[biz-engineer]

Quote: Full frame basically gives you one stop improvement over APS-C. Only if you sacrifice DoF for an improvement in SNR. I'm starting to wonder if people say this just to irritate me. APS-c ƒ5.6, 100 ISO, 1/100s, is the same as FF ƒ8, 200 ISO, 1/100s. Same total light, same Signal to Noise Ration, same DoF, same shutter speed, = same picture. Where is the extra stop better for FF?

Ok you are right. In situations where DoF is not an issue, full frame gives you one stop improvement over APSc.

[normhead]

Ok you are right. In situations where DoF is not an issue, full frame gives you one stop improvement over APSc.

Music to my ears....

The thing is, for landscape guys, DoF is often important. For us, all we can get is higher resolution... a metric of unproven value in the overall scheme of things. Sometimes I think this is a safety valve issue. "I'll go higher res, just case it matters, because sometimes it might."

[biz-engineer]

Ok you are right. In situations where DoF is not an issue, full frame gives you one stop improvement over APSc. Music to my ears....

That would mean that beyond the 70-200 (300mm or more FL) , there are more drawbacks than benefits of using a full frame camera if you have an APSC such as a K-3.

[Rondec]

Music to my ears....

The thing is, for landscape guys, DoF is often important. For us, all we can get is higher resolution... a metric of unproven value in the overall scheme of things. Sometimes I think this is a safety valve issue. "I'll go higher res, just case it matters, because sometimes it might."

In situations where either depth of field isn't important or, you have remembered to bring a tripod. Once you remember your tripod, you can have whatever depth of field at whatever iso you want.

[Ian Stuart Forsyth]

Then why do all the sites that have comparison images looks so similar? Why do we have no clear cut examples? This is nonsense. You assume there are visible imperfections in the lens to be magnified. Wouldn't that be the first thing that would bet established in such a theory? If the visible imperfections round off to zero, then twice zero is zero. But I'm game....

Sometimes, I wonder if folks are just making stuff up, ... or just taken in by and repeating stuff other people made up. "It sounds logical, so, there fore I'll repeat it as fact."

Can you even tell which is which, forget about find lens imperfections in one, that have been magnified in the other?

K-3 or D610, tell us which is which and show us these imperfections.



Forgive my impatience, but I've heard this kind of hogwash repeated over and over for about 4 years now, it's getting old.

We have no idea how the above images where processed and really not telling us the resolution difference between formats , more importantly this is dwarfed by the way the images where captured and how this correlate’s with how we take images in the really world of photography.
When IR decided to hold the same F/8 for both FF and APSC that puts the FF images with over a stop less DOF than the apsc camera and Just as important the position of the camera as it was moved also having further nulling images for resolution testing.

How does this play out when using the IR studio image as a gauge to resolution ?

About as useful as using a K3 with a 70mm at F/5.6 taking a shot at 5 meters away from the target and then using the same K3 this time with a 45mm lens at F3.7 at the same time moving the camera 1.66m closer to the target. We have 2 very different photographs with different DOF, perspective and most importantly absolute blur( the rate at which blur increases as you move away from the point of focus)

Why is this important?

Resolution, or more importantly how we see resolution (degree of blur). Here we have 2 images, 1 FF 70mm F/8 and 1 APSC 70mmF/8. The FF image is captured using a shallower DOF and on top of that the rate of Blur (absolute blur) one see’s increases at a much greater rate than that of the APSC camera.

Conclusion how can one use IR as a basis to evaluate resolution when both images have such different characteristics in the blur and our goal is to try and measure some kind of resolution from that blur?

Like I said "About as useful as using a K3 with a 70mm at F/5.6 taking a shot at 5 meters away from the target and then using the same K3 this time with a 45mm lens at F3.7 at the same time moving the camera 1.66m"

In really world photography some of the tools we use to capture images are movement, DOF, FOV, perspective and absolute blur. Why not base how we rate resolution on these fundamental properties we use in photography? Not some arbitrary images when we cut 1/3 of the DOF, alter the perspective that also alters the rate at which blur occurs.

“ sure my 1986 civic just as fast as your 911, hey before we race could I load your 911 with 2000 pounds of rocks? ”

Overlook all that is said above and let’s have a look at your own resource for resolution testing IR.

Just as I have pointed out to you before (Your reference )

See no difference ? I do
K3 d610 images are converts of raw with no sharpening or corrections using DXO

And sharpened both using topaz deblur 0.83 radius at the same time

I cannot help but notice that the D610 is sharper throughout the shallower DOF and even improves more as you move towards the camera, hinting that its front focused
I have never liked the d610 as it had a rather strong AA filter
Now the D750 all the images are converts of raw with no sharpening or corrections using DXO

topaz deblur 83 applied to the above photo

even with the shallower DOF and more absolute blur the D750 is better across the DOF

both scaled to a 6mp image image was scaled as a single image using the above photo with the same processing done as a single image

At 6mp image you can still see an advantage and I was done with 6mp images in 2005

Both scaled to 2.5mp image was scaled as a single image using the original photo with the same processing done as a single image

Still see a slight advantage and seldom ever use a 2.5mp image
Clearly FF with a shallower DOF and more absolute blur was better (even when the framing was at a different size between the images the full frame needs to be closer to the target), Now the big question that needs to be answered, what if they were shot using the same FOV DOF and perspective I am think we would see an even larger difference.

[normhead]

Actually, it's not that clear, the images are pretty much identical, I'd be quite happy to show these images to people to show them how little difference there really is. I'd say, listen, if you squint really hard you can see a difference. I can't think of a single successful image that required squinting really hard. But it is typical of the kind of work people tend to bring to these arguments. Little differences magnified to be hugely important.

I make a point of taking the images straight from IR with a screen copy, because I don't know if DxO or any other software does as good a job on K-3 images as it does on D610 or D750 images. I trust that whatever IR does, they try to be fair to both cameras. I see no advantage to running the images through additional software of unknown effect, that's just another unknown added to the equation. So, I really can't recommend your methodology, except for maybe committed users of DxO.

[Rondec]

We have no idea how the above images where processed and really not telling us the resolution difference between formats , more importantly this is dwarfed by the way the images where captured and how this correlate’s with how we take images in the really world of photography.
When IR decided to hold the same F/8 for both FF and APSC that puts the FF images with over a stop less DOF than the apsc camera and Just as important the position of the camera as it was moved also having further nulling images for resolution testing.

How does this play out when using the IR studio image as a gauge to resolution ?

About as useful as using a K3 with a 70mm at F/5.6 taking a shot at 5 meters away from the target and then using the same K3 this time with a 45mm lens at F3.7 at the same time moving the camera 1.66m closer to the target. We have 2 very different photographs with different DOF, perspective and most importantly absolute blur( the rate at which blur increases as you move away from the point of focus)

Why is this important?

Resolution, or more importantly how we see resolution (degree of blur). Here we have 2 images, 1 FF 70mm F/8 and 1 APSC 70mmF/8. The FF image is captured using a shallower DOF and on top of that the rate of Blur (absolute blur) one see’s increases at a much greater rate than that of the APSC camera.

Conclusion how can one use IR as a basis to evaluate resolution when both images have such different characteristics in the blur and our goal is to try and measure some kind of resolution from that blur?

Like I said "About as useful as using a K3 with a 70mm at F/5.6 taking a shot at 5 meters away from the target and then using the same K3 this time with a 45mm lens at F3.7 at the same time moving the camera 1.66m"

In really world photography some of the tools we use to capture images are movement, DOF, FOV, perspective and absolute blur. Why not base how we rate resolution on these fundamental properties we use in photography? Not some arbitrary images when we cut 1/3 of the DOF, alter the perspective that also alters the rate at which blur occurs.

“ sure my 1986 civic just as fast as your 911, hey before we race could I load your 911 with 2000 pounds of rocks? ”

Overlook all that is said above and let’s have a look at your own resource for resolution testing IR.

Just as I have pointed out to you before (Your reference )

See no difference ? I do
K3 d610 images are converts of raw with no sharpening or corrections using DXO

And sharpened both using topaz deblur 0.83 radius at the same time

I cannot help but notice that the D610 is sharper throughout the shallower DOF and even improves more as you move towards the camera, hinting that its front focused
I have never liked the d610 as it had a rather strong AA filter
Now the D750 all the images are converts of raw with no sharpening or corrections using DXO

topaz deblur 83 applied to the above photo

even with the shallower DOF and more absolute blur the D750 is better across the DOF

both scaled to a 6mp image image was scaled as a single image using the above photo with the same processing done as a single image

At 6mp image you can still see an advantage and I was done with 6mp images in 2005

Both scaled to 2.5mp image was scaled as a single image using the original photo with the same processing done as a single image

Still see a slight advantage and seldom ever use a 2.5mp image
Clearly FF with a shallower DOF and more absolute blur was better (even when the framing was at a different size between the images the full frame needs to be closer to the target), Now the big question that needs to be answered, what if they were shot using the same FOV DOF and perspective I am think we would see an even larger difference.

I would say they are awfully close. Certainly the difference is not enough to make or break this image. I understand that on the forum folks find it necessary to attack different size sensors. "Full frame is the best!" "Not if you need more depth of field." And so on. The reality is that at low iso there isn't a lot of difference. We forumites wax eloquently about these things, but if you take the final product, in many cases there is little enough difference.

Few photographers are truly wringing every ounce of dynamic range or sharpness out of their sensor. Few are shooting to the limit with regard to iso. Few truly benefit from extremely shallow depth of field. Most images are made in the middle with only a few photos being made at the extremes.

I know that you do shoot at the extremes with your wildlife photography and have spent more money in gear than most APS-C photographers would spend in twenty years, but not everyone needs that kind of performance.

[Mattox]

Well with all this technical discussion and negativity going on I just spent the morning out at 6am taking waterfall panorama's with a friend. I used my k5 and he used his Nikon 800. From first hand results my k5 took fantastic shots. I don't know if it was me or the k5, but all this technical chatter, does it really make that much of a difference? What's more Important is to go out and shoot I say

[normhead]

Well with all this technical discussion and negativity going on I just spent the morning out at 6am taking waterfall panorama's with a friend. I used my k5 and he used his Nikon 800. From first hand results my k5 took fantastic shots. I don't know if it was me or the k5, but all this technical chatter, does it really make that much of a difference? What's more Important is to go out and shoot I say

Exactly... images missed while talking on line are gone forever. You never get them back.

[GlassJunkie]

Time to be the heretic.. It's not sensor size vs. Print size, its pixels. A 24mp APSc mapped to print at A2 will print exactly the same resolution quality as FF (SNR aside)...period. Empirical math. SNR (particularly with high ISO) may be a concern, based on commercial/aesthetic requirements... Can't argue that... Value proposition is cost/ size/ weight vs SNR. Also (never talked about by FF supporters) is cross talk at a given pixel pitch (IMHO APSc wins) vs materials capabilities. As sensors evolve, SNR @ high ISO will likely disappear... So FF vs APSc as a debate will look like Ford vs Chevy, or Glenlivet vs Glenfiddich... <img src="https://www.pentaxforums.com/forums/images/smilies/smile.png" border="0" alt="" title="Smile" smilieid="1" class="inlineimg" /> Sorry, couldn't help it...<img src="https://www.pentaxforums.com/forums/images/smilies/wink.png" border="0" alt="" title="Wink" smilieid="4" class="inlineimg" />

[Glorfindelrb]

How someone calling himself GlassJunkie can ignore lens part of the real resolution ? There is no perfect lens out there and the resolution you should look at is more the PMpix evaluated by DxO or a similar measurement than a number of pixels. They show that no lens is resolving a 24MP APS-C sensor. The best lens for sharpness measured on a K3 is the Sigma 35mm F1.4 with 16 Perceptual Mpix on a K5IIs it's 13 PMpix.
I feel the DA 70 ltd being pretty sharp and it's rendering 11 PMpis on a K3 and 10 PMpix on a K5IIs.

Adding pixels helps but less and less. Thinking the lens resolution as a fixed limit and the lens to be perfect before reaching this limit is wrong. Getting 24 PMpix is feasible with a good glass on a 36MP FF (the Sigma 35mm F1.4 --> 30 PMpix), it's maybe possible with a 30+MP APS-C but the glass is not on the market and would cost A LOT.

The true question is "do we need that resolution ?" and the answer is probably no in 95% of our shots. The size/weight/cost arguments are also real.

[gazonk]

Getting 24 PMpix is feasible with a good glass on a 36MP FF (the Sigma 35mm F1.4 --> 30 PMpix)

DxO actually measured the Sony FE 90mm macro to 32 P-Mpix! (and to 18 P-Mpix on the A5000)

[D1N0]

Bigger pixels on FF will be easier to resolve. 36mp on ff is a comparable pixel size to 16mp on aps-c. DXO measures the resolution of lens camera combinations. So when you want to compare lens resolution you have to pick camera's with the same mp/sensor size.

[biz-engineer]

Bigger pixels on FF will be easier to resolve. 36mp on ff is a comparable pixel size to 16mp on aps-c. DXO measures the resolution of lens camera combinations. So when you want to compare lens resolution you have to pick camera's with the same mp/sensor size.

Sure, a larger sensor relax the constraints on the lens.

[Glorfindelrb]

DxO actually measured the Sony FE 90mm macro to 32 P-Mpix! (and to 18 P-Mpix on the A5000)

I took the Sigma because it was my example on K3, there is a lot of recent lenses above 24 P-Mpix on 36+MP FF cameras. (It's 32P-Mpix on a FF A7R with 36MP)

Bigger pixels on FF will be easier to resolve. 36mp on ff is a comparable pixel size to 16mp on aps-c. DXO measures the resolution of lens camera combinations. So when you want to compare lens resolution you have to pick camera's with the same mp/sensor size.

Thats my point.