Diffraction defeated by pixel count - 17mm, f22

Discussion of lenses, brand or independent, uses and merits
Forum rules
No more than three images or three external links allowed in any post or reply. Please trim quotations and do not include images in quotes unless essential.
David Kilpatrick
Site Admin
Posts: 5985
Joined: Sat May 19, 2007 1:14 pm
Location: Kelso, Scotland
Contact:

Diffraction defeated by pixel count - 17mm, f22

Unread post by David Kilpatrick »

Diffraction does lose sharpness when you stop down even below f8 or f11 on small sensors. But the Alpha 900 sensor pixel pitch is not going to see diffraction losses until f11 - and with so much detail on the huge image, you can safely use f22 for extreme depth of field.

Here is the 17-35mm KM (D) f2.8-4 aka Tamron in drag, used at 17mm and f22. ACR processing used -17/+17 on the CA removal and Highlight fringe suppression, but really, the image was very clean to start with. Some capture sharpening and some final web sharpening (100, 0.7, 0) have been applied as this helps improve the detail which is softened by stopping down to the minumum aperture of the lens.

Image

Original file - full size - be warned this is a 6 megabyte plus image and will take some time to build from pBase if you click ORIGINAL as your viewing choice!

http://www.pbase.com/davidkilpatrick/image/103392090

Right now, I'm happy that the 17-35mm will do perfectly good service on the Alpha 900 and can be used on it just as I would have used it on film.

David
User avatar
pakodominguez
Minister with Portfolio
Posts: 2306
Joined: Tue May 22, 2007 5:38 pm
Location: NYC
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by pakodominguez »

David Kilpatrick wrote: Here is the 17-35mm KM (D) f2.8-4 aka Tamron in drag, used at 17mm and f22. ACR processing used -17/+17 on the CA removal and Highlight fringe suppression, but really, the image was very clean to start with. Some capture sharpening and some final web sharpening (100, 0.7, 0) have been applied as this helps improve the detail which is softened by stopping down to the minumum aperture of the lens.
David,
Can't we expect any better from this lens? I mean images CA free? I don't like the idea of PP every single pic I'll take with the 17~35 f2.8-4... Will the Fisheye have less CA problems?

We had a discussion last winter at the Minolta Yahoo groups (http://tech.groups.yahoo.com/group/Mino ... age/143159) about the constraint of wide angles due to the lens-to-film/sensor flange: our dear Chuck Cole stated that, wider than 28mm is really complicated to get a good wide angle. I used to like my 20mm on film, but now I don't really know anymore...

In other hand, all your pics looks good but not that punchy colorwise. I guess the fall sunlight in Scotland isn't helping, or this camera hasn't produce richer colors than that? The few samples I took yesterday didn't really helped me about this because I couden't take the camera farther than the sidewalk around the Sony Plaza (Madison and 55st, midtown Manhattan) and everything was in the blue cast...
Pako
------------
http://www.pakodominguez.photo/blog" onclick="window.open(this.href);return false;
David Kilpatrick
Site Admin
Posts: 5985
Joined: Sat May 19, 2007 1:14 pm
Location: Kelso, Scotland
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by David Kilpatrick »

pakodominguez wrote:
David Kilpatrick wrote: Here is the 17-35mm KM (D) f2.8-4 aka Tamron in drag, used at 17mm and f22. ACR processing used -17/+17 on the CA removal and Highlight fringe suppression, but really, the image was very clean to start with. Some capture sharpening and some final web sharpening (100, 0.7, 0) have been applied as this helps improve the detail which is softened by stopping down to the minumum aperture of the lens.
David,
Can't we expect any better from this lens? I mean images CA free? I don't like the idea of PP every single pic I'll take with the 17~35 f2.8-4... Will the Fisheye have less CA problems?

We had a discussion last winter at the Minolta Yahoo groups (http://tech.groups.yahoo.com/group/Mino ... age/143159) about the constraint of wide angles due to the lens-to-film/sensor flange: our dear Chuck Cole stated that, wider than 28mm is really complicated to get a good wide angle. I used to like my 20mm on film, but now I don't really know anymore...

In other hand, all your pics looks good but not that punchy colorwise. I guess the fall sunlight in Scotland isn't helping, or this camera hasn't produce richer colors than that? The few samples I took yesterday didn't really helped me about this because I couden't take the camera farther than the sidewalk around the Sony Plaza (Madison and 55st, midtown Manhattan) and everything was in the blue cast...
Well, the colours look fine to me - very much as they should be. I do not emphasise colours artificially, and a grey boat on a brown river with a bit of green grass does not have a big colour palette. Shots like the close-up of the oars and net with the 28-75mm have very little colour, but I like the rendering of the neutrals - far more difficult for the camera to get right than saturated colours. And this is, also, in the two-hour period before sunset with low sun and naturally yellow colours.

The 17-35mm has less CA than the sample of the CZ 16-35mm I tried, and less CA to correct than my CZ 16-80mm, Sony 18-250mm, or Sony 16-105mm, or 11-18mm KM or Sony lenses. Overall it's the best wide angle I have now, when used on the A900 - lowest distortion, lowest CA correction needed (+17-17 on a 24 megapixel output, as opposed to +50-0 needed for the 11-18mm at 11mm, and +30-0 needed for the CZ at 16mm).

Unless final production samples of the 16-35mm f2.8 eliminate strong moustache shape distortion at 16mm plus fairly crisp colour fringes, the 17-35mm D strikes me as the best buy I ever made. I can tell you now, because I have received samples and worked with the files, that the 17-35mm D on the A900 is far superior for corner to corner coverage and sharpness than the Canon 17-35mm or 17-40mm L. It is not as good as the Nikon 14-24mm, and I have only used the Canon 16-35mm f2.8 on their tiny 1.6X factor sensors so I don't know how that works on full frame.

David
User avatar
pakodominguez
Minister with Portfolio
Posts: 2306
Joined: Tue May 22, 2007 5:38 pm
Location: NYC
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by pakodominguez »

David Kilpatrick wrote: The 17-35mm has less CA than the sample of the CZ 16-35mm I tried, and less CA to correct than my CZ 16-80mm, Sony 18-250mm, or Sony 16-105mm, or 11-18mm KM or Sony lenses. Overall it's the best wide angle I have now, when used on the A900 - lowest distortion, lowest CA correction needed (+17-17 on a 24 megapixel output, as opposed to +50-0 needed for the 11-18mm at 11mm, and +30-0 needed for the CZ at 16mm).

Unless final production samples of the 16-35mm f2.8 eliminate strong moustache shape distortion at 16mm plus fairly crisp colour fringes, the 17-35mm D strikes me as the best buy I ever made. I can tell you now, because I have received samples and worked with the files, that the 17-35mm D on the A900 is far superior for corner to corner coverage and sharpness than the Canon 17-35mm or 17-40mm L. It is not as good as the Nikon 14-24mm, and I have only used the Canon 16-35mm f2.8 on their tiny 1.6X factor sensors so I don't know how that works on full frame.

David
I know colors is about personal taste, the color on your pics aren't bad, just too neutral for what I use to look for -I avoid over saturation, but I like rich colors: you know, I came from the tropic... Since my work is to get good looking prints, the way I prepare files for printing is not the same than the way you work on pics you'll watch in the screen.

Don't take me wrong about the 17~35 f2.8-4, it's a lens I like very much, but that I'll use less on full frame: I like wide angles but I don't like to abuse... Do you thing the new CZ 17~35 will resolve better CA? and, It will resolve it 3 times (1000 US$) better? that's the difference on the price tag... Do you own the 20 f2.8? had you tried on the A900? How about the 100 f2.8 SF? I'll be going next week, early in the morning so I'll get a better light, to the Sony Plaza and try different lenses on the camera. I only used the 17~35 f2.8-4 and the 35~200xi (the new I got last month) and CA is strong. The 35-200xi is a little to soft on the corners.

Canon lenses (17~35 and 17~40) aren't a reference. They aren't even good on APS cameras...
Pako
------------
http://www.pakodominguez.photo/blog" onclick="window.open(this.href);return false;
David Kilpatrick
Site Admin
Posts: 5985
Joined: Sat May 19, 2007 1:14 pm
Location: Kelso, Scotland
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by David Kilpatrick »

Pako, here is a 17mm shot as an in-camera JPEG so no CA reduction applied. It has been resized to A700/D3/D700 pixel size (12 megapixels). How could you be disappointed by this?

Image

http://www.pbase.com/image/103402323

I have tested dozens of combinations - for example I have the Olympus 7-14mm here right now - and I can tell you that in terms of absolute detail rendering (identical print size) and CA relative to the format, even with the cheap and now obsolete 17-35mm f2.8-4 D KM lens, the A900 is winning all the way.

It does appear to have very aggressive vignetting reduction in-camera - sensor shading - and this seems to be focal length sensitive, as the 70-300mm SSM vignettes more than the 17-35mm D!

I have only used a prototype 16-35mm, and it had much sharper 'defined' CA just like the 16-80mm CZ. Unless the final production is much better than the prototype I won't be buying. The 20mm f2.8 I have tried; it's very poor on the A900. The image has severe distortion and CA. This was suspected from the start because of the age of the lens. The 17-35mm, at 20mm, is much better.

David
Javelin
Emperor of a Minor Galaxy
Posts: 1856
Joined: Wed Jul 16, 2008 4:51 pm

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by Javelin »

I have the KM 17-35 and I like it very much. I carry it with my beercan as a good combination. that may change now that I have the 28-135 though.
User avatar
pakodominguez
Minister with Portfolio
Posts: 2306
Joined: Tue May 22, 2007 5:38 pm
Location: NYC
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by pakodominguez »

I'm not disappointed -I like this lens. You can follow a couple of discussion between Ray (Sybercitizen) on the Minolta Yahoo groups while I defend this lens while he thinks the 17-80 is much better (...)
This is what I got yesterday.
17~35 f2.8-4
ftp://ftp2.phototeka.net/phototeka/A900 ... -small.jpg
Full size:
ftp://ftp2.phototeka.net/phototeka/A900 ... ualDRO.JPG

35~200xi
ftp://ftp2.phototeka.net/phototeka/A900 ... -small.jpg
Full size:
ftp://ftp2.phototeka.net/phototeka/A900 ... NRauto.JPG
I have to work more with this camera + my lenses, but I won't be able to buy the camera before the end of the year...
Pako
------------
http://www.pakodominguez.photo/blog" onclick="window.open(this.href);return false;
David Kilpatrick
Site Admin
Posts: 5985
Joined: Sat May 19, 2007 1:14 pm
Location: Kelso, Scotland
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by David Kilpatrick »

Looks to me as if the 35-200xi, which I know is quite bendy at the long end but very sharp indeed on film, will work well on the A900 too. I have one here so may do some tests - but I have far too many different lenses to try!

David
m-a
Acolyte
Posts: 4
Joined: Thu Sep 25, 2008 1:05 am

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by m-a »

David Kilpatrick wrote: Unless final production samples of the 16-35mm f2.8 eliminate strong moustache shape distortion at 16mm plus fairly crisp colour fringes, the 17-35mm D strikes me as the best buy I ever made. I can tell you now, because I have received samples and worked with the files, that the 17-35mm D on the A900 is far superior for corner to corner coverage and sharpness than the Canon 17-35mm or 17-40mm L. It is not as good as the Nikon 14-24mm, and I have only used the Canon 16-35mm f2.8 on their tiny 1.6X factor sensors so I don't know how that works on full frame.
David
Nitpick - how do you mount an EF lens on your A900? 8)

Seriously - I'm surprised to read Can*n's 17-40 L (which is 1:4 constant AFAIR) were worse than the 17-35/2.8-4(D), but then again perhaps too large a market share isn't much of an incentive for prime quality (even if there's an L).

Incidentally, a used 17-35/2.8-4 (with KoMi branding) in mint condition crossed my way a couple of weeks ago, and except for the egg-like bokeh on film (less visible near the center, so less to worry 'bout on the DX sensors) I've found nothing to complain about, it's not the lenses fault that most DSLRs crop around the center -- and if I expect to need real WA, it's "load film into the Dynax" - providing I remember to take the film from the freezer in time ;-)

Matthias
artaphot
Acolyte
Posts: 5
Joined: Sat Nov 15, 2008 11:31 am

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by artaphot »

David Kilpatrick wrote: ...

It does appear to have very aggressive vignetting reduction in-camera - sensor shading - and this seems to be focal length sensitive, as the 70-300mm SSM vignettes more than the 17-35mm D!
Any "hard" information on this? You knowof course that the beercan-generation 70-210 and 75-300 have very strong vignetting both on film as well as on A900? Regarding vignetting, I can't see any difference between film and A900 - therefore I would be very careful when making claims such as "It does appear to have very aggressive vignetting reduction in-camera".
David Kilpatrick wrote: I have only used a prototype 16-35mm, and it had much sharper 'defined' CA just like the 16-80mm CZ. Unless the final production is much better than the prototype I won't be buying.
David
Similar first impression as I had. But let's see. The first production samples may arrive end of January.
David Kilpatrick wrote: The 20mm f2.8 I have tried; it's very poor on the A900. The image has severe distortion and CA. This was suspected from the start because of the age of the lens. The 17-35mm, at 20mm, is much better.

David
Obviously something's wrong with your 2.8/20mm. Look at the careful comparison I did (2.8/20mm, 3.5/17-35mm G, 2.8/17-35mmG):
http://artaphot.ch/index.php?option=com ... &Itemid=43" onclick="window.open(this.href);return false; (scroll down for 20mm test)

Look at this 6000x4000px sample of my 20 y old 2.8/20mm:
http://artaphot.ch/index.php?option=com ... Itemid=105" onclick="window.open(this.href);return false;

A sample of the Sony SAL 2.8/20mm I checked recently in a local shop seems to be even better than mine (new coatings, circular aperture).

Greetings ;)

Steve
sparaxis
Initiate
Posts: 42
Joined: Fri Nov 07, 2008 10:30 pm
Location: Baltimore USA

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by sparaxis »

There seems to be quite a lot of variability with this lens. In the pre-digital days the German Magazine "ColorFoto" did mind bendingly thorough test reports. The earlier ones...(Somewhere in a box in the basement) rated the lens as poor, worse than the Canon and Nikon equivalents.
Their last "pre-digital best list" gives it 71 points vs 67 for Canon, and 70 for Nikon. (Sample variation perhaps).
On the other hand, they show the 24mm as getting 73 points. MHohner's site also shows the 24 as a bit better.
(Mine is very good too)

Amateur Photographer tested the lens in 86 and called it excellent and again in 94 when it was only very good.
Of course, in those days Amateur photography's tests were still rather informal. Going back far enough...Who remembers the Ship used as the target of way too many tests?
Perhaps David has a sub-standard 20mm and Steve a sub-standard 24mm.

<grin>

(I have my good bad and ugly list - maybe I'll write about them someday!)

Alan
User avatar
alphaPDX
Heirophant
Posts: 80
Joined: Mon Oct 13, 2008 8:53 pm
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by alphaPDX »

David Kilpatrick wrote:Pako, here is a 17mm shot as an in-camera JPEG so no CA reduction applied. It has been resized to A700/D3/D700 pixel size (12 megapixels). How could you be disappointed by this?

I have tested dozens of combinations - for example I have the Olympus 7-14mm here right now - and I can tell you that in terms of absolute detail rendering (identical print size) and CA relative to the format, even with the cheap and now obsolete 17-35mm f2.8-4 D KM lens, the A900 is winning all the way.
I have the Tamron version of this & like it with my a200. I had heard less than stellar comments about this on FF cameras, so hearing this is good news. I'm not upgrading soon, but it is a comfort to know that this lens and my Minolta 35-105N have a future when I'm able to move to FF.
Jim R, Oregon -- a200 + lenses & stuff
User avatar
pakodominguez
Minister with Portfolio
Posts: 2306
Joined: Tue May 22, 2007 5:38 pm
Location: NYC
Contact:

Re: Diffraction defeated by pixel count - 17mm, f22

Unread post by pakodominguez »

sparaxis wrote:There seems to be quite a lot of variability with this lens...
...
Who remembers the Ship used as the target of way too many tests?
Perhaps David has a sub-standard 20mm and Steve a sub-standard 24mm.
Well, apparently I got a "sub-standard" 20mm too, difficult to get the focus right. And, when in focus, too much CA for my taste. Went to EBay a couple of weeks ago and sold for $300+ (I did paid $145 for it 4 years ago) that money will finance the Tamron 70-200 f2.8.

I can tell you that I'm not that happy about that, the 20mm was my favorite lens (on NiNikon) back in the film days.

Regards
Pako
------------
http://www.pakodominguez.photo/blog" onclick="window.open(this.href);return false;
Post Reply

Who is online

Users browsing this forum: No registered users and 19 guests