[RobA_Oz]

I think this thread needs another update to the K1 website.

[Fogel70]

Obviously not you, Fogel ... you have common sense! :-)

But what is the point in comparing crops with same numbers of pixels from from different cameras? It's like comparing same numbers of cylinders on engines and saying that a 4-cyl 200bhp engine has the same performance as a 8-cyl 400bhp engine if each cyl perform the same.

This type of comparison will favor lower pixel count as you will be using a larger part of the sensor when comparing, and sensor size will be irrelevant. So a smartphone could outperform a medium format camera if the smartphone use larger pixels.

I believe most users like using comparison method that will mirror real world use of the cameras.

[Gimbal]

Yet both of these have this Normalization to 8Mp crap!

Please look at the Screen rather than Print stats, and you will find the K-5 is also similar to the mighty 810 in dynamic and tonal ranges. This will be a rude shock to certain forum members.

That each pixel behaves pretty much the same isn't chocking, But you don't seem to understand that having more then twice as many of them does have an impact on the final quality of the picture.

So the only chocking part is that you don't get the idea behind the normalization "crap".

[mee]

chocking? Did you mean choking or shocking?

[normhead]

That each pixel behaves pretty much the same isn't chocking, But you don't seem to understand that having more then twice as many of them does have an impact on the final quality of the picture.

So the only chocking part is that you don't get the idea behind the normalization "crap".

My best guess was that comparing a D810 and K-5 or D810 image was that you wouldn't be able to tell the difference until you use more than 3000-x2000 pixels, although it could be as high as 3300 x2100 (7 MP). On images cropped to less than 6 MP you can't tell the difference.

I've done no work on a K-3 , so but since the K-3 has about 23% percent more resolution than a K-5 then I'd expect that the number for a K-3 would be closer 4000x3000 pixels, or 10-12 MP.

SO I expect you won't be able to tell the difference between a D810 image and a K-3 image on a 4k screen, unless you crop.

Since the quality of the image is extremely dependant on the quality of the output device (the camera is probably less than half the equation) , and whether or not it makes efficient use of added resolution, talking about a D810 image being better than a K-3 or even a crop D810 image needs some parameters. It's not generally true, because generally we are looking at 1024 or at the most 1280 images,where I can assure you it's a very rare circumstance where it makes any difference at all. But if you have an actual set of circumstances where you can show it's true, I'm all ears and eyes.

jsherman has one set of comparison images, where he shot an environmental portrait of his son, @ 50mm on a 35mil sensor was more pleasing than 35mm shot on APS-c, just because the bokeh was so much more pleasing. But that's one pair of images out of millions of posted images, so if anything, I'd use that example of how rare it is that you benefit from shooting 35mil over APS-c.

[beholder3]

But what is the point in comparing crops with same numbers of pixels from from different cameras?

That each pixel behaves pretty much the same isn't chocking, But you don't seem to understand that having more then twice as many of them does have an impact on the final quality of the picture.

It depends on the magnification which one feels is his normal use case.

If you compare differing magnifications from sensor to viewing (explanation see my post here https://www.pentaxforums.com/forums/16-pentax-news-rumors/310344-pentax-645z-...ml#post3473157) = viewing the same "print size" from different sensor sizes at same viewing distance, then you are right and clackers is wrong. That is usually the typcial approach cited for comaprisons, even though it has a big BUT:
This usually throws away any pixel count advantages on resolution = no visible difference between 12MP and 36MPx. If you paid for 36 Mpx then it was a waste.

The other scenario is the same massive magnification for differing sensor sizes with same pixel size(!) (e.g. K5 vs D800). This is usually cited as the "100% view" on screen. You do need a strong magnification of the 36Mpx FF sensor to be able to see/benefit from the extra resolution. Sadly this automatically increases perceived noise to the point where the "FF advantage" is nil*.

It all does depend on the assumed viewing mode. And no viewing mode is better than the other, as they all have their reasons for some.

*It is nil anyway for apples-to-apples comparisons of equivalent pictures at same viewing distances same print sizes. It is only not zero when losing sharpness/image quality in the depth of field as a trade off.

[Gimbal]

chocking? Did you mean choking or shocking?

Try to figure it out. If you don't succeed, ask an adult.

---------- Post added 01-01-2016 at 05:57 PM ----------

Since the quality of the image is extremely dependant on the quality of the output device (the camera is probably less than half the equation) , and whether or not it makes efficient use of added resolution, talking about a D810 image being better than a K-3 or even a crop D810 image needs some parameters. It's not generally true, because generally we are looking at 1024 or at the most 1280 images,where I can assure you it's a very rare circumstance where it makes any difference at all. But if you have an actual set of circumstances where you can show it's true, I'm all ears and eyes.

I agree in so much that if you don't push against any boundaries you might not se any improvements at all. But is it really any point in comparing anything if you set up the rules such that no improvements can show?

Like comparing a -66 VW Beetle with a new Porsche and it turns out that they are equally fast, when both are driven at 50km/h that is.

---------- Post added 01-01-2016 at 06:05 PM ----------

The other scenario is the same massive magnification for differing sensor sizes with same pixel size(!) (e.g. K5 vs D800). This is usually cited as the "100% view" on screen. You do need a strong magnification of the 36Mpx FF sensor to be able to see/benefit from the extra resolution. Sadly this automatically increases perceived noise to the point where the "FF advantage" is nil*.

You still have a lot more pixels which means a lot more information. But if you crop them away to end up with the same print size, then you have thrown away the advantage.

---

Last edited by Gimbal; 01-01-2016 at 11:27 AM.

[Fogel70]

It depends on the magnification which one feels is his normal use case.

If you compare differing magnifications from sensor to viewing (explanation see my post here https://www.pentaxforums.com/forums/16-pentax-news-rumors/310344-pentax-645z-...ml#post3473157) = viewing the same "print size" from different sensor sizes at same viewing distance, then you are right and clackers is wrong. That is usually the typcial approach cited for comaprisons, even though it has a big BUT:
This usually throws away any pixel count advantages on resolution = no visible difference between 12MP and 36MPx. If you paid for 36 Mpx then it was a waste.

The other scenario is the same massive magnification for differing sensor sizes with same pixel size(!) (e.g. K5 vs D800). This is usually cited as the "100% view" on screen. You do need a strong magnification of the 36Mpx FF sensor to be able to see/benefit from the extra resolution. Sadly this automatically increases perceived noise to the point where the "FF advantage" is nil*.

It all does depend on the assumed viewing mode. And no viewing mode is better than the other, as they all have their reasons for some.

*It is nil anyway for apples-to-apples comparisons of equivalent pictures at same viewing distances same print sizes. It is only not zero when losing sharpness/image quality in the depth of field as a trade off.

If you want to compare the output of the cameras you need to normalize the output, It's possible to combine both method above, FI by viewing the D800 at 100% and K-5 at 225% (or K5 at 100% and D800 at 44%),

By viewing both at 100% you are not comparing image quality between then, you compare two different crops of the final image. What is the purpose of this comparison? What conclusion can you draw from this comparison?

[normhead]

If you want to compare the output of the cameras you need to normalize the output, It's possible to combine both method above, FI by viewing the D800 at 100% and K-5 at 225% (or K5 at 100% and D800 at 44%),

By viewing both at 100% you are not comparing image quality between then, you compare two different crops of the final image. What is the purpose of this comparison? What conclusion can you draw from this comparison?

It is a valid comparison, because you can only use what your output device needs. If the most your output device, printer or screen can handle is 72 DPI, then it doesn't matter if your camera produces 100 dpi, or 300 dpi. People seem to want to keep these discussions isolated from discussions of printer and screen qualities, and that's where the nonsense starts.

Like comparing a -66 VW Beetle with a new Porsche and it turns out that they are equally fast, when both are driven at 50km/h that is.

Exactly... if what you want to do is drive at 50 km per hour because of traffic, and you have 4 people and baggage to carry, the Porsche is going to really suck. I mean those poor folks are going to be excruciatingly uncomfortable.
They'll be uncomfortable in the Beetle, they'll be deformed by the Porsche.

The problem is, the object in photography is to be able to create output beyond the camera. There is absolutely no benefit to producing "quality" that your output device can't utilize.

Think a screen that's 1280x800.
The absolute most it can display is 1280 distinct lines horizontally and 800 distinct lines vertically. Choosing a camera that will produce 7000 lines horizontally (D810), over one that will produce 6000 (K-3), lines horizontally or even the 5000 line of a K-5 , doesn't increase the resolution of the out put device, and makes no difference at all to the output. It has to reduce both images by factors of 4 or 5. The difference is, digital deals with reduction not enlargement. The problem many photographers have is they think this is still film, where you enlarge your image to display it at a visible size. In digital, we are reducing the image most of the time.

The film equivalent would be taking an 8x10 film image reducing it to 35 film size, and then comparing the images. If what you want is a 35 mm contact sheet type image, shooing the original in 8x10 film offers no advantage. When you reduce the size, you lose all the advantages of the added resolution in the original.

IN film there is that cutoff point where, if all you are going to make is a 4x6, you may as well shoot 35mm instead of 8x10.
In digital it's about 24 inches by 36 inches, for 36 MP over 24 MP as far as I can tell. In digital you have to go really large before you see any benefit. In film sensor size was way more critical, and the difference in sensor size were way more pronounced.

Arguing APS-c against 35mil in digital is like arguing 645 against 6x7 in film. They are both way beyond consumer level in their out put capabilities, and far less than the best out there. For 4x6 or 8x10 prints, they are both extreme overkill.

In any case, I digress.

[jimr-pdx]

When I retire at the end of March, this may be my retirement present to myself. (I avoided he use gift as it may confuse our German speaking members.)

I retired early to care for my wife - but I do hit 59½ in August, so watch out IRA funds!

I love the history segment on their page. Many say Pentax is late to the party, but all the decorations at the party belong to Pentax - metering, reflex mirror, &c. Imagine today's legal system back then, no one else could make an SLR if Pentax had today's lawyers suing for infringement..

[savoche]

chocking? Did you mean choking or shocking?

I trust your Swedish is flawless

[Fogel70]

It is a valid comparison, because you can only use what your output device needs. If the most your output device, printer or screen can handle is 72 DPI, then it doesn't matter if your camera produces 100 dpi, or 300 dpi. People seem to want to keep these discussions isolated from discussions of printer and screen qualities, and that's where the nonsense starts.

What kind of camera you can make use of using different type of output is a incompletely different discussion VS how to compare IQ on two different cameras. How does viewing images at 100% help in knowing what camera fulfills you need?

As you say, most users do not need the best possible camera, but to find out what camera fulfill your needs, you need to know how to compare camera to know what to expect. Someone might skip to upgrade to a new camera as they see very little improvements comparing their old 2MP P&S with a 24MP APS-C camera viewing images at 100%, even though they might have seen improvement viewing images on their 4K TV.

I still struggle with finding it meaningful comparing images at 100%.

BTW, is there any major noise performance gains to upgrade from Nikon D2H to D810?

[monochrome]

Like comparing a -66 VW Beetle with a new Porsche and it turns out that they are equally fast, when both are driven at 50km/h that is.

When I drove a VW in the early 70's there was a story going around it was the fastest car on earth.

For the first ten feet.

Everything depends on the parameters of the argument, which is why we can't agree on anything.

[reh321]

What kind of camera you can make use of using different type of output is a incompletely different discussion VS how to compare IQ on two different cameras. How does viewing images at 100% help in knowing what camera fulfills you need?

As you say, most users do not need the best possible camera, but to find out what camera fulfill your needs, you need to know how to compare camera to know what to expect.

In some cases, I'm guessing that a little thing called pride is involved. I know of a guy who used to accompany the pictures he posted on-line with the tagline that his Nikon produced 36MP images, the largest available; he wasn't very happy with me when I suggested he put "FF" between "largest" and "available". Of course, when you're posting on-line, with restrictions similar to what we have here, 36MP vs 24MP is "a difference without a distinction", as a former boss of mine used to say.

[mee]

I trust your Swedish is flawless

Doesn't have to be.. we're on a web forum that uses English to communicate. Use a word once incorrectly and I can overlook it, more than once incorrectly demands mention as it was clearly not a typo. Of course, unless you enjoy watching people misspell words. As for me, I'd rather be corrected as to enhance my ability to communicate in the future. It is about caring for your fellow forum goers.