[normhead]

It is wrong to suggest that images not presented as there were the original comparisons of non pixel shift with both cameras this indicated quite clearly what you would expect in the field comparing each system (camera and lens combination). As stated earlier comparing two different MP count bodies with two different lenses and not treating the data appropriately (and differently!) for each is problematic.

But it's not equally wrong to suggest that the results of artifacting occurring on 2D line drawings in some way will represent every day photography in 3D space?

You do know there's a video we are discussing there?

---

Last edited by normhead; 12-04-2017 at 07:19 PM.

[TonyW]

....You do know there's a video we are discussing there? This whole paragraph suggest you didn't watch it.

Yes a video that wastes several minutes of my time complaining of conspiracy theories and the ridiculous notion that post processing is cheating - what the hell does he think happens to a raw image when it goes through the editing pipeline?

Are this guys original raw available for download anywhere?

The main thrust that high ISO noise is somehow better does not appear to be supported from the raw files I have looked at at the various ISO.

Your response suggests that you have not bothered to download and evaluate comparable images to ascertain for yourself the noise characteristics.

[shardulm]

Obviously so when Pixel Shift applied, however that is not the case as far as I can see with standard images.

However If you are referring to high ISO being significantly better would you care to share your image sources and the objective testing comparing cameras?

Well I thought we were discussing the results in the video. But if you think d850 has better iso performance than K-1according to the video then great. I have no qualms. Enjoy your best thing since sliced bread.

[UncleVanya]

Yes a video that wastes several minutes of my time complaining of conspiracy theories and the ridiculous notion that post processing is cheating - what the hell does he think happens to a raw image when it goes through the editing pipeline?

PP makes it hard to evaluate the outcome fairly but your point is valid in that no one uses the files without the opportunity for some PP. I think the PP he complained about was the resizing of the file and downsampling of the files so that they appeared lower in noise than the actual files.

Grabbing RAW samples and doing your own processing and comparing results is probably the best method to determine how well each camera's output works for you.

[Ian Stuart Forsyth]

My assumption was that he must compare raw otherwise it makes a nonsense out of testing. Resampling should help to disguise noise.

But comparisons should be properly studied - should the raw data be handled individually to get the best out of each image or should it be left as is SOOC. Attached a quick comparison without any intervention in post by me, sharpening default ACR. Nikon D850 reduced to match K1 pixel count long edge. If you were to just judge on this then the left image (850) shows more resolution than the centre (K1) but the right most image shows best resolution and lowest noise. D850 noise is mainly luminance whereas there is a lot more colour noise in the standard K1 image and the image appreciably softer.

You have to also consider using the same constant exposure as different manufactures saturate the sensors differently, for Nikon at the same iso setting they expose 2/3 of a stop smaller giving additional headroom for the same iso

[UncleVanya]

You have to also consider using the same constant exposure as different manufactures saturate the sensors differently, for Nikon at the same iso setting they expose 2/3 of a stop smaller giving additional headroom for the same iso

Good point - but if we are going to that level I assume we need to account for T-stops not just F-Stops. This is getting complicated. Phew - just buy one and shoot already! LOL

(PS - Ian - thank you I am not saying your advice isn't welcome - I'm giving up and saying they both are great and few people will ever need to decide which is better for anything other than their collection of gear and budget.)

[Ian Stuart Forsyth]

Good point - but if we are going to that level I assume we need to account for T-stops not just F-Stops. This is getting complicated. Phew - just buy one and shoot already! LOL

(PS - Ian - thank you I am not saying your advice isn't welcome - I'm giving up and saying they both are great and few people will ever need to decide which is better for anything other than their collection of gear and budget.)

When we start going down the path we always have to consider what is actually being tested is it the cameras predetermined output or is it something else.

Take for example the D850 taking a shot at iso 63 it uses a exposure very close that of the K1 at iso 100, now if I ignored what the manufactures
wants you to use and use one stop larger exposure there is a difference

depending on the user that can be important

We also have to consider how much of the false detail located in the DPR corner is from having a sharper lens in that corner. If we stop the lens down to a level where diffraction can replace the badly needed AA filter what would we see in that part of the image ?

---

Last edited by Ian Stuart Forsyth; 12-04-2017 at 11:32 PM.

[normhead]

Well, going over to IR my favourite resource, the D850 definitely has more detail on non-pixel shift images. SO I'm going with the D850 is the better overall performance but the K-1 can top it in some circumstances.

But that's irrelevant. The K-3 was good for me, (and remember I sell prints up to 30" by 20" the K-1 is serious overkill. The D850 is ridiculous overkill. But,as I said, the final determination can't be made until I see a raw files comparison from the D850. I don't know if I'm dealing with more aggressive noise reduction in the Pentax jpeg engine, the small amount of increased detail in the D850 images looks like it could be compensated for by adding a bit of sharpening in raw, like less than 10%.

There is one thing I am absolutely certain of. For what you get, a D850 is not worth the price of 2 K-1s, for even the most demanding photographer. The only area the increased resolution could make a difference, because there is no detail in the D850 image that isn't in the K-1, would be possibly product photography, And in most product photography, pixel shift is easy to implement and the D850 is again on the short end, for twice the price.

I'm going with that until I see a proper test done with raw files.

Why is the guy in the video so enthusiastic. If you found a camera for half the cost that was capable of taking better images, wouldn't that excite you? My conclusion looking at the D850 images was for what you get, the camera is seriously over-priced.

Some see the D850 as the latest thing since sliced bread and a camera everyone should have. I see it as the camera to hold up to point out what a awesome value the K-1 is.

Using pixel shift you can getter images than this camera that costs twice the price. I guess some people like to waste their money.

---------- Post added 12-05-17 at 09:30 AM ----------

But it's not equally wrong to suggest that the results of artifacting occurring on 2D line drawings in some way will represent every day photography in 3D space?

Still waiting for an answer here TonyW. You're quick to point out the error in other's judgement criteria and slow to acknowledge the errors in your own.

Do you have even one piece of evidence to suggest artifacts created copying 2D line art can be applied to performance rendering 3D real life objects?

Since that's pretty much the basis of your argument, I assume you would have addressed it, unless of course this is based on conjecture.

Without addressing that the images you posted mean nothing. The real world does not depend on skinny little lines to create shadow effects. This is made even worse by your insistence, unsupported that pixel shift can't be used in all circumstances, yet the example you picked was one where pixel shift could be used every time. That point would be more believable if you'd picked an image where pixel shift couldn't have been used. IN actual fact, there are very few landscapes where pixel shift can't be used, even those with motion.

Of my keepers from my walk yesterday, 7 out of 13 keepers were pixel shift images, the ones that weren't were mostly pictures of my dogs. SO sure you can say "pixel shift can only be used in some circumstances.". I could counter and say, sometimes you can use pixel shift for the majority of your images." Fact is, on my walk yesterday more than half my images would have been noticeably worse using a D850, and of the images I'd actually consider printing large, 100%.



Full size, you can count the pine needles on the branches across the river. How much detail do you need? Is the best camera not the camera that gives you the chance to take the best image?

Your anti-Pentax bias is getting the better of you here.

[TonyW]

Well I thought we were discussing the results in the video.

Yes and what the guy said there was conspiracy and further he presented evidence to show that the K1 far superior for high ISO noise BEFORE Pixel Shift applied. Due to this I queried the validity of his statements and TBH sounded like pot calling kettle. The evidence presented in raw files I have compared show nothing worth bothering about pixel peeping at 100% and I would go as far as to say there is unlikely to be anything seen in comparing actual print.

I asked if the guy had released the raw to support his theories so far no reply.

You stated about 'wiping the floor' but you do not declare your criteria, if that is with Pixel shift ok, but cannot be supported without pixel shift as there is nothing to choose between them. If you disagree I will ask again for you to present your image comparisons as evidence.

But if you think d850 has better iso performance than K-1according to the video then great. I have no qualms. Enjoy your best thing since sliced bread.

It would be appreciated if in future you could refrain from trying to put words into my mouth.

I will let slide for now the seeming snide comment you made and direct you to the gist of what I am saying which is that there is no worthwhile difference to be seen with noise between these two cameras at most ISO I have looked at (WITHOUT PS!!). Please feel free to direct me and others to your source of raw file comparison in standard mode of your own or others

I have no axe to grind in this comparison as I am not currently interested in either the Nikon or Pentax so there is no such thing as best since sliced bread. What I am interested in is getting to the truth about differences, and if they are worthwhile for the individual but find it difficult to get to this due to fanboyism. I thought it mainly raised its ostrich like head in other fora such as Nikon v Canon but see that it exists here as well.

[normhead]

It would be appreciated if in future you could refrain from trying to put words into my mouth.

Right back at you.

But it's not equally wrong to suggest that the results of artifacting occurring on 2D line drawings in some way will represent every day photography in 3D space?

Still waiting for an answer here TonyW.

3rd time asked. I won't put words in your mouth, if you answer the question.
Otherwise I might take the liberty to suggest what you might be thinking if you refuse to enlighten us. You've been given lot's of opportunity to avoid having someone else "put words in your mouth."

This is not my bias, this is the plainly observed truth. It's your bias that keeps you from seeing it.

So no complaints if I make something up... right?

Or do you really expect everyone to just accept this ridiculous example of photography, an example so bizarre I can't find a single single image like it in over 30,000 files, and you expect us to accept that as proof of what camera has more or better resolution?

As I said, the results I have seen were not done with raw data and could be applicable only to jpeg shooters. The result you posted could only be of use to art copy artists, and they could use pixel shift 100% of the time.

Simple fact... the K-1 is the better camera system for the image you posted. There may be other images taken in different settings that show the superior image taken with a D850 elsewhere, out in the real world. Your assertion that this shows the D850 would be better somewhere else based on the two artifacts images, both with artifacts, just one worse than the other, is unsupported anywhere in your post. You are assuming that. But based on the images you posted, anyone would be a fool to buy a D850 instead of a K-1 for that purpose. Twice the cost, less resolution.

[Rondec]

From what I can tell, they are actually both really close in terms of dynamic range and noise, but Nikon has better resolution. How much bigger could you print a D850 file than a K-1 file? I have no idea, but my guess is that it isn't as much as what it sounds consider you are going from 36 to 45 megapixels.

The reasons to get the D850 have to do with buffer size, auto focus module, video performance, and the other little details that Nikon improved as compared to the D810. If you put the D850 sensor in a D810 body with D810 specs, I doubt most D810 users would upgrade.

[TonyW]

Well, going over to IR my favourite resource, the D850 definitely has more detail on non-pixel shift images. SO I'm going with the D850 is the better overall performance but the K-1 can top it in some circumstances.

SO once again you are in violent agreement with my assesment.

I'm going with that until I see a proper test done with raw files.

And yet when presented with the source of such test files you steadfastly refuse to even look?

Why is the guy in the video so enthusiastic.

Why is he presenting data in the first part of the video about excess noise and not releasing his raw files for examination. Comparison raw images have been put up here and elsewhere that shows his assesment to be flawed.

Still waiting for an answer here TonyW. You're quick to point out the error in other's judgement criteria and slow to acknowledge the errors in your own.

Do you have even one piece of evidence to suggest artifacts created copying 2D line art can be applied to performance rendering 3D real life objects?

Since that's pretty much the basis of your argument, I assume you would have addressed it, unless of course this is based on conjecture.

Without addressing that the images you posted mean nothing. The real world does not depend on skinny little lines to create shadow effects.

You have failed to prove errors in my judgement so far and you steadfastly hold onto your beliefs without actually looking at the image I posted and compared the noise performance without PS. You are free to download the same raw images from Dpreview. The fact that the comparison centred on 2d is immaterial and artifacts were not what we were looking at but noise and secondary resolution.

Your anti-Pentax bias is getting the better of you here.

This is rather a dissappointing response from you and without the benefit of any supporting evidence and is truly well off the mark and a likely sign of true fanboyism.

You are free to see for yourself from published raw data the noise signatures of raw at various ISO's. And I will say once again in case you missed it from several times already mentioned. The K1 is a great camera as is the D850 and the K1 scores points for IQ for sure when Pixel Shift employed (as do all cameras that use PS).

The presentation of noise being better controlled in the K1 (without PS) is as far as I can tell from using actual raw comparisons flawed and brings into question the reviewers knowledge, integrity and purpose. To present such highly subjective views without giving the audience access to the original data is pretty poor.

I am only prepared to do this one more time (unless presented with real alternative comparative data). Attached is the noise signature of the K1 and the D850 at ISO 12,800 (from images freely available to you! The images are non PS). These images brought into ACR and all noise and sharpening set to zero.

What I would like to know is which one is better than the other and for what reasons, and when printed (actual print size shown in reduced image) can we see the difference?

EDIT: Just seen this comment and I really must have missed this nonsense the first time around

As I said, the results I have seen were not done with raw data and could be applicable only to jpeg shooters

Really and you think you can judge by using prebaked data that by its nature has discarded a load of information. Please revisit when you can present proper data and a reasoned argument based on the result of that data comparison

---------- Post added 12-05-17 at 10:17 AM ----------

PP makes it hard to evaluate the outcome fairly but your point is valid in that no one uses the files without the opportunity for some PP. I think the PP he complained about was the resizing of the file and downsampling of the files so that they appeared lower in noise than the actual files.

Grabbing RAW samples and doing your own processing and comparing results is probably the best method to determine how well each camera's output works for you.

I understand that some do not like the downsample or upsampling of files to compare but ultimately it is going to happen anyway once the acid test The Print is made.

I agree that getting the raw samples and processing yourself is certainly the best method and the one that I used here to arrive at my own conclusions. Then we get into the realms of which converter is best to use etc

-

[normhead]

OK, I give up, this will be the fourth time I've asked this question, and you keep dodging it.

Why do you think artifacts in 2D line drawings predict performance in real world 3D objects? Do you have any evidence to support this hypothesis.

[TonyW]

You have to also consider using the same constant exposure as different manufactures saturate the sensors differently, for Nikon at the same iso setting they expose 2/3 of a stop smaller giving additional headroom for the same iso

I agree and AFAIK the samples are shot under the same conditions using the same exposure. The difference being the lenses used and the fact that they may apply WB correction and exposure correction to bring the images into the same. Not ideal of course, but I believe acceptable to most.

Typical example:

Image comparison: Digital Photography Review

[normhead]

OK, I looked at the comparison as posted at Every ISO from 25600 down to 100, and i have rarely seen so little difference as there is between the D850, K-1 and Canon 5DMkIII. Most of them are indistinguishable from each other. I'm even less impressed with the D850. Based on the IR still lifes I was expecting better.

I guess unless you are trying to do thread counts on fabrics on straight OOC jpegs, there really is no difference. Which also points out how meaningless line drawings are in evaluating real world images.

Do you really see some of those images as better than the others? I don't think I could identify which was which in a blind test. SO, pretty much time wasted even looking. Didn't change my position one iota, and pushed me back saying anything a D850 can do a K-1 can do, but not everything a K-1 can do can be done with a D850.

I guess that's the trouble with subjective evaluations.