It's Not The Camera...

...It's the lenses... 

OK, we all know that it's the photographer that makes the photos, not the camera.  Yada yada yada. 

And after that, we've heard that your money is better spent on glass.  Despite that rejoinder I spent some money on a camera last month.  Now I'm wondering about my lenses...

 For the past decade I have been shooting "crop factor" DSLRs. These are cameras with a sensor smaller than an actual 35mm film frame, so the area covered by the images is smaller than it would be if the film plane was "full frame."  With the 1.5x crop factor, a "normal" 50mm lens delivers the frame of 75mm.  Now, mind you, the optics don't change.  You don't get the additional magnification you would get with a 75mm lens - you get the same perspective you would get with a 50mm lens -- you just get less of it.  

The issue is well illustrated in the photo below of the Dunluce Castle ruins in County Antrim, Northern Ireland (a place I hope to visit and photograph one of these days), which Google found for me at the Peter Cox Photo Courses website.   The photo was shot with an "ultra-wide"  24mm lens; the entire photo shows you what you can get with a "full frame" sensor. The white rectangle shows you what you get with a "crop-factor sensor.   It just ain't the same. 

I have been jonesing for a "full frame" sensor precisely because of what this photo illustrates: I want the vast expanse of an ultra-wide angle that lets you compose foreground and background across the expanse of the entire frame.  

The deterrent has been cost.  "Full frame" cameras (in Nikon parlance, "FX" format) like the Nikon D700  typically cost upwards of $1,000 more than their crop factor (what Nikon calls "DX" format) brethren like the Nikon D300s, which has been my principal implement of destruction  for the past almost four years. 

Last year, Nikon began to introduce a new generation of DSLRs, starting with the very high-end FX (like $6k) Nikon D4 and the D800/800e, which run around $3k.  

Then late last year they introduced the D600 - an "entry level" FX body that sells for about what the D300s sold for when I bought it 2009.  In anticipation of a return trip to Scotland  at the end of May, I bought a D600 last month and I have been trying to come to terms with it ever since. 

By "come to terms" I mean that I've been disappointed in some of the results.  I'm expecting the photos that come out of this camera to be sharp, detailed and noise free.  And for sure, some of the images that have come out of the camera have matched that description.  

But some of the images have been been... well, fuzzy.  Soft.  Lacking in fine detail.   So I've set about to figure out why. 

March 19, 2013:  After 10 years of digital photography, finally a "full frame" DSLR! 

The problem is illustrated in the photo below, which was shot at the new Cotten Music store in Nashville.  The light in the space is mostly kinda dim, so I had to crank the ISO on the D600 up to 1250; then I started with my  Nikon 50mm f1.8 - kinda the bargain basement of "normal" lenses.   Of course, I haven't had a "normal" lens since I started using DSLRs.  So shooting with a 50 on a full frame is a revelation. You can have all kinds of fun with foreground/background and depth of field - like focusing on the headstock of this vintage National "Triolian" resonator guitar: 

For this shot, I set the autofocus detector in the center of the frame, focused on the headstock, and then recomposed to include the guitar player (the relatively small spread of autofocus points in the viewfinder is whole other issue that I'm not getting in to here; this is about the lenses, not the camera). 

At first glance, the whole frame looks pretty nice: focused on the headstock, the guitar player out of focus filling the rest of the frame.  But lets take a closer look at the details: 

Nikon D600:  ISO 1250; Nikon 50mm f/1.8 lens: 1/60th at f/1.8

Here is the detail in the headstock of the guitar at 100% enlargement.  Look at the characters in the label.  Does that look sharp to you?  It looks legible to me... but not sharp. 

Now, admittedly, that fuzziness could be attributable to any number of characters:  hand held at 1/60th of a second while crouching on my knees for starters; the moderately high ISO.  Or just maybe its something else.  Like the lens itself... 

So let's look at another shot.  This is the lovely Kim Peery Sherman, co-owner of Cotten Music, as seen through some of the fabulous inventory of vintage guitars found at the store.  

I thought I'd done a great job of capturing Kim's expression. I'd just told her "Kim, you just sold a $30,000 guitar!" she broke into laughter and this was shot in a moment after "peak laugh."  

I was fairly pleased with myself for the angle and the moment.  And then I zoomed in to see how it looks in the details: 

Nikon D600: ISO 1600; Nikon 50mm f/1.8 lens; 1/30th at f/2.8

Same shot, detail at 100%.  Again, there could be a lot of factors at work here to cause the lack of detail: hand held 1/30th of a second, moderately high ISO.  But by now I'm starting to wonder if there isn't some other factor at work here.  Like... not the camera, but the lenses themselves. 

So... Lens Testing

So last night I set up a little stand in my office and proceed to test three of my "prime" (fixed focal length, not zoom) lenses to see how they measure up under controlled conditions.   Per some  instructions I found on YouTube, I placed a Coke can on a pedestal, put my camera on a tripod, set up an outboard flash (Using Nikon's Creative Lighting System or CLS),  set the ISO at the camera's native setting (ISO 100) and used a 2sec delayed self time to release the shutter to minimize shake.  I started with the 50mm f./1.8 lens since that was the one that was confounding me at Cotten Music.  I focused on the dead center and used a shutter speed of 1/125 for all these tests.  CLICK ON ANY IMAGE TO ENLARGE IT IN YOUR BROWSER. 

Nikon 50mm f1.8

50mm f/1.8 lens @ f/1.8 Now, this may look sharp at first.  But let's go to 100%... 


...and look at what the lens renders when it's wide open at f/1.8  You think that might explain the fuzziness in the Cotten Music stuff?
From the YouTube tutorial, I found a chart that lists the sharpest apertures for all of the lenses in the Nikon line.  It says that this lens is supposed to be sharpest between f/4.0 and f/16.  So let's see what we get at f/8: 
50mm f/1.8 @ f/8: At full frame, you can already see the lens is sharper at this aperture...
...and at 100% it's waaaay sharper than at f/1.8, though still not exactly crisp.  
50mm f/1.8 lens @ f/16.  So let's go a couple more stops, this time to f/16.
Same, at 100%: I would say that's fairly sharp, and the depth of field at this aperture even renders the lettering on the side of the can readable.

But f/16 is an awfully small aperture to have to close down to to get a sharp image from a lens that's supposed to be able to focus at f/1.8!

I went on and tried a couple of more lenses.  I've recently become re-acquainted with the Nikon 105mm Micro (aka "macro") f/2.8, which I was almost ready to put on eBay because I hadn't used it in such a long time.  But I pulled it out a few nights ago and liked the depth of field (or lack of it at wide apertures) and may use it again.  So let's see how this lens does with our Coke can: 

Nikon 105mm f/2.8

Nikon D600;  ISO 100; Nilon 105mm "Micro" f/2.8 lens @ f/2.8


Same shot; detail at 100%.  Close enough for government work?   Let's see what we get at f/8:
Same shot, detail @ 100%.  Sharper than the f/12.8, but more a matter of degree than magnitude like the 50mm.  Nice depth of field too.
Nikon D600;  ISO 100; Nilon 105mm "Micro" f/2.8 lens @ f/8
And finally, some tests with my revered Nikon 24mm f/2.8 - the lens I have waited 10 years to get the full "panoramic" ultra-wide angle effect from.  

Nikon 24mm f/2.8

Nikon D600;  ISO 100; Nikon 24mm  f/2.8 lens @ f/2.8.  The Coke can is 15-3/4" away from the focal plane of the camera.  Now that's "wide." 

New Image(s)

This is 100% detail from the 24mm lens at f/2.8.  Detail is obviously harder to discern at this focal lenth.  Particularly looking at the lettering on the side of the can, I'd say this one is reasonably sharp.  
And here's the same setup at f/8.  
Not a whole lot of detail at this focal length - or difference in the apertures. 

So, what's a fella to do? 

It seems to mw that the conclusions to draw from  this litte experiment are:

     1) There's nothing wrong with the camera, in fact I shot the same series of tests last night with my D300s, and, apart from the crop factor issue, got essentially the same results;

     2) There really is a difference in sharpness from one aperture to the next, and the "sweet spot" for any                            lens is well within the extremes at either end of the aperture range; 

     3) That 50mm f/1.8 lens?  That's clearly an issue.  I've had the lens for a long time - possibly most of the 10 years that I've been shooting Nikon DSLRs.  Maybe it's just taken one-too-many blows to the focus ring.  It's not an expensive lens, less than $150, so maybe I've gotten all the life I can reasonably expect from it.  Perhaps I should consider upgrading to the 50mm f/1.4.  That's about a $450 proposition, and the extra half-stop might mean it will be sharper at, say f/2.8 or f/4.  

That chart with the f/stop details says that the 50 f/1.8 should be sharp from f/4 - f/16 - and clearly that is not the case with this particular item.   The 50 f/1.4 should be sharp from f/2 up to only f/5.6.  That's not much sharpness range for a $400+ lens.  

So as long as you're reading this, I'd appreciate any feedback you care to share, on this whole subject of lenses, apertures, and autofocus.  You can post comments here, or send an e-mail to driver at 49chevy dot com.  

Thanks for following along...