Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Nikon 17-35mm f2.8 AFS (The one you've all been waiting for) Review

0.28M or 0.9ft for the focus distance, according to the users manual. :)
 
Evin, how did you create these tif files from the test?
So far I was looking at these files on my 1920X1200 24” Dell screen and all was looking good.
However today I took few pictures of yours and I fed it in to my 52”Bravia so I wanted to see how it will look for the average HDTV viewer as a kind of RED camera shot that was paused and one paused frame is displayed on the HDTV screen at a time. It was not a pretty picture.
I went back to my Dell screen and I magnified your picture 200% so I got approximately the same size of the pixels as on my Bravia. The same thing popped up.
The left of the details is visibly greenish the right is reddish. The power cables hanging across the backyard (the out of the window shot) are digitized in to the very bad staircase effect. The little white roof above the window (inverted V shaped) across the backyard has clearly blue edge on the left side of it and the red edge on the right side.
Is this my tif reader doing it or your conversion of Nicon RAW to the tiff was not done in the best possible way.

BTW HDTV is fed via HDMI interface and DEL screen is fed via DVI
 
  • Thread starter
  • Moderator
  • #63
First of all those Tiffs are 4K, so by blowing them up 200% you are seeing them at a distance no one will ever see them, essentially a foot or two from a 40' movie screen. Scale them to 1920x1080 for display on your Bravia and you will get an idea of what they looklike when displayed in HD. Your dell monitor is not any more res than your Bravia but will scale better, but just understand that at 200% (double native res) nothing looks good! Because every one image pixel is now being represented by two diplay pixels, hence stair-stepping off color crap.
 
Evin, what I meant by 200% + scale up is that I did magnify your picture bit more then 2x times from what I have seen on the Dell screen that is 24” when whole your picture was seen entirely on the Dell 1920x1200 screen. I did change scale 10% +- trying to match pixel per pixel, because I was thinking that pixel scaling is not working properly. I even use two different tiff viewers and the same thing happened.

At the end I had one pixel from your tiff file represented by 1 pixel on the Dell screen exactly.

My Bravia is 52” so 52/24= 2.16
I was trying to match the size of Dell screen with the Bravia diagonal size by doing that.
Another words if my Dell screen would be 2.16 times bigger and had the same pixel size it would be not 1920x1200 screen but 4147x2332 watched from 2’ distance.
I am watching LCD TV from 5’-6’ distance and my Dell 24” screen from 2’
I agree no one is watching 52’ screen from 2’ distance but it is what actually happened when I magnify your picture on my Dell screen 2.16 times.

I have one question though, how did you get from Nicon RAW to the tiff?
Did you use Nicon utilities coming with the camera or something else?
 
  • Thread starter
  • Moderator
  • #65
Adobe Camera Raw from Photoshop CS2. It's very hard to tell exactly what's going on in your display set up, but on my 30" apple cinema display the images look stunning, there are some small issues with the wide open shots as I stated in the review, but most of those will be hidden by the moving image. Other than that, I'm not sure whhat else I can do to help.
 
Evin, these are cutouts at pixel per pixel mapping.
Do you see left edges with the bit of red and right with the bit of green?
I pull out my Cannon 100-400 and shoot the building 300 meters away at 400-f7.
Canon has sweet focus spot around 300 so 400 is the bad spot but I wanted to get this blue bar of the window and the white box there.
I see the same red/green smudging, though less pronounced and on top as on the bottom as well. Look at the top of the white container.
Left side of the blue strip is red and right is bit green. I wonder if it is debayer effect or something else. Now when I put it on the 52”screen I can see it. Maybe I am getting paranoid here. I guess it will be significant only for huge, cinema type, presentation.
Frankly speaking I wish my current HDTV movies would be as good as this.

I know it is only one pixel wide distortion, but one pixel out of two mean that 4k picture will be 2K one. I guess one more argument to shoot in 4K and downers to 2K for perfect picture. This statement is true only if it is not a lenses problem though.

BTW read this article, especially about color fringing and influence of it on green screen keying, maybe it will help you to figure out what is going on here.
I found a lot of good references on ISOE website.
http://www.zeiss.de/C12567A8003B8B...-KMu_SMPTE2002/$File/SMPTE2002_DigiPrime.pdf

Andrew
 
  • Thread starter
  • Moderator
  • #67
All lenses display a bit of Chromatic Abberation but it is usually only visable in very high contrast areas (White trim over a dark background) most modern tele lenses control it quite well using ED and UD elements, wides have a much more diffucult time at it, the 17-35 does display some CA but controls it quite well. It is also easy to correct in processing by applying a small amount of color channel re-registation, it does not affect actual resolution so it will not render your 4k image 2K. And from the math 50% of 4k is actually 3K because 4K is 4x 2k not 2x.

Also, the mysterium sensor is an important factor, it may not have as much CA issues or it may have more, we'll have to wait and see.
 
Evin Grant said:
All lenses display a bit of Chromatic Abberation but it is usually only visable in very high contrast areas (White trim over a dark background) most modern tele lenses control it quite well using ED and UD elements, wides have a much more diffucult time at it, the 17-35 does display some CA but controls it quite well. It is also easy to correct in processing by applying a small amount of color channel re-registation, it does not affect actual resolution so it will not render your 4k image 2K. And from the math 50% of 4k is actually 3K because 4K is 4x 2k not 2x.

Also, the mysterium sensor is an important factor, it may not have as much CA issues or it may have more, we'll have to wait and see.
Click to expand...

Thanks! Evin for patience, you calmed down my paranoidus greatus attack.
That is why zoom deals bit better with CA, also I was using f7 and you were f4 but this should affect only geometrical aberrations not CA I think.
Not bad shot as for 300 yards, if I did put my zoom on the tripod probably we could read what this person is drinking in this bed.
 
Go to KEH.com. I just bought one from them. They should ship internationally.
 
You might want to take care if buying used. There were quality control issues when the lens was introduced. Though from what I understand they were resolved fairly quick.
 
Evin-
In your test, the image falls apart at F2.8. I mean, you can't separate the bricks in the building at 35mm @ F2.8, and the eye pics clearly lack resolution at F2.8. I don't understand how this lens can "easily resolve 4k+" if the images don't really hold up in these little stills....imagine those 2.8 shots on a 40 theatre screen...they would look really soft, don't you think?
I also think it will be difficult to consistently pull focus on these lenses...the optics aren't the issue with quality so much as the mechanicals.
 
  • Thread starter
  • Moderator
  • #76
  • Thread starter
  • Moderator
  • #77
steevo435 said:
Evin-
In your test, the image falls apart at F2.8. I mean, you can't separate the bricks in the building at 35mm @ F2.8, and the eye pics clearly lack resolution at F2.8. I don't understand how this lens can "easily resolve 4k+" if the images don't really hold up in these little stills....imagine those 2.8 shots on a 40 theatre screen...they would look really soft, don't you think?
I also think it will be difficult to consistently pull focus on these lenses...the optics aren't the issue with quality so much as the mechanicals.
Click to expand...

I don't think you understand the test, those are 100% crops. They would represent a very small section of the movie screen and so would not "Fall apart" as you say. And all lenses are softer wide open, but this lens is better than most wides and although you might see a slight diffrence on an identical shot to shot comparison you would not pick it up if you cut from a shot at f2.8 to a diffrent angle at f5.6.
 
Evin Grant said:
I don't like MTFs there are no sample images in that test. It basically just tells you numbers and numbers don't get projected on movie screen.
Click to expand...

I usually look on the numbers first. MTF graphs for different f stops.
Then I look on the picture. Very unlikely that bad number lenses will perform good.
If I see that MTF drops sharply in the zone away from the center of the picture or below f5 then usually there will be problem ahead. If zoom do not perform at wide angle end, then I will try it anyway since it may work in 80% of the range very good.

Example is Canon EF-S 10-22mm works very well above 14.

Also 135 type lenses on S35 could perform exceptionally well since the sensor is two times smaller so you do not go all the way to the edge of the illumination circle of the lenses. Again it will be difficult to get 135 sensor size lenses for less then 18 mm fl so this way or the other we have to pay extra for 100 degree view lenses.
For fl of 20 to 80 there are plenty of lenses that perform more that sufficient for 4K sensor especially if lenses are for 135 sensor size.
 
  • Thread starter
  • Moderator
  • #79
MTF numbers can point you in a direction but are rarely usefull for deciding between two pro quality optics. In most cases, unless there is a design or manufacturing flaw these lenses will perform well. It's the nuanced details in the real tests that make or break a lens.
 
I'm sorry, I'm not a photoshop guy- what is a 100% crop?
Not to be nitpicky, but "all lenses are softer wide open" is not true, unless you're just talking about still photography lenses. The Zeiss Digiprimes are optimized to perform best wide open, and they do...they're beautiful!
"numbers don't get projected on movie screen." True, but if you look at the MTF data, you can determine the real sweet spot in any lens. If you look at those graphs from photozone, it seems like the real sweet spot in this lens is at F5.6....most lenses are optimized to be used 5.6-8, but it does vary. I'd think you'd want to know this before going out and shooting, especially with Red in 4k...you don't want the lens to be the limiting factor when you're trying to resolve a 4k image.
 
You must log in or register to reply here.
Back
Top