Does anyone have (or know where to see) samples of IQ180 at F/22?

[jcoffin]

Even though I agree that the wall itself looks a bit OOF, I think the vertical board slightly left of center is pretty telling. If you look in the f/11 section, it has quite a well-defined surface texture. There's only a little of it showing in the f/16 section, so it's hard to say for sure, but I'd say it looks decidedly softer. In the f/22 section, we get only a hint of the surface being rough at all -- where the f/11 section shows clear saw marks, the f/22 section looks like somebody spent some time sanding it, so even though there's a bit of "waviness", there's no real texture left.

 

[kdphotography]

Wwwhat? No brick wall? :ROTFL:

I'm with Wayne and Doug here. Sometimes you let yourself get a bit antsy or worked up from pixel peeping too much, and perceived softness particularly with higher res backs. Yes, you still need to use good photography care in capture, but you'd be amazed how well those images print up....

ken

 

[Steve Hendrix]

Yeah I'm not an MFDB owner so I can't really play around with it to know. I want to know the F/22 answer because it will be what I plan to shoot with the most. On my Canon F/22 is definitely softer but acceptable for landscape photography.

Thanks for the sample image. Much appreciated! I know there's no sharpening done here, but I can't help to think that it looks very soft. Was it handheld or was it on a sturdy Tripod/head with MLU? Was there a focusing micro offset (i.e. the focus plane isn't perfectly aligned) that you're aware of? Did you use the Phase body (SLR) or Cambo? Or is the softness normal and will become 'properly' sharp once sharpened?

Sorry with all the questions a bit excited to see I wonder if you can try to get the F/11 as sharp as possible and compare against F/22. Even shooting something inside is ok as long as it's on tripod with MLU. I just did this experiment with my Canon an hour ago

I think if you've been satisfied with the Canon images at f22, you'll probably be ok with IQ160 or IQ180 images at f22. Shooting at f22 isn't necessarily a deal killer, it typically has the effect of the loss of critical crunch, rather than being totally soft, if it is a sharp lens whose optical resolution is not overly weighted in the wider apertures.

In my experience, there are definitely optical designs that favor the resolution that is captured at the wider end of the aperture range (and vice versa). So, that can play a role also.

All that said, the question in my mind is what you shoot that most of your images require f22? Whatever it is, my assumption is that you are aiming for maximizing your depth of field? If so, perhaps there are workarounds that can produce a sharper image (if that is your objective) at a more optimal aperture. This is frequently a consideration when we discuss camera/lens choices with clients who are trying to maximize depth of field.

The goal, in terms of the equipment, is often to try and move the target aperture to say, f11, and configure everything so that is possible and still attain an equivalent depth of field. Of course, there are limitations to this, but limitations don't always stop photographers from achieving their objective, do they? One advantage of a high resolution sensor is the ability to shoot at different distances and still end up with a high resolution image.

To me, the critical sharpness (crunchiness) is lost (or begun) with almost all professional camera/lens combinations at f22. So, the effort is most often pointed towards - what can we do to get away with f16, or f11, etc.


*Edit - Ken makes an important point here also, that we can get hung up looking at the image on the computer monitor/display and often, in terms of noise and sharpness, the print is quite different (usually in your favor).


Steve Hendrix

 

[Jack]

the difference between f/16 and f/22 was pretty pronounced.

It is. And when we did our IQ 180 test, I took one shot at f16 with the 55LS, and the negative aspects from diffraction are clearly visible at 100% view as compared to f11; f22 would be significantly worse. However that said, IMHO a print made from that f16 capture would pass most close scrutiny, and the slight amount of added DoF is probably more beneficial than the slight loss of fine detail. However, f22 would probably go the other way...

 

[Jim2]

It is. And when we did our IQ 180 test, I took one shot at f16 with the 55LS, and the negative aspects from diffraction are clearly visible at 100% view as compared to f11; f22 would be significantly worse. However that said, IMHO a print made from that f16 capture would pass most close scrutiny, and the slight amount of added DoF is probably more beneficial than the slight loss of fine detail. However, f22 would probably go the other way...

This is exactly why I'm beginning to think that it's one of the reasons many landscape photographers still use 8x10 - the larger surface of the 'sensor' is more forgiving to bigger diffraction, something that DMF cannot physically deal with, in fact it only gets worse as the Mega pixels are getting higher.

 

[Steve Hendrix]

This is exactly why I'm beginning to think that it's one of the reasons many landscape photographers still use 8x10 - the larger surface of the 'sensor' is more forgiving to bigger diffraction, something that DMF cannot physically deal with, in fact it only gets worse as the Mega pixels are getting higher.

That said, I'll bet there's a lot of wonderful 8x10 film images shot with 8x10 lenses that, printed at large size, don't look any sharper than an 80MP file shot at f22 with any optimal modern lens printed at the same size.


Steve Hendrix

 

[svema]

Hello Jim. Your question very difficult. Probably you are familiar with it: http:// www.luminous-landscape.com/tutorials/resolution.shtml
But it is not forgotten that the photo is light and a shade. And so in a shade just all the worst also is shown.
Knowing I all it was afraid to change the Р45 + on Р65 + and has bought it only because it is the maximum size of a sensor control. My experience on Р65 + it F11. I remove a landscape and have come to Sinar р3. But it is my choice.

 

[coulombic]

Two points -- first, I have an Aptus-II 12 and wouldn't mind shooting a barrage of apertures and lenses to give you an accurate sense of how the decreased pixel element size affects diffraction.

Second, why has no one mentioned hyperfocal distance? It seems the case that he's intending to shoot landscape images -- likely with a very wide-angle lens. I presume a 28mm, or something along those lines. If I focus my 28mm hyperfocally and shoot at f/16-f/18, the whole scene is reasonably sharp, maintaining excellent DoF. Certainly, if he's shooting a longer focal length, and attempting to maintain corner-to-corner DoF, there's going to be an issue, but with a shorter focal length lens, assuming he knows where to focus, this seems possible without resulting to f/22.

From my experience, as well, f/22 on a P65+, which was my previous back, wasn't horrible, but it was notably less-sharp than f/16-18. I've yet to attempt it with my Leaf, but I'll try to post some samples tomorrow. Any requests on the focal length?

Also, with the higher resolution, as diffraction becomes more of an issue, the effective resolution starts to ebb away. How much data you're actually recording at f/22 might not be any more than a P65+ at f/16. I've not done the math, but it seems like a worthy consideration to undertake. If you plan to consistently shoot at smaller apertures, it might benefit you to stick with a sensor that won't show the limitations of the optics as readily.

 

[MGrayson]

Disclaimer: The following is a purely theoretical statement and is no use if it doesn't hold up in practice.

[:lecture:]Diffraction destroys some information (the convolution kernel is, unfortunately, zero in places), but much less than sampling does. A well enough sampled image that has been blurred by diffraction could in theory be well reconstructed through deconvolution sharpening. This may totally fail in practice and give ugly images.. I don't know. I have yet to take a photo where diffraction was the dominant source of lack of sharpness:ROTFL: But it is a fairly regular and uniform source of image blur, and so should be tractable. [/:lecture:]

Matt

 

[Jim2]

Two points -- first, I have an Aptus-II 12 and wouldn't mind shooting a barrage of apertures and lenses to give you an accurate sense of how the decreased pixel element size affects diffraction.

Thanks for the offer to take some examples with Aptus II 12. Mal's examples above seem like a good format for comparison - F11 vs F16 vs F22.

On 35mm (Canon) Focal Lengths that I use vary from 35mm, 70, 105, or longer.

Also, with the higher resolution, as diffraction becomes more of an issue, the effective resolution starts to ebb away. How much data you're actually recording at f/22 might not be any more than a P65+ at f/16. I've not done the math, but it seems like a worthy consideration to undertake. If you plan to consistently shoot at smaller apertures, it might benefit you to stick with a sensor that won't show the limitations of the optics as readily.

My question would be: A scene with the same framing/composition and the same F stop (in this case F/22), would it be better taken by an 80 MP back or a 60MP back, considering that the diffraction is affecting the 80MP back more adversely due to its smaller pixel size.

 

[malmac]

Mal, I'm not sure if this is something you want to do (for fun!)

- Capture tethered (to save time)
- Set to AF at the barn area that we are looking at 100% crop
- Take a shot - check shot result
- Then switch lens to manual focus and adjust focus forward, take a shot, adjust focus back, take another shot

See whether the AF is really doing its job properly?

Jim2

Though your suggestion was reasonable. Loaded software onto laptop, then realised I don't have a firewire port on laptop. Not able to shoot tethered at present. Damn.

Decided I would try and reshoot the shed with a number of focus settings to see if I could get a sharper result more through luck and trial and error. So here is my very unscientific test results.

My conclusion -
Inconclusive result. No direct coparison with another camera so hard to say what is possible.
The portion of the image being reviewed is realtively small when one zooms in. I guess if you zoom in far enough any digital image will get soft. I expected that the image would start to sharpen or blur more as I changed the manual focus, perhaps I may not have gone far enough to get to the sweetest focus point.

Any comments welcome.


Thanks Mal

 

[Jim2]

Mal,

Thanks for doing this. I cannot see an 'obvious' difference. Nothing so striking. I was just thinking, as you said that maybe due to distance from the camera that's the best we can get? Perhaps try moving closer say to within 10m of the barn and see if we can get more sharpness. I noticed the artefacting of the jpg too. Maybe it's the lens that's not sharp..... The classic shots people make seem to be of a brick wall - see http://www.luminous-landscape.com/reviews/cameras/2010_mini_medium_format_shoot_out.shtml

 

[SCHWARZZEIT]

This is exactly why I'm beginning to think that it's one of the reasons many landscape photographers still use 8x10 - the larger surface of the 'sensor' is more forgiving to bigger diffraction, something that DMF cannot physically deal with, in fact it only gets worse as the Mega pixels are getting higher.

Not necessarily, if you stop down to f/22 on 645 to get enough DOF you would have to use f/90 on an 8x10" to get similar DOF. The level of detail in one-shot photography can be ultimately limited by your DOF requirements due to diffraction.

-Dominique

 

[Ebe]

Test I did with p40+, after upgrading from p30
http://forum.getdpi.com/forum/showthread.php?t=18137

 

[Jeremy]

Not necessarily, if you stop down to f/22 on 645 to get enough DOF you would have to use f/90 on an 8x10" to get similar DOF. The level of detail in one-shot photography can be ultimately limited by your DOF requirements due to diffraction.

-Dominique

Can you throw some math at me on how you got these numbers?

If we're talking hyperfocal distance using "normal" lenses (80mm on 645 and 300mm on 8x10) with a subject distance of around 45m (based off of the picture in this thread) I imagine the required f-stop on 8x10 to match the "acceptable close focus" figure on f/22 with full-frame 645 is probably between 32 & 45--nowhere near f/90. (This is assuming we also want infinity in focus.) This is based off of experience shooting both medium format and 8x10. If I get the time I'll pull up the hyperfocal equations as I'm now intrigued.

 

[SCHWARZZEIT]

Can you throw some math at me on how you got these numbers?

If we're talking hyperfocal distance using "normal" lenses (80mm on 645 and 300mm on 8x10) with a subject distance of around 45m (based off of the picture in this thread) I imagine the required f-stop on 8x10 to match the "acceptable close focus" figure on f/22 with full-frame 645 is probably between 32 & 45--nowhere near f/90. (This is assuming we also want infinity in focus.) This is based off of experience shooting both medium format and 8x10. If I get the time I'll pull up the hyperfocal equations as I'm now intrigued.

With a subject distance of 45m you wouldn't use f/22 on 645 for DOF reasons with a normal lens. With a view camera you would tilt to increase DOF so there is no reason to go to f/90 unless you're in confined space.

The math I used was simply that the 8x10" image format is about 4.5 times wider than 645 in each dimension. Thus the equivalent aperture for similar DOF is 4.5 times smaller. This calculation also covers the linear scaling of the COC.

-Dominique

 

[Jack]

If you do the math, the smaller formats will gain DoF faster than they loose to diffraction at aperture X -- meaning they are more efficient. However, they also need to be enlarged more, so in reality it's becomes a sort of trading of deficits -- and it's been long accepted that MF through 4x5 is the sweet-spot happy medium.

8x10 is preferred for movements over 4x5 because you can more easily see the effects of focus, tilts and swings on the GG than with 4x5. MF is preferred at the other end over 4x5, because they need less or no movements to carry DoF acceptably, where 4x5 does -- and you're focusing though a magnified finder, or with some focus confirmation electronics on board. (But if you do need tilts on MF, it is a real PITA to see them clearly on a ground-glass. Hence the huge benefit the IQ series Phase backs will offer to tech camera users with it's usable 100% frame view.)

 

[Wayne Fox]

Mal,

I think your tests would be more effective if you would show the same section of the wall for each f/stop, not the pie chart thing you are doing. No 1 to 1 correlation of detail between the sections.

 

[malmac]

Wayne


Thanks for the feedback.

I sent one of the images to my camera lab and had them print out some 12x8 inch sections from the centre and the side of image from the f8 setting.
The results are pleasing but still have some testing to do with the 65+ on the Cambo Wide DS with a 35mm Schneider lens to see how that works. Still dont have a laptop or a viewfinder so focus will come down to measuring the distance to the subjet.

Thanks again for your help. A lot to learn but that is all good stuff.



Mal

 

[goesbang]

All this math and physics and science is hurting my brain.
I think I'll take my ALPA STC , 35mm SKXL, 80 MP Aptus and go out in the real world and take some beautiful pictures at whatever aperture I darn well feel like.....
Cheers,