[gazonk]

That's probably a better name.

The problem with this "noise equivalence" is that it only works for theoretical sensors. DoF and FoV equivalences are of a different and absolute nature, but noise depends on the sensors at hand. For instance, the K-5 has better DR than the 5d mk iii at iso 200, so ISO 200 on the 5d mk iii is in no way "noise equivalent" to ISO 100 on the K-5!

[ElJamoquio]

The problem with this "noise equivalence" is that it only works for theoretical sensors. DoF and FoV equivalences are of a different and absolute nature, but noise depends on the sensors at hand. For instance, the K-5 has better DR than the 5d mk iii at iso 200, so ISO 200 on the 5d mk iii is in no way "noise equivalent" to ISO 100 on the K-5!

Huh? DR is only implicitly related to noise.

The ISO/noise equivalence has to do with the number of photons of light available per unit of the image... so really it should be a 'light gathering equivalence', as the aperture is more of a 'DOF equivalence' and the focal length is a 'FOV equivalence'.

Once we normalize for these factors it becomes easier to see which sensors are better and which are worse, say, at the 'electrical engineer level'. It gives us more of a basis to extrapolate designs from one camera to the next.

[gazonk]

Huh? DR is only implicitly related to noise.

The ISO/noise equivalence has to do with the number of photons of light available per unit of the image... so really it should be a 'light gathering equivalence', as the aperture is more of a 'DOF equivalence' and the focal length is a 'FOV equivalence'.

Yes, I'm aware if this. But if you look at dxo's charts, you'll see that the noise equivalence roughly holds for the s/n ratio, but that the DR of the 5d mk iii "flattens" at lower isos. I take that as evidence that it lacks the capability to fully utilize all the available photons.

[ogl]

Mystical K-5n - I've heard about this model 6 months ago....Is it myth, fantasy or it will be real?

[twitch]

Shouldn't you then name it noise equivalent?

Yes (although ISO also impact DR, coour bit depth and probably other things), and equiv aperture could be better called DOF equiv, and equiv focal lenght known as FOV equiv.

[gazonk]

Yes, I'm aware if this. But if you look at dxo's charts, you'll see that the noise equivalence roughly holds for the s/n ratio, but that the DR of the 5d mk iii "flattens" at lower isos. I take that as evidence that it lacks the capability to fully utilize all the available photons.

Here's what low DR means in terms of shadow noise. I took the dpreview test shots for 5d mk III and K-5, both in ISO 100 RAW versions, pushed exposure on both by 7 steps (equivalent to ISO 12800), then scaled both to 3000x2000 pixels and finally cropped on of the darkest areas of the dpreview test scene:

First, K-5:


Then, 5d Mk III:


As some users have pointed out, the K-5 is "the ISO-less camera" - the 5d mk III not so.

[bossa]

This will be my only reply.
I don't want to add on anymore to this FF/D800/D600 nonsense in a happy Pentax new release thread.

This is again in theory.
Practically, I've not seen 2 or more folks shoot with APS-C and FF cams on a photo outing with the FF one producing obviously better results than the APS-C users.
If there is such a case, it has only been for very controlled situations like model shoots.

24mm at f5.6 is at f5.6 too in terms of light gathering.
Candids, streets or a more environmental shot done in less ideal brightness, the DA15/4 at f4 would be better.
SR too on a Pentax camera.


Actually, I usually carry a 85/2 or 77ltd.
Perhaps you could crop a 50mm on FF to be a 70 on APS-C but the 70 on aps-c can also be cropped to be 130mm, and thats with a~2yr old 16mp camera. On a 24mp one, perhaps more.
I'm interested to see if this works out for you (as in purposely using a 50mm as a 70mm equivalent on APS-C as a travel solution).
I usually crop when necessary when reviewing the shot or have a preconceived notion that a shot needs to be cropped to 'something' later in PP, but can't judge what is 70mm APS-C on 50mm FF just by looking thru the viewfinder.


Encumbrance (not only weight) is different. The clunk of size and the weight means a bag that is 1 size bigger and a bit more strain from carrying around.
Easy for those who drives around, not that desirable if one does not.
Of course we humans are adaptable and can get used to the weight, but given a choice, I'd rather not go there.







I understand what you mean (I do shoot 35mm film and also MF 6x7; On 67, a 180mm lens focusing the distance like a ~90mm on 35mm film for portraits is really nice)
What I'm just saying is that shallow DOF is not everything in photography that is not still life, fine arts, portraits (including family ones).
For travel/landscapes/streets/documentation etc it matters much less and can be an advantage for comfort due to encumbrance and sometimes 'visibility' from a larger camera.

Like I've mentioned above, practically, I've not seen APS-C and FF shot on the same outing with the FF one producing obviously better results than the APS-C users. (Neither did I ever see Canikon branded cameras perform any better in terms of AF and such )

The lack of cheaper choices for Pentax to match the equivalent FOV on FF, I blame Pentax.
On Canikon, there are cheap 28, 35, 50, 85.
I'll have to say that the Pentax ones are smaller, typically better built and optically good though.
If the FA35/2 is used over the FA31ltd cited as an example, then the price difference becomes larger.






Ok. Enough from me.

I agree with all of that.

I just shot a 'still life' in crop mode on the D800E/Sigma 50 @ f/1.4 and K-5/DA*55 @ f/1.4. I backed off the '55 to try and frame it the same way (not very well unfortunately) and I found the Pentax gear produced a much sharper image but bokeh was different. I was also unable to see the subject (a coin) clearly in the D800 VF due to it being smaller in overall FF image so focusing was a little iffy.

The Nikkor 50/1.4G I had before the Sigma was rubbish compared to the Sigma and even worse compared to the DA55. The Nikon costs $750 here and the DA55 about 850. I got the DA for 600 used and the Sigma was 500 new. More than I paid for the Pentax version I had.

As you say, more often than not, for landscapes and tourist shots, you want more DOF so there probably isn't really an advantage with FF systems due to size etc.

All the FF lenses equivalents are a magnitude larger in size so the PAS-C definitely becomes the preferred option in some situations.

[bossa]

Here's what low DR means in terms of shadow noise. I took the dpreview test shots for 5d mk III and K-5, both in ISO 100 RAW versions, pushed exposure on both by 7 steps (equivalent to ISO 12800), then scaled both to 3000x2000 pixels and finally cropped on of the darkest areas of the dpreview test scene:

First, K-5:


Then, 5d Mk III:


As some users have pointed out, the K-5 is "the ISO-less camera" - the 5d mk III not so.

Thanks that was interesting.

From what I've seen in a few online tests the D800 also leaves the 5D3 for dead with shadow detail. It would have been cool to see that camera here in your presentation.

[zapp]

A quote from the German DSLR forum:
translated:
"... 1.) the last K5s arrived from the Far East
2.) As of May, the K5n will be available (with articulated screen)
3.) As of July/summer there will be a K3

...and the big bang:

a truly professional DSLR (K1?) is due in autumn WITHOUT APSC (!)"

because the original content is in Deutsch which is beyond my understanding, I guess the English is translated by Google translate, which may not exactly express it's original meaning.

Below is the original address, hope this is true although the possibility is low.

K30, K5n, K3 and maybe full-frame at Photokina...: Pentax SLR Talk Forum: Digital Photography Review

I am completely underimprssed by the K30. It is a slightly better, not in all terms, K5, so where is the news? Any interesting feature I would like to see on a yet to be announced FF Pentax. or a camera one level up? The AF was not vastly improved, styling stinks, and a little more ISO leverage is not really great news. This is a stand still with tiny improvements packed in a new box. This is the end of the K5 - articulated screen would be nice, but not a reason to buy a new camera. Typically the market segment of the K30 would be a technology carrier. Imagine a K3 full frame camera. Although great, the K30 will be the volume model and should showcase new technology rather than the pro edition.

[gazonk]

I am completely underimprssed by the K30. It is a slightly better, not in all terms, K5, so where is the news?

The news is that the K30 is the K-r replacement, not the K-5 replacement. The fact that several people mistakes it for a K-5 replacement says everything about how good it is

[gazonk]

Thanks that was interesting.

From what I've seen in a few online tests the D800 also leaves the 5D3 for dead with shadow detail. It would have been cool to see that camera here in your presentation.

I tried the same for D800 (pushed 7 steps in Aperture, resized to 3000x2000 and cropped) and got a very good result - but slightly confusing. It's confusing because while it looks very good indeed (the amount of detail in the threads and the Baileys bottle letters before the downsizing is breathtaking), some detail seems lost in the black here - what's that? RAW NR at ISO 100? Or maybe just that Aperture's defaults are different for each camera.

[Emacs]

You seem to suggest that ISO100 with FF sensor has a different sensitivity as ISO100 with APS-C,

It's all about effective SNR. We have basically the same sensor as in K-5. But with FF sensor you have 2.25 times more pixels for the same subject. For each sensel we have a probability: P(|px-mx|<ε), where the mx is the "real" value, and px is the value that is taking with the sensel, the ε is the acceptable accuracy. The P is different for different ISOs, of course. So, you have 2.25 times more pixels to show exactly the same information. And thus you have 2.25 times more chances to get the proper value.

[bossa]

I tried the same for D800 (pushed 7 steps in Aperture, resized to 3000x2000 and cropped) and got a very good result - but slightly confusing. It's confusing because while it looks very good indeed (the amount of detail in the threads and the Baileys bottle letters before the downsizing is breathtaking), some detail seems lost in the black here - what's that? RAW NR at ISO 100? Or maybe just that Aperture's defaults are different for each camera.

Yeah it's still not as smooth as the K-5 output and it appears as though there's some processing going on where a filter is blocking certain information that doesn't get blocked until an extreme situation like this comes along..

[Emacs]

Yeah it's still not as smooth as the K-5 output and it appears as though there's some processing going on where a filter is blocking certain information that doesn't get blocked until an extreme situation like this comes along..

Really?
I found when examine D800 RAW I can pull up to 5 stops, compared to K-5's 4 stops with acceptable noise and colors.

[Aristophanes]

I think almost everyone here understands the aperture 'equivalency'.

An aperture is aperture. f-stop is f-stop. There is no "equivalence".

What you are referring to are the DOF, CoC, FOV, and hyperfocal distances in relation to the sensor size and how the optical configuration draws the picture of a properly exposed photograph, much based on subjective criteria such as OOF and bokeh.

ISO "equivalence" is also problematic in that higher ISO's reduce DR as well as adding noise. And noise comes in different varieties.

It's like saying we all understand how fish taste, disregarding that salmon doesn't taste like tuna.

Obviously, many people here don't understand.