What's all this about corner softness

Lee Jay · Sep 2, 2007

Rick Knepper said:
ljfinger said:

Then again, I think sharpness is overrated and

Click to expand...

How does one overrate sharpness?

A tack-sharp shot of a boring subject is a much worse shot than a butter-soft shot of an interesting subject. In other words, lack of sharpness doesn't ruin a shot (I'm talking about lens sharpness here, not massive miss-focus or motion blur) the same way a boring subject does.

This shot is soft because of noise caused by low-light and low color temperature (not much green or blue wavelengths in the light source) and by lack of contrast in the lighting. Yet, I like it anyway.

This shot is tack-sharp, but boring.

--
Lee Jay
(see profile for equipment)

GaborSch · Sep 2, 2007

who have not heard of

1. landscape photography,

2. architectural photography,

3. reproducting of painted arts, drawings, etc,

4. stitched panoramas (the sharpness from the first to the last pixel is extremely important here).

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

jouni · Sep 2, 2007

no text

Amin Sabet · Sep 2, 2007

Let me know if you object to my hosting/posting crops of your image, and I will take them right down.

ljfinger said:
I just don't see it. The grass, the fence pickets and the roof
shingles look virtually the same on the left and right edges as they
do in the center.

--
http://aminphoto.blogspot.com

joe mama · Sep 2, 2007

Rick Knepper said:
Rick Knepper said:

However, I feel that sharpness is over-rated because of the
sacrifices required to get it. How many people ask at which f-ratio
their lens is at it's sharpest? These people care more about
sharpness than DOF, which is odd to me.

Click to expand...

Rick Knepper said:
Don't know about most people. I ask for a lens to render sharpness in
an image where it is supposed to be sharp - in the in-focus parts.
Playing with the depth of field is indeed whole 'nudder thang.

I concur.

Rick Knepper said:
So, in such circumstances as described above, and maybe we'll call
them narrow circumstances in an effort to be fair and balanced, if
your lens is still producing images that are sharp in the center and
grows noticeably softer towards the edges at f8, I feel the lens is
not right for that type of photography. This narrow circumstance is
very similar to the flat wall tests some folks do.

Having to use a DoF function to produce an effect that the lens is
incapable of physically is indeed frustrating.

Again, I concur.

Rick Knepper said:
Rick Knepper said:

The 50 / 1.2L is relatively sharp wide open, but the 50 / 1.4 is not.
Regardless, I used the 50 / 1.4 wide open all the time, not to get a
sharp image, but to get a shallow DOF. Would I have preferred a
sharp image? Of course! But not at the expense of deeper DOF.

Fortunately, the 50 / 1.2L is considerable sharper wide open and
allows the best of both worlds. That's a big reason I got it. But
it's a lot sharper at f / 5.6 than f / 1.2 and yet that plays no
role in my photography.

Click to expand...

Rick Knepper said:
Not following you here. Maybe that's because I don't shoot wide open
often, and all of this seems written as rather "wide open- centric".

Just saying that sharpness plays second fiddle to DOF.

Rick Knepper said:
Further anecdotal proof. Canon specifically mentioned edge
performance in their press release for the 16-35 II so obviously, a
significant number of folks were concerned enough about the issue for
Canon to react. Probably the same for the 14L.

Certainly agree with that, as well. Check out Cedric's gallery:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=24630383

I think he did a fine job of presenting images where edge sharpness is important.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

Lee Jay · Sep 2, 2007

Amin Sabet said:
Let me know if you object to my hosting/posting crops of your image,
and I will take them right down.

Nah...they're just test shots - do what you want to them.

I do prefer to try to look at sections of the image with similar lighting contrast and content, however. Left-to-right on the crops.

--
Lee Jay
(see profile for equipment)

Rick Knepper · Sep 2, 2007

Of course, you'll have to do some work yourself. All shot RAW, converted with DPP, all with WB temp of 6500k, no sharpening or other processing whatsoever. You can run your own PP on them whatever that may entail (so that my "bad" PP choices don't get in your way).

Lee Jay · Sep 2, 2007

Rick Knepper said:
I did these for a 1.6x vs. FF equivalency excercise but they may have
some value here. If they don't, then please disregard.

They don't have the same FOV, thus they are not equivalent at all.

--
Lee Jay
(see profile for equipment)

Amin Sabet · Sep 2, 2007

ljfinger said:
Amin Sabet said:

Let me know if you object to my hosting/posting crops of your image,
and I will take them right down.

Click to expand...

Nah...they're just test shots - do what you want to them.

I do prefer to try to look at sections of the image with similar
lighting contrast and content, however. Left-to-right on the crops.

I found it hard to match lighting and content since areas of similar content had different lighting and the two extreme corners in the foreground lack detail. Nevertheless, I think the crops you chose tell the same story as the crops I chose. There is a clear drop in resolution/acutance with distance from the center. Other factors to consider with any such test are that 1) edge performance is not as bad as corner performance, 2) extreme corners are generally much worse than near corners, and 3) one would prefer not to have to use f16 every time, as Rick pointed out. That said, the edge/corner performance shown in your photograph is absolutely good enough to suit my standards in any application. Rick just has higher standards than I (and apparently you) do.

Lee Jay said:
--
Lee Jay
(see profile for equipment)

--
http://aminphoto.blogspot.com

Lee Jay · Sep 2, 2007

Amin Sabet said:
ljfinger said:

Amin Sabet said:

Let me know if you object to my hosting/posting crops of your image,
and I will take them right down.

Click to expand...

Nah...they're just test shots - do what you want to them.

I do prefer to try to look at sections of the image with similar
lighting contrast and content, however. Left-to-right on the crops.

Click to expand...

I found it hard to match lighting and content since areas of similar
content had different lighting and the two extreme corners in the
foreground lack detail.

Yup.

Amin Sabet said:
Nevertheless, I think the crops you chose
tell the same story as the crops I chose. There is a clear drop in
resolution/acutance with distance from the center.

I see it more as "barely" than "clear" but it's a judgement call.

Amin Sabet said:
Other factors to
consider with any such test are that 1) edge performance is not as
bad as corner performance, 2) extreme corners are generally much
worse than near corners, and 3) one would prefer not to have to use
f16 every time, as Rick pointed out. That said, the edge/corner
performance shown in your photograph is absolutely good enough to
suit my standards in any application. Rick just has higher standards
than I (and apparently you) do.

I guess he does. In any print or on-screen view of normal size (I have a 1920x1200 24" screen) I can't see it at all. Here's another sample.

Lower left corner (barely in the DOF):

--
Lee Jay
(see profile for equipment)

GaborSch · Sep 2, 2007

The FOV has nothing to do with this subjec. The comparison is valid.

The edge of the 300D shot is is sharper than that of the 5D even at f/8. Not only that, but the same section of the scenery, for example the building at the left edge close to the top is sharper on the 300D, even at f/8. I guess this is because the very same detail is farther from the lens center on the 5D.

However, had this test been designed for the purpose of this thread, you should have fixed the focusing on the lens and shoot with the very same adjustment on both cameras.

I find it suspicious, that at f4 the 300D image is sharper even at the center than the 5D. The center of the image is on water or at the edge of the water and thus the focusing may not be accurate. The best would be to focus manually.

One more issue: you posted "no sharpening in DPP". I would like to see confirmed, that the sharpening has been set to 0 in both cases , otherwise the in-camera setting applies.

-
Gabor

http://www.panopeeper.com/panorama/pano.htm

joe mama · Sep 2, 2007

Rick Knepper said:
That one had some serious softness at the corners.

I have these, from a test I did a while back to show the corners of the 16-35 / 2.8L on the 5D:

http://www.pbase.com/joemama/163528l_on_a_5d

Do you mean this pic from the above link?

http://www.pbase.com/joemama/image/68520036/original

I forget the f-ratio I shot it at, and it's on a different harddrive (when I convert from RAW, I save as a tiff, and when I save a tiff as a jpg using IrfanView, it yanks the EXIF, for whatever reason).

Anyway, the extreme bottom corners are whacked, for sure, but, like I say, I don't know what the f-ratio of that shot was. Best to look at the other test pics on the page for a better understanding of the lens performance.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

Amin Sabet · Sep 2, 2007

ljfinger said:
I see it more as "barely" than "clear" but it's a judgement call.

When I called it a "clear" difference, I meant clearly discernible. That doesn't mean I think it's a marked difference. Both photos you posted have good enough edge-to-edge performance for me. In fact, my Tamron 17-35 has clearly worse corner performance at f16 than your 17-40, yet it too is good enough for me, even after correcting the barrel distortion. However, in a 12x18 print at a reasonable viewing distance, I would be able to see the drop in performance towards the edges on the first shot you posted. I didn't print it, but I've printed enough to know how things translate from the screen to print for me. For the second example, I'd have to see a higher-res version to be sure. Three of your crops are pretty central, and the bit of the two edge crops you gave are pretty small.

--
http://aminphoto.blogspot.com

Lee Jay · Sep 2, 2007

GaborSch said:
The FOV has nothing to do with this subjec. The comparison is valid.

The edge of the 300D shot is is sharper than that of the 5D even at
f/8.

Did you compare at constant final image size?

GaborSch said:
Not only that, but the same section of the scenery, for example
the building at the left edge close to the top is sharper on the
300D, even at f/8. I guess this is because the very same detail is
farther from the lens center on the 5D.

No, because at the same focal length the 300D, even at only 6.3MP, puts more pixels on the subject than does the 5D, and you're looking at 100%.

--
Lee Jay
(see profile for equipment)

Lee Jay · Sep 2, 2007

Amin Sabet said:
ljfinger said:

I see it more as "barely" than "clear" but it's a judgement call.

Click to expand...

When I called it a "clear" difference, I meant clearly discernible.
That doesn't mean I think it's a marked difference. Both photos you
posted have good enough edge-to-edge performance for me. In fact, my
Tamron 17-35 has clearly worse corner performance at f16 than your
17-40, yet it too is good enough for me, even after correcting the
barrel distortion. However, in a 12x18 print at a reasonable viewing
distance, I would be able to see the drop in performance towards the
edges on the first shot you posted. I didn't print it, but I've
printed enough to know how things translate from the screen to print
for me. For the second example, I'd have to see a higher-res version
to be sure. Three of your crops are pretty central, and the bit of
the two edge crops you gave are pretty small.

What do you think about this for a test.

Take what amounts to an infinitely sharp shot and compare that to the actual output of the lens in question on a single shot. The infinitely sharp shot would be a minimum 3x3 panorama of the same angle of view (obviously with a longer lens) stitched and then reduced to the same final size as the single shot. You'd have to match apertures close enough to ensure that DOF wasn't an issue.

--
Lee Jay
(see profile for equipment)

Steen Bay · Sep 2, 2007

Is corner softness better on 35mm film than on a FF-sensor, because of the differences in "technology"?

Lee Jay · Sep 2, 2007

Steen Bay said:
Is corner softness better on 35mm film than on a FF-sensor, because
of the differences in "technology"?

If the corners are better on film, the center is too, because of the smaller grain size of certain films (especially B&W films). What film is supposed to do better in the corners is have less light falloff, but I haven't seen this in practice. I keep meaning to do a controlled test but the film cans are still sitting on here my desk.

--
Lee Jay
(see profile for equipment)

Amin Sabet · Sep 2, 2007

Seems like overkill to me Lee Jay. Mark over at 16-9 Lens Tests has done a nice job comparing lenses without going through that sort of thing.

An example of one of his corner assessments with the 16-35 II: http://www.16-9.net/lens_tests/canon1635ii_a/c1635ii_a5.html

Interestingly, he went through multiple copies of the 17-40L before finding one that still made him happy.

Here's his comparison with the good copy of the 17-40L: http://www.16-9.net/lens_tests/canon1740_nikon1735/c1740vn1735a.html

Interestingly, this best 17-40L was a near match for the Zeiss in the extreme corners at f16. Based on your first sample, I'd guess that your copy isn't as good as his. This makes sense given that he went through a bunch of copies to find it.

Personally I don't care very much about lens performance in the extreme corners, but for those whom it matters, I like the way he compares them.

ljfinger said:
What do you think about this for a test.

Take what amounts to an infinitely sharp shot and compare that to the
actual output of the lens in question on a single shot. The
infinitely sharp shot would be a minimum 3x3 panorama of the same
angle of view (obviously with a longer lens) stitched and then
reduced to the same final size as the single shot. You'd have to
match apertures close enough to ensure that DOF wasn't an issue.

--
Lee Jay
(see profile for equipment)

--
http://aminphoto.blogspot.com

GaborSch · Sep 2, 2007

ljfinger said:
GaborSch said:

The FOV has nothing to do with this subjec. The comparison is valid.

The edge of the 300D shot is is sharper than that of the 5D even at
f/8.

Click to expand...

Did you compare at constant final image size?

Lee Jay said:

Not only that, but the same section of the scenery, for example
the building at the left edge close to the top is sharper on the
300D, even at f/8. I guess this is because the very same detail is
farther from the lens center on the 5D.

Click to expand...

No, because at the same focal length the 300D, even at only 6.3MP,
puts more pixels on the subject than does the 5D, and you're looking
at 100%.

--
Lee Jay
(see profile for equipment)

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

Rick Knepper · Sep 2, 2007

That's at f2.8, certainly not the type of image for peeping the edges.

But, yes that was the pic.

joe mama said:
Do you mean this pic from the above link?

http://www.pbase.com/joemama/image/68520036/original

I forget the f-ratio I shot it at, and it's on a different harddrive
(when I convert from RAW, I save as a tiff, and when I save a tiff as
a jpg using IrfanView, it yanks the EXIF, for whatever reason).

Anyway, the extreme bottom corners are whacked, for sure, but, like
I say, I don't know what the f-ratio of that shot was. Best to look
at the other test pics on the page for a better understanding of the
lens performance.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos.
If you wish to use any of my photos for any purpose other than
editing in these forums, please ask.

What's all this about corner softness

Forum Pro

Veteran Member

Senior Member

Veteran Member

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Veteran Member

Forum Pro

Veteran Member

Forum Pro

Veteran Member

Forum Pro

Forum Pro

Veteran Member

Forum Pro

Veteran Member

Veteran Member

Forum Pro

Keyboard shortcuts