[carolina_sky]

When would you use pixel shift with a high iso or is this just a theoretical comparison?

In lower light situations primarily. Can only do a 30s exposure with PS (so 30x4). If you need to bring out some deep dark shadows, you probably need to go higher than 100 ISO.

[Simen1]

When would you use pixel shift with a high iso or is this just a theoretical comparison?

Just theoretical. Maybe it could be an alternative to shutter speed bracketing?

If you need to bring out some deep dark shadows, you probably need to go higher than 100 ISO.

.. or just bring up the shadows in post so you preserve the highlights in the process.

[carolina_sky]

.. or just bring up the shadows in post so you preserve the highlights in the process.

I know there has been some discussion of K-1 as being close to ISO invariant. I guess that means that what you suggested should work. Is this the case? Is ISO 800 IDENTICAL to ISO 100 with +3 stops of shadow boosting? I went back looking and I haven't done any actual side by side compare of different ISOs of dark scene that I can pull out here. Something to try...

Another scenario which I ran in to recently was some night time shooting on a bridge with vehicular traffic (in Lucerne). I could not get a sustained 2min gap between vehicles crossing the bridge and causing vibration. So I chose to shoot 8s at 400 ISO instead of 30s at 100 (in PS which meant 32s shoot time at 400 ISO and 2 min at 100). And voila no vibration. High wind might be a similar kind of scenario. So I think there are valid reasons for shooting higher ISO with PS (I wouldn't personally bother with anything higher than 1600).

[Simen1]

I know there has been some discussion of K-1 as being close to ISO invariant. I guess that means that what you suggested should work. Is this the case? Is ISO 800 IDENTICAL to ISO 100 with +3 stops of shadow boosting?

Not identical, for better and worse. On the positive side pulling ISO 100 three stops up (with curves) preserves highlights much better. On the negative side, you will get a fraction of a stop more noise.

[carolina_sky]

Not identical, for better and worse. On the positive side pulling ISO 100 three stops up (with curves) preserves highlights much better. On the negative side, you will get a fraction of a stop more noise.

Clearly one has to balance highlight and shadow. I wouldn't jump up the ISO unless I had some way or plan (filters, etc) to control highlights. My speculation was rather theoretical since I use ISO 100 in about 98% of my static scene shooting.

[richmondthefish]

I'm still impressed by how much the pixel shift feature can do for still scenes. Try DPReview's side by side comparison with the mighty 25x more expensive Phase One XF 100 Mp. Here are my analysis:

I've found Pixel Shift to be really good in product type photography for pixel peeing but I still haven't been too satisfied with using it for anything else. I've gotten some decent results in landscape but the final image wasn't really better then what I was using without it. It seems to be useless with people even when you try to get the person to be as still as possible with a fast shutter.

[Nicolas06]

I also find it odd, when folks talk about Canon specialty lenses, to one of my Canon user buddies, when I'm able to drop my 40XS and 21 ltd. into a small case and stick it in my pocket constitutes specially gear.

Canon has similar specs 24 and 40mm. The 24mm lens as much better borders than our 21 on similar specked APSC sensor and the 40mm support officially FF format.

So while I agree there no DA15 or DA70, you managed to choose the 2 pancakes Canon can perfectly compete with

And yes, I have DA21 and I really notice the border performance with its field curvature characteristics. That the only remaininng lens I want to upgrade actually, but there nothing in Pentax land except FA20 that now very expensive and hard to find. If you ask I altready tried FA24 f/2 (huge and terrible at anything bellow f/5.6).

Certainly then if DA21 and DA40 count as specialty gear, the 24 and 40mm from Canon should count too!

[Nicolas06]

As per previous discussion, the consensus seems to be that you really couldn't tell the difference between a difference of 5 points or less on the DxO tests. You probably should not be reading too much into the numbers of those at the top, they don't mean much.

100% agree, actually DxO say it themselve.

[Nicolas06]

Sun light surely produce light ray from all septrum, but not at all with same intensity, and that also depend of the what time of they day you speak of:



Then surely halogen isn't perfect, but is not so bad match but lack near 400mm wavelength...





But the camera sensor are not that sensitive anyway theses wavelength neither:


And that's our eyes sensitivity:



Even if Sun emit significantly more blue near 400nm than halogen lamps, the low sensitivity of colors sensors and our eyes to this kind of light make it look it is not that bad as one would make us think.

DxO make one of the finest raw demoising algorithms, they have many scientists and skilled technician working for them and doing analysis. It is funny to see how anybody here, feel the right to explain how wrong they are without having even a small fraction of their knwoledge and practical experience in the field of camera and lens testing.

For all we know they are perfectly aware and corect they measurement for it... Or simply this blue light spectrum is simply not applicable at the way we think it is.

To start with ou eye see the 400mm barely and even 450m with 3-4Ev less intensity than green.

[biz-engineer]

It is funny to see how anybody here, feel the right to explain how wrong they are without having even a small fraction of their knwoledge and practical experience in the field of camera and lens testing.

A lot of people were born before you, and you learned most of what you know from others, if you had to learn it by trial and error, you'd know a lot less, and you'd have died of the injuries of the failed experiments, fortunately your parents and society was there to prevent as many accidents as possible, or you are a super genius who knew it all at the time you were born? :-)

[D1N0]

Beats me why this thread is alive again...

[Nicolas06]

A lot of people were born before you, and you learned most of what you know from others, if you had to learn it by trial and error, you'd know a lot less, and you'd have died of the injuries of the failed experiments, fortunately your parents and society was there to prevent as many accidents as possible, or you are a super genius who knew it all at the time you were born? :-)

While it sound quite profound, it is trivia that apply to anybody.

if you have an actual message there, maybe you should be more explicit.

[D1N0]

Ok it is has become the Nicolas wants to outsmart and condescend everybody thread now.

[Nicolas06]

Ok it is has become the Nicolas wants to outsmart and condescend everybody thread now.

So then if somebody say theses guy are dumbs they don't know their business look they even use halogen lights that mean they completely messed up... That's not consdencending to anybody. Surely not to DxO.

But then if somebody else says hey look they studied at school for that longer than most of us (on the subject) and that's also they daily job so maybe they know a thing of 2 about it and maybe that blue light thing isn't as terrible as some make it appear, then it is condescending to everybody thread?

[BigMackCam]

Folks - please let's get back to the subject without bickering... otherwise, the thread serves little purpose and I'll have to close it.

Thank you!