case it would be hoped that the test data would represent comparing apples with apples. IMHO, Bill Claff, Imaging resource, DPR and DX0 are among those that provide us with consistent data and while they may not agree exactly with each other the tests do seem to be comparing like with like.

The data I have used is from DPReview Studio Shot comparisons Here downloading the original raw files and processing via ACR.
No editing other than ACR defaults (sharpening) applied to these images. No noise reduction applied and images saved to web in CS6

I have looked at several ISO comparisons from the Nikon D700 and the D800E and just settled on showing the results from the 12800 ISO shots side by side just using screen grab as this good enough for purpose at this time.

Often we talk about downsampling as a way to minimise noise and although there is some truth in this it conveniently overlooks that we may have to upsample to get the size we require if native pixel count is different between cameras.
For instance a Nikon D800 series has a native resolution of 7360 x 4912 pixels. Should we wish to print to a Canon or Fuji printer then we want to be using a resolution of 300 PPI. This will result in a 24.5" x 16" print.
But if we have used the Nikon D700 the native resolution of 4256 x 2832 pixels only allows us a 14" x 9" print @ 300 PPI.

This is how the two images appear at 100% zoom side by side in PS. Note that 100% zoom is an unrealistic view of a print being a 3x magnification using the 300 PPI example above
D700_D800E_100.jpg
Subscribe to see EXIF info for this image (if available)

Should we wish to upsample the D700 file to match the D800 this is what we can expect to see - it should be obvious which is which
Upscaled-D700_12800.jpg
Subscribe to see EXIF info for this image (if available)

Now downsampling the D800E to match the D700. Still at 100% zoom so not as realistic as making a proper print. What you should see on screen and be improved in the print is a reduction in noise and an apparent increase in resolution. Worth noting while downsampling can reduce noise it also must reduce resolution although the reduction can give the appearance of a sharper image
Downsampled_D800E.jpg
Subscribe to see EXIF info for this image (if available)

If we downsample the full size 12800 ISO D800 for web it is interesting to see that even without any noise reduction applied there is little evidence of a noisy image at least viewing on a 100 ppi monitor screen. This is a crop 100% from a web sized image 1024 pixel wide
D800EDownsample-to-1024-for-web.jpg
Subscribe to see EXIF info for this image (if available)

I believe the consensus is that IF noise reduction needed that this is applied prior to image downsampling to maintain the best IQ.

Hope that this may be of some interest
 
Joined
Nov 11, 2005
Messages
3,187
Location
West coast of FL (formerly Westchester County, NY)
"Quite often the comments are made after one camera upgraded to another and the owner may feel he should try to justify to the world why he has spent the entire budget of a third world country on his new acquisition and just look at the huge benefits which will make me a better photographer.

Rarely do you see anything other than subjective opinion. Although I am sure they must exist I cannot recall seeing any objective evidence put forward to support an authors view of why the Nikanon Super XXL MkII is so much better than the older Nikanon not SoSuper XXL MK I. That is to say those making the claims not putting forward any photographic evidence to prove the point - say two images taken at the same time with the same lens from the same position of the same subject etc.

Therefore many claims seem to be very subjective with no basis in reality as objective comparisons have never been made much less shown.

The truth is that even with the presentation of 'objective' evidence it is often hard for us to accept this evidence if contrary to our belief."


Absolutely right on point. Fact is, most times (especially on other forums which shall remain nameless) people ask "which camera is better" get bombarded with replies from people who have never done an ounce of testing. I might add in a totally obnoxious manner that I have read answers that lead me to believe that the responder has no idea what he/she is talking about and might be classified as "stupid" (I use that adjective because the mods here would censor the one I had in mind). The internet is rife with people who insist that their opinions are gospel and no amount of testing or logic to the contrary will alter their opinion. (Try asking whether a protection filter is a good idea and sit back and watch the fierce arguments from both sides trying to bludgeon the other side into agreeing.)

I do understand the point of a relatively new photographer asking for opinions on which camera may suit him best, but the replies seem to often come from someone who is enamored by his view of the world and refuses to consider that others may disagree.

Full disclosure - If I were to be totally honest, I am sure that somewhere along the way, I have been guilty of the offense that I am criticizing here.
 
Joined
Mar 13, 2013
Messages
2,014
Location
vancouver, canada
Tony and Rick you two are way too sensible...

Ok, should I use a protective filter on my expensive lens with a Teflon coating? :beaver::beaver::)

I remember years back being at a Pentax event, Way back when Pentax was Pentax. They were introducing the super multi coated filters. The presenter lit a cigarette and butted it out on the filter protected lens while commenting that this would not scratch the filter. Someone in the audience dryly commented: " I use an ashtray for my cigarettes. "

Cheers,
alexis and Georgie Beagle.

The up sampling and down sampling examples are very informative and really tell the story. I approve,. - Georgie Beagle
 
Joined
May 5, 2005
Messages
25,771
Location
SW Virginia
Tony, I find nothing to disagree with in what you have written, but I'm not sure exactly what your point is. Is it that trying to compare cameras with two different resolutions is fraught with ambiguities? Certainly that is not a controversial statement.

While many I know pooh pooh the ideas of any testing let alone trying to go for objective testing it is nevertheless a necessity for many of us and particularly for those that have worked in a highly competitive commercial field.

I certainly agree with you there. As a scientist I cannot understand the attitude of those who reject any sort of objective tests or comparisons and simply say "go out and shoot". That attitude can certainly only lead to subjective opinions which are often very incorrect.

As an amateur photographer without any test equipment it is difficult for me to do objective comparisons, but sometimes they are possible. While going through pictures from a recent trip to Italy with a D7200, I was disappointed by the rather high noise levels in some interior low-light photos taken at fairly high iso values. This led me to question whether I should have taken my D610 instead, so I decided to try to make an objective comparison. Luckily these are both 24 MP cameras which makes such a comparison much easier. Some of my results are available in post #31 in this thread. To me, these comparisons not only showed the superiority of the D610 in these situations, it helped me to see how much difference there really is. As a result I will be taking a FF camera (now D750) on my next trip.
 
Last edited by a moderator:
Joined
Jan 15, 2010
Messages
4,001
Location
UK
Hi Jim, yes I guess part of my point is that comparing cameras can be really difficult in as much as often we are presented with what only amounts to subjective opinion, some based on actually working with the cameras in question, but nevertheless still subjective as rarely are side by side tests undertaken. I like your examples and would say that your testing is the exception that proves the rule :)

Originally I was asked for my opinion on which camera from an old colleague looking to get into a camera system for some 'serious' work. Initially he was thinking of a used D700 as he had seen high recommendations on the net and via a local dealer, but was concerned that 12MP may not be enough if he wanted to produce very large prints. As budget did not really enter into the scene my off the cuff remark was that he should perhaps look at the D810 and other manufacturers equivalents.

I have seen similar requests here and on other fora rarely are these request met with examples. So while I had already done some digging I thought it might be of use to others and if nothing else give some idea of information sources.

I think the whole testing scenario is fraught with ambiguities not least of which is comparing different sensor resolutions. Then there is the question of what the user expects out of the camera and the main usage not forgetting the final output destination. For instance it is very unlikely that any real benefit to be had with a 36MP camera vs a 12MP if all you are going to be doing is viewing on the web or making the odd 10" x 8" print.

In doing the comparisons all I concentrated on was doing upsampling and downsampling to look at IQ as other areas such as handling benefits would need to be understood. I also pointed out that viewing images at 100% crop was not the best way to judge end quality particularly if going to print due to the zoom being around a 3x magnification. Correcting for noise at this size needs a gentle hand if we are not to erode IQ. Further if these images viewed on a high resolution monitor the noise and resolution characteristics could appear quite different. The bottom line is that to make a real judgement one should not rely on the screen but make a print i.e. if the destination is to be print.

One of the comparison I did and not shown originally involved the D810 and another make same MP count. An interesting comparison perhaps at relatively high ISO and obviously one better IQ but again to be meaningful resize for web or print.
Nik810_K1.jpg
Subscribe to see EXIF info for this image (if available)
 

Growltiger

Administrator
Administrator
Joined
Apr 26, 2008
Messages
13,620
Location
Up in the hills, Gloucestershire, UK
I agree that the whole idea of 100% crop for comparison between cameras is pretty much useless. In fact it is likely to lead to incorrect conclusions.
100% crop is very helpful for a quick check of focus and quality of one's photos, but that is all.
 
Joined
Sep 19, 2011
Messages
1,398
Location
Ireland
Real Name
Eamonn
viewing images at 100% crop was not the best way to judge end quality particularly if going to print due to the zoom being around a 3x magnification.

Educate me here please on that statement...would it depend on the resolution of the monitor too?

thanks.
 
Joined
Jan 15, 2010
Messages
4,001
Location
UK
Educate me here please on that statement...would it depend on the resolution of the monitor too?

thanks.
Yes it does depend on the monitor and the printer requirements. For ease and based on what I use the assumption is that the monitor resolution is 100 PPI (mine is 98ppi as an example) and we will output to a Canon or HP printer that declares itself as 300/600 PPI. Epson declares 360/720 PPI to the OS

So sticking with Canon printer we will be ideally sending 300 PPI (or if native resolution greater 600 PPI) for print so if we view at 100% zoom in Ps or LR we are looking at an image that is actually 3x magnified view of print size. If we try to view at actual print size then the zoom ratio would need to be 33.33% leaving us viewing an image that will contain only 100 pixels per inch standard monitor @100 PPI

We cannot judge print quality at actual size 33.33%, due to only having 1/3 of pixels available on our monitor. And viewing at 100% zoom showing the 3x magnification makes it difficult to judge sharpening and noise reduction - until practical experience gained.

Still IMO 100% zoom should be used to evaluate both sharpening and noise reduction. Capture and creative Sharpening first should probably look a little crunchy at 100% that is a little more than you would expect needed. Conversely NR will need a lighter touch as noise will reduce greatly when output at final size either print or screen. In some cases you may not need to reduce noise at all as it will not be seen in print but this will depend on output size.

For the newer higher resolution monitors then the zoom ratio is going to be different moving the goal posts somewhat. From what I have seen reported our images are likely to appear sharper and noise reduced due to the higher pixel count. I seem to recall that Andreas who posts here reported this.

However we are still left with scaling issues with 4K monitors trying to judge as we are not getting WYSIWYG until such time as our monitor resolution is 300 - 360 PPI i.e. 1 monitor pixel = 1 print pixel. I will probably still complain then wanting 600 - 720 PPI monitor to equal the fine print res. of printers :D
 
Joined
Nov 10, 2008
Messages
3,972
Location
Chicago
All depends on how large a print and how high the iso needs to be. Then there is in camera noise reduction that can not be turned off.

I suggest Image Resource Comparometer as a very good start. Large images from any two modern cameras can be put up on the screen at the same time and the images are same subject, lighting, iso, of your choice.

D3 and D700 are same camera inside for imaging. Sharpness can be maintained with proper noise reduction, i.e. filter, stylize, find edges. Make an edge mask
 
Joined
Feb 17, 2008
Messages
2,641
Location
Virginia Beach, Virginia USA
Real Name
Bill Mellen
I agree that the whole idea of 100% crop for comparison between cameras is pretty much useless. In fact it is likely to lead to incorrect conclusions.
100% crop is very helpful for a quick check of focus and quality of one's photos, but that is all.

Exactly! What matters the most is how the image looks in its intended display size / print.
 
Joined
Jan 15, 2010
Messages
4,001
Location
UK
I cannot recall any actual references other than remembered quotes from 'imaging experts' stating that downsampling improves noise appearance and can make images appear sharper on screen and in print plus my own practical testing.

I think it is important to appreciate that we are talking about making an image appear sharper through the process of downsampling from the full sized original. In the process the downsampling algorithms must 'discard' pixels and interpolate what should be seen, also in the process it is likely that sharpening applied.

A quick example that hopefully demonstrates these effects. Image supplied by Dpreview image comparator tool : Nikon D700 @ ISO 12800.
Top image shown at 100% screen view. This equates to a magnification factor of 300% if you are viewing this on a standard monitor (around 100 pixel per inch).
Middle image crop downsampled to actual print size when sent at 300 ppi to Canon/HP printers. Note: You can only get an idea of size comparison here as you do not have enough screen resolution to judge detail - but I would expect noise to be improved as well as apparent sharpness, although it does depend on other printing factors
Bottom image - scaled for web at 1000 pixels wide

Original Nef brought into ACR no NR applied just default sharpening. Duplicated this for print and web views DPR LINK

DownsamplleD700_12800.jpg
Subscribe to see EXIF info for this image (if available)
 
Last edited:

Latest threads

Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Nikon Cafe is a fan site and not associated with Nikon Corporation.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Forum GIFs powered by GIPHY: https://giphy.com/
Copyright © Amin Forums, LLC
Top Bottom