[BigMackCam]

And do the comparison with a less than ideal lens - say an old zoom - I think that is when I notice it most.

Putting aside the explanations you and Ian kindly provided - which I promise I'll come back to (having just made a big mug of strong, freshly ground, calculation-busting coffee ) - this sounds like the same difference I notice when using the same less-than-stellar lens on my Samsung GX-10 vs the K-3. The GX-10's 10MP sensor is more forgiving of the lens... which is due to the bigger photosites / lower density, yes?

[Rondec]

You all are going over my head quite a bit.

I guess I feel this way. If I am going to crop most of my photos to APS-C or close to it -- even if I have similar number of pixels left (as with a K5/K-1 comparison) I'm still better off using an APS-C camera. The frame rate will likely be faster (the K-1 is 6.5 fps in crop mode) and you aren't paying for a bunch of sensor surface area that you aren't using. So, if your main interest is sports and wildlife and you can't afford a 400mm lens, it is very reasonable to choose an APS-C camera as a good compromise.

I do think it is clear that when pixels get really dense (like on cell phone cameras) it gets much harder to get pixel level sharpness and using a smaller sensor doesn't really make your lenses longer, despite all of the micro four thirds folks arguing to the contrary. An iphone X has lenses that are 4.25mm and 6mm in length. So that 6mm lens is supposed to give an equivalent image to a 50mm lens on a K-1. Not sure about that but I guess at least with regard to angle of view that is correct...

[GUB]

Putting aside the explanations you and Ian kindly provided - which I promise I'll come back to (having just made a big mug of strong, freshly ground, calculation-busting coffee ) - this sounds like the same difference I notice when using the same less-than-stellar lens on my Samsung GX-10 vs the K-3. The GX-10's 10MP sensor is more forgiving of the lens... which is due to the bigger photosites / lower density, yes?

It is probably more accurate to say that the increased density of the K3 pixels hasn't boosted the IQ as much as expected due to the low resolution of the lens. And the ccd has brought qualities to the image (compared to cmos) that over-ride the pixel gain.
And maybe this doesn't happen so much with a sharper lens (the pixels have more to work on).
And bed time here!

[GUB]

I guess I feel this way. If I am going to crop most of my photos to APS-C or close to it -- even if I have similar number of pixels left (as with a K5/K-1 comparison) I'm still better off using an APS-C camera. The frame rate will likely be faster (the K-1 is 6.5 fps in crop mode) and you aren't paying for a bunch of sensor surface area that you aren't using. So, if your main interest is sports and wildlife and you can't afford a 400mm lens, it is very reasonable to choose an APS-C camera as a good compromise.

It is funny how we see the same thing but different ways. Yes for birding etc the lack of reach in the K1 is annoying but you can just crop it and have a K5 image. And in macro work the loss of dof is frustrating until you just pull back and create an apsc image .

[Rondec]

It is funny how we see the same thing but different ways. Yes for birding etc the lack of reach in the K1 is annoying but you can just crop it and have a K5 image. And in macro work the loss of dof is frustrating until you just pull back and create an apsc image .

I like using a K-1 a lot, but I seldom use really long lenses. My longest full frame lens is the DFA *70-200. Longest APS-C lens is the DA 55-300 (original version). If I'm going somewhere where I plan to shoot telephoto a lot, I do use the 55-300 on the K-3. It isn't perfect, but I get faster frame rates, quicker buffer emptying and smaller file sizes with the K3 versus the K-1. If I can use the 70-200 with minimal cropping, it will give significantly better image quality on the K-1 than the 55-300 on the K-3.

(Of course, I can also shoot the DFA *70-200 on the K3 too).

[someasiancameraguy]

Get the APS-c if you want responsiveness. Get the FF if you want ultimate IQ. That's basically the whole debate :P

[Pål Jensen]

Get the APS-c if you want responsiveness. Get the FF if you want ultimate IQ. That's basically the whole debate :P

Not the whole. You forgot size, weight and cost....

[Serkevan]

That is right but it is the lens and its limitations that is the main driver in the resolution. That is Ian's point.

That also plays an important part, of course. At a given final image size the smaller sensor will lead to more magnification, which means that the limits of the lens' resolving power are stretched further.

[Hidrieus]

I recalled reading an article in an Outdoor Photographer magazine, and found it in my small stack of print magazines. It's in the October 2019 edition, and available online as of January 2020:

Does Camera Sensor Size Matter? - Outdoor Photographer

The author, Josh Miller, a professional outdoor photographer, discusses image quality and other attributes of various camera formats, and suggests that all formats have their advantages. Here are several quotes:

"Fast forward 10-plus years, and full-frame is still the primary choice of most professionals like myself. But in these intervening years, camera makers have really been pushing the limits of what smaller sensor cameras can do, allowing for less-expensive and lighter-weight cameras and lenses and sensors that, though smaller, are still capable of professional-level work."

"In my experience with current cameras in these various formats, any camera with at least 20-megapixel resolution will make great prints up to 20×30-inches or larger, assuming you are shooting at reasonable ISOs with quality lenses and good technique."

"With all the sensor formats being so good, I wouldn’t actually make sensor size my No. 1 determining factor when choosing to invest in a system. I would decide how good is good enough in terms of image quality and then look more broadly at the lenses and accessories being offered with the system."

"When choosing your next camera, ask yourself these questions: Are you willing to pay the costs to have the absolute best image quality on the market, or have we reached a point where your image quality needs are met, and it’s more about ease of use and making photography fun? Is that full-frame or even medium-format camera really worth the extra hassle of lugging it into the field, or will something smaller do the trick? In the end, we all know the best camera is the one you are willing to carry to the most locations."

- Craig

https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2Ftse3.mm.bing.net...6pid%3DApi&f=1

For me the issue is not image quality, noise or dynamic range that defines the difference between 16x24 versus 24x36. It is lens distortion. A 85mm f/1.8 at 24x36 can not be substituted by a 56mm f/1.4 at 16x24.

[BarryE]

I recall reading a technical article a while back regarding the different algorithms available to developers, many patented and proprietary. One that struck me as interesting confused the visual metaphor of sensor sites being buckets of different depths and surface area. The buckets concept works quite well to explain the ability of the different sensor technologies to gather incoming photons. Large, deep buckets give greater dynamic range - think medium format. Small, shallow buckets - think smartphones. A large deep bucket takes longer to become saturated and highlight details lost than a smartphone. However, if the processor is fast enough the bucket at a pixel site can be augmented by their adjacent pixel sites/photo diodes. So with sophisticated algorithms a small bucket can hold highlight detail by utilising neighbouring silicon. The whole concept of buckets soon becomes rather weak as processing power offsets the physical limitations imposed by small sensors. Thus the whole picture becomes more and more complicated as smartphones eat away at the advantage of large slabs of silicon on DSLR/MF cameras with greater research and development dollars being available to them.

Well at least I thought it was interesting ...

[GUB]

That also plays an important part, of course. At a given final image size the smaller sensor will lead to more magnification, which means that the limits of the lens' resolving power are stretched further.

But if the K3new turned out to be 36mp apsc then it would require no more magnification than the K-1. But the detail supplied by the lens would still be less than half that of the K-1. (see post 100 - megadots).

[BigMackCam]

https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2Ftse3.mm.bing.net...6pid%3DApi&f=1

For me the issue is not image quality, noise or dynamic range that defines the difference between 16x24 versus 24x36. It is lens distortion. A 85mm f/1.8 at 24x36 can not be substituted by a 56mm f/1.4 at 16x24.

Not wanting to bring up so-called "equivalence" (I've avoided it for the second time in a week... I still bear the scars from previous occasions ), but...

The reason for the perspective distortion in those linked examples is that the different focal lengths were used on the same format camera, with the shooting distance adjusted so that the subject's head fills the frame. If, instead, you shoot a full frame camera with the 85mm f/1.8 lens then, from the same distance, shoot an APS-C camera with the theoretical 56mm f/1.4 lens you refer to, the subject's head will fill roughly the same amount of the frame and you'll get more or less the same perspective distortion. It's the choice of lens for the format you shoot that dictates the distance you need to shoot at, and the distance (not the lens) that results in a lengthening or flattening of the subject's features.

What you couldn't achieve with that theoretical 56mm f/1.4 lens on APS-C is the very shallow depth of field obtained with the 85mm f/1.4 shot wide open on full frame at the same distance. So, in that sense, full frame has an advantage in giving you the ability to achieve shallower depth of field when all other elements of the shot are the same...

[Kevin B123]

But if the K3new turned out to be 36mp apsc

I think we know it is a 26MP camera:

Pentax K-3 III - Pentax K-mount DSLRs - Pentax Camera Reviews and Specifications

[mikesbike]

https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2Ftse3.mm.bing.net...6pid%3DApi&f=1

For me the issue is not image quality, noise or dynamic range that defines the difference between 16x24 versus 24x36. It is lens distortion. A 85mm f/1.8 at 24x36 can not be substituted by a 56mm f/1.4 at 16x24.

Yep, lens usage is a big factor. I've had only a very small opportunity to put my new K-1 II to use, but I have been taking along my KP as well, with a corresponding quality lens to get the same framing, and have found so far the quality between the two comes off remarkably close. The KP tracks behind the K-1 II very well. Not surprisingly, this agrees with other reports. But when it comes to trying to get with certain APS-C lenses what can be had with certain FF lenses, in some cases, it just can't happen.

But then, when it comes to fast action or longer telephoto reach, APS-C can be advantageous.

[Ian Stuart Forsyth]

But if both formats have the same amount of pixels then the pixel density is different, right?

This is correct

---------- Post added 11-30-2020 at 08:27 PM ----------

I should at least see the difference.

It's rather hard to find data as many times we have one camera with a AA filter and the other without
Here is the Tamron SP 35 1.8

here it is on the D7100 at 24mp

And here it is on the D3X at 24mp, and we know this camera has rather strong AA filter
What is interesting to see is that when this lens is shot wide open at f1.8 or even stopped down to F11 that this combo can resolve nearly as much resolution as this lens at its best f-stop a on a dx body

We can also give you a starker contrast as to the difference plays out when using lenses with the same FOV



Here is the 24mm 1.8 on a dx 24mp camera



Here is the 24-120mm zoom and the FX 24mp body
For me this is rather poor preforming large FL zoom and it holds its own at similar DOF to the DX body and prime