Sony 10 – 18mm f/4 E Lens Review
by Mark D Segal
The Sony NEX 5N and NEX 7 are superb cameras. Raw files from the NEX 5N can deliver stunning 13X19" prints quite easily; so I never regretted buying this camera for its extreme portability and discretion when traveling. It packs a lot of quality in a very small package. The only frustration has been the relatively small number of lenses that are made for it - an improving story as the product line matures. Yes, I do have the (costly) LA-EA2 adapter and it does extend the lens options considerably, but the pickings are thin in the super wide-angle zoom category (lenses in the range of 10~18mm zoom). I've tried three different ones using the adapter and returned them, because sharpness at long ends of the images was simply inadequate (to my liking) to justify the high prices one pays for them. Two of the three were also quite heavy and all of them used with the adapter added much bulk to the camera, impairing the beauty of handling it.
While on the subject of returning lenses – do yourselves a favour and buy from retailers who will take them back and refund your money without fees if you're not happy after a short trial period. This is particularly important for these zoom lenses that are challenging to design and manufacture, and experience indicates that not every piece is necessarily optimal.
So I was patiently awaiting the day when Sony would produce a super wide-angle zoom with an E-Mount for the NEX. Last week they finally hit dealers' shelves and my patience was rewarded. The NEW Sony 10~18mm super wide angle zoom really has some good things going for it.
What's to Not Like?
I'll dispense with this first, because there is so little to not like about this lens, we may as well get it out of the way: the housing and dials are made of plastic. Well, OK, there are various high-tech plastics on the market these days that are supposed to be very durable, so maybe it's just a bit of snobbishness and a legacy of handling really robust steel casings and controls since back in the 1950s, but I can't help seeing it as kind of cheap to produce a $900 lens with plastic casings and dials. The flip side if this probably contributes to one of its endearing qualities, to which I turn.
What's to Like?
Weight – or the absence thereof. Weighing in at 225 grams (8 oz., according to the specs, but it feels like less), it's very light and doesn't need an adapter. In fact, using this lens feels no different than using the 18~55mm kit lens, which is quite a relief. Second, it's smooth – rotating the dials is silky and silent. Third, it's f/4 at all focal lengths – no compromise on aperture related to focal length. Fourth, the lens shade is compact. Fifth, it has stabilization, auto-focus and auto-exposure, all of which integrate with the camera, and work as intended. Sixth, the focal length markings on the lens barrel are very distinct, hence easy to read.
The figures immediately below compare the sizes of the w/a zoom on the left with the 18~55mm kit lens on the right, and show the camera fitted with the w/a zoom. So you can see what I mean.
1. Super w/a and 18~55mm kit lenses compared 2. Super w/a on Sony NEX 5n
Delivering the Goods
OK, form and feel done-with, let's turn to the substance: does it make good photographs? In a word: Yes.
I decided that for 900 dollars (plus taxes) I should give this lens a good workout to decide whether it's a keeper. I ran several series of shots each with a slightly different purpose:
(1) The proverbial brick wall to test for texture rendition, curvature, distortion and resolution from close-up - challenging for a wide lens with flat-field subject matter.
(2) The front of a house with sharp contrasts and mild textures to test for chromatic aberation (CA), edge rendition and detail resolution from a short distance (within 15 meters).
(3) Broad downtown cityscapes, to test for CA, resolution, and distortion from a longer distance (well over 50 meters).
(4) A shopping mall interior, to test for resolution and distortions across nearby and distant subject matter in the same image.
I didn't photograph test charts and didn't use a tripod, because I wanted to test the lens under conditions of normal usage (for me and I think most people using these cameras), which is handheld capture of real-world subject matter. What interests me, and I expect you, is whether the system as a whole delivers good photographs of real subject matter under routine operational conditions.
For each subject category I ran a matrix of focal lengths and apertures: 10, 12, 14, 16 and 18 mm focal lengths, each at f/4, f/5.6, f/8, f/11 and sometimes f/16 apertures. In all, I made about 150 images. No, I wouldn't even think of posting all of that here - too much work, too boring and unnecessary; but I do show below what I believe are good representative samples of the lens' performance. All results of course are from the particular copy of the lens that I purchased.
A word about post-capture evaluation: these all started life as Sony raw files about 16MP each. I view them in Lightroom 4.2 using Lightroom's default sharpening, because again, I know under normal operational conditions most digital captures need at least an initial round of capture sharpening. For all images I applied Lightroom's automatic CA correction tool, because it's needed, and it moderately improves overall image definition, as expected. I must say, there were very few of these images that needed any further defringing - the Lightroom CA tool is very cool indeed. The comparison below shows a typical case of CA (left) and its corrected counterpart (right). (Note: all illustrations in the articles are converted from raw to JPEG at 80% quality in sRGB colour space for web viewing.)
3. Red and Green arrows point to uncorrected CA 4. CACorrected using Lightroom 4's CA correction tool (one check-box)
The photo is at 10mm f/5.6 and the scene is toward the upper right corner of the buildings about 50~75 meters distant. This extent of CA we're seeing here is typical of most images in the test runs and very moderate, but deserves correction.
All told, the lens performs well at all focal lengths. At all focal lengths, the outer edges of the landscape orientation are somewhat less sharp than the more central areas, but this is expected, and with sharpening, most of it is usable image detail. Areas of varying (but usually small) size in the corners can be less usable. Detail rendition is best in the range of f/5.6 to f/11, again very much as expected. f/4 is a bit more challenging at the edges, while at f/16 the effect of diffraction becomes mildly visible. Viewed in Lightroom Process 2012 with all luminance and colour adjustments at zero, default sharpening, and White Balance "As Shot", the images have good sharpness, contrast and saturation "out of the box". In the samples below, except where mentioned, the only post-capture correction I performed in this evaluation context is for CA.
The Brick Wall Comparisons:
Starting with the worst-case scenario, here's 10mm, f/4, the whole image, followed by a snippet from the center (left image) and from the lower right side (right image). There is slight barrel distortion, not surprising for such lenses and mild enough to be corrected easily with Lightroom or Photoshop Lens Correction tools. The overall sharpness is acceptable but not stellar, the center being sharper than the side edges.
5. 10mm, f/4, full image
6. 10mm f/4 detail at 100% from center of photo 7. 10mm f/4 detail at 100% from lower right side of photo
Changing the aperture from f/4 to f/5.6, there is a noticeable up-tick of overall image crispness, making it very good indeed (see images of the whole photo and the center and right-side snippets below).
8. 10mm, f/5.6, full image
9. 10mm f/5.6 detail at 100% from center of photo 10. 10mm f/5.6 detail at 100% from lower right side of photo
The right side is not 100% of the crispness of the center, but this is normal for the kind of lens, distance and subject matter at hand; for this class of lens the result is nonetheless very good and can be further improved with output sharpening. The results at f/8 and f/11 were ever so slightly sharper, but the difference would be hardly noticeable converted to JPEG for screen viewing. I show below the result for 10mm f/16, to illustrate the slight onset of diffraction's impact on resolution - if you can see it by comparison of center snippets just above and below. The image is still OK, and amenable to further improvement with output sharpening, but just not quite as good as its f/5.6~f/11 comparators.
11. 10mm f/16 detail at 100% from center of photo
Finally, it's worth having a look at image resoluition half way up the zoom scale - 14mm, the photo below at f/8. Notice the reduction of barrel distortion. Below it is a snippet showing resolution of the same image from the center and most of the right side. Results at 16mm and 18mm are similar, but of course the subject matter is closer up and the difference of detail resolution between center and sides less apparent.
12. 14mm f/8 full image
13. 14mm f/8 center to right side at 100% magnification
The House Front Comparisons:
Mercifully leaving the world of brick walls to something a bit more interesting, another challenging shot for a wide angle zoom, given the great range of distance from where I stood to the roof, and the distance from which I wish to capture fine texture. The first set of illustrations below are the worst-case scenario of 10mm at f/4. The photo was focused on the front door.
14. 10mm f/4, full photo; (and my shadow making the photograph)
15. 10mm f/4 detail at 100% from lower right 16. 10mm f/4 detail at 100% from center of photo 17. 10mm f/4 detail at 100% from upper left
For comparison, the set below are the same except for the change of aperture to f/5.6, where, consistently with the brick wall shots, there is improvement of overall sharpness.
18. 10mm f/5.6, full photo
19. 10mm f/5.6 detail at 100% from lower right 20. 10mm f/5.6 detail at 100% from center of photo 21. 10mm f/5.6 detail at 100% from upper left
Finally, given the distance between the focal point (front door) and the roof area, the shot at 10mm f/11 (below, left) shows how the lens renders the stucco texture with the increased depth of field, and before diffraction becomes a noticeable factor. To bring out the texture in the samples just below, I changed Lightroom's sharpening from the Default to the "Scenic" preset and toned-down the highlights so that the detail isn't obscured by brightness (the raw file has no clipped highlights). The photo to the right is at 16mm f/11, to show the result of greater focal length and reduced aperture, using the same focal point - the front door.
22. 10mm f/11 detail at 100% from upper left 23. 16mm f/11 detail at 100% from upper left
The Cityscape Comparisons:
Our viewers who know Toronto will recognize Dundas Square in the photo immediately below, made at 10mm f/4. The only editing is Lightroom default sharpening, CA removal and a slight reduction of Black intensity to correct a small amount of shadow clipping. Of course the image is keystoned because of the extreme wide angle and the perspective from which the photo was made. The three photos below the full scene are details from the same image, center foreground, upper left and upper right sides.
24. 10mm f/4, full photo
25. 10mm f/4 detail at 100% from center foreground 26. 10mm f/4 detail at 100% from upper left 27. 10mm f/4 detail at 100% from upper right
The same scene, photographed at 10mm f/11, the details below show the center portion of the buildings in the background and the improved sharpness of the center foreground, indicating satisfying depth of field.
28. 10mm f/11 detail at 100% from center back 29. 10mm f/11 detail at 100% from center foreground
Finally for the cityscapes, I show what could be a very typical shooting situation - 14mm, f/11. Here I did a bit of additional adjusting to lighten the shadows and darken the highlights, allowing the image's inherent detail to show through more easily. The images below it are magnified segments, showing some fall-off of quality in the lower left (I brightened the JPEG to better show the result), but commendable quality shortly beyond that area. The lower left foreground is of course much closer to the lens than the focal point of the image, which was in the center area across the plaza to the right of "Milestones". The detail of the center area, #30 below left, also shows the detail picked-up of the buildings a block or two away in the background to the left of "Milestones".
30f. 14mm f/11, full photo
30. 14mm f/11 detail at 100% from center back 31. 14mm f/11 detail at 100% from left edge foreground
The Shopping Mall Comparisons:
To remind, the purpose of these comparisons is to see how the lens handles a wide-angle sweep of a scene with a lot of narrow depth. The sides in the foreground are much closer to the camera (indeed almost beside me) than the distant centre. I was positioned at the middle of a foot-bridge that joins the two sides of the mall interior at third-floor level, providing an excellent view of everything from the basement to the domed roof, and probably photographed by countless numbers of people. That's OK, it serves my purpose - originality of the subject matter isn't part of this agenda. Because of where I was positioned, the keystoning of these images was minor, so I coundn't resist the temptation to adjust it, but keeping all the image information inside the frame. The only other adjustments are Lightroom default sharpening and CA correction. I increased ISO from my usual default of 200 (for most of the outdoor shots) to 400 for these interiors, because the lighting is lower and regardless of the image stabilization, I'm not comfortable using shutter speeds of less than 1/50th of a second for hand-held photography. Motion blur is the last thing we need for testing the performance of a lens.
As usual, the first photo below is the worst-case scenario of 10mm at f/4. I must say, when I first saw this result I was truly impressed with it. The focal point, by the way, is down by the third Indigo sign, yet the material on the side walls, almost beside me at this focal length, is truly very clear.
32. 10mm f/4, full photo with keystone correction in Lightroom 4.2
33. 10mm f/4 detail at 100% left side foreground 35. 10mm f/4 detail at 100% from center back 34. 10mm f/4 detail at 100% from right foregroiund
The center snippet at 10mm f/5.6 below seems a tad sharper than its counterpart above at f/4. Results at f/8 and f/1 are similar, with slightly better depth-of-field.
36. 10mm f/5.6, detail at 100% from center back
Turning to the samples photographed at 14mm, f/5.6 below, the overall crispness and detail are really very good, seen from the entire photo, or the center snippet.
37. 14mm f/5.6, entire photo, no perspective corrections
38. 14mm f/5.6, detail at 100%, center
The very lower left corner exhibits some distortion and comparatively less resolution, as expected; however it takes little movement above that area for the image to come into quite satisfactory focus with minimal distortion, as seen in the two snippets below.
39. 14mm f/5.6 detail at 100% lower left corner foreground 40. 14mm f/5.6 detail at 100% from left side, foreground
The photos shot at higher apertures required increased ISO - to 800 - at which point noise begins to interfere with an uncluttered appreciation of the lens; save for the noise, image resolution is fine at least through f/11 - and with better depth of field than at f/5.6, focus of the foreground corners is improved, as shown in the final illustration below.
41. 14mm f/11, detail at 100%, lower left corner foreground
I believe this small sample of my large number of test images provides a representative overview of what I achieved with this lens. It is the best of the four super wide-angle zooms I've tried, and I consider it a keeper.
Mark D. Segal
Toronto, November 2012
Mark is a regular contributor to the Luminous-Landscape, particularly in the areas of scanning, printing and equipment testing. Mark recently wrote a book on scanning with SilverFast 8, including integrated workflows with SilverFast 8, Adobe Photoshop Lightroom and Adobe Photoshop, published by LaserSoft Imaging and available on the SilverFast website.