Tamron SP AF 90mm f/2.8 Di Macro : Review


Introduction

The Tamron SP AF 90mm f/2.8 Di Macro is a highly popular lens amongst amateur photographers looking forward to take themselves into the realm of macro photography. I will be showing you the results of my own tests on a Nikon D700 full-frame DSLR. This lens belongs to the SP (Super Performance) line of lenses made from Tamron, that generally have been delivering very good performance and (especially) great value for money. This lens has been replaced recently by the VC USD version which adds Vibration Compensation optical stabilization system and a silent and fast Ultra-Sonic Drive motor with the purpose of being compatible with the newest entry-level Nikon DSLRs.
Here’s how it looks with the D700:

Announced in 2004, this particular iteration (model 272E) was optimized for digital sensors, hence the Di (Digitally Integrated) designation, and received a few cosmetical changes. Internally, this lens has 12 elements in 9 groups, none of them being of any special kind, and has 9 straight aperture blades. It doesn’t have an AF motor, so it will not focus automatically on the newest entry-level DSLRs. Speaking about focus, selecting between auto and manual is done through a push-pull clutch mechanism, which isn’t exactly friendly on the field, but mainly when it doesn’t auto-focus on the exact point at first shot and one needs to override it slightly. But when doing macro photography, most of the time it’s better to use manual focus anyway, and although the focus ring isn’t damped by any means, it is easy to turn the focus ring around since the mass of the moving elements is low. The lens is very light and compact, and with a filter thread of 55mm and weighing only about 400g, it makes a very portable lens to carry around, either on the field doing macro or on the streets shooting candids unconspicuously.
The lens is mostly plastic on the outside and has a metal mount. There’s also a rotating switch on the side that limits the minimum or maximum auto-focusing distance, depending on the current focusing point – it can either limit AF between 0.44m and infinity, or between 0.29m and 0.42m, which is nice to have to speed up the AF time. The front element doesn’t rotate while focusing, but the lens extends a lot between infinity and 0.29m, which is a bad disadvantage when shooting small insects from very close distances. In this matter the Micro-Nikkor AF-S 105mm f/2.8 G ED VR, which I also reviewed, is priceless. Like most lenses made for Nikkor until 2004, the Tamron has an aperture ring and allows to lock it at f/32 for AF, otherwise the camera would show an fEE error on the display.
Today, the Tamron can only be found on the used market, and at a price of about 250€ it’s a tremendous value for money for seriously starting getting into macro photography.


Technical Specifications

Focal length 90mm
Maximum aperture f/2.8 (far distance) – f/5.6 (nearest)
Minimum aperture f/32 (far distance) – f/64 (nearest)
Field of view 27 degrees (on FX)
Weight 405g
Dimensions 97 x 72mm (148 x 72mm, extended)
Optical construction 12 elements in 9 groups
Aperture blades 9, straight
Filter diameter 55mm
Minimum focus distance 432mm (270mm from the front element, focus limiter on), 290mm (95mm from the front element, focus limiter off)
Hood 2C9FH, round
Mount Nikon F


Mechanical Characteristics

Zoom ring n/a
Focus ring Plastic and metal with rubber finish, no infinity stop
Focus throw 260 degrees (focus limiter off)
Focus motor No
Optical stabilizer No
Front element rotation while zooming n/a
Front element rotation while focusing No
Internal focusing No
Lens extension while focusing Yes, up to 51mm
Lens extension while zooming n/a
Maximum magnification 1:1


Handling

With the D700, the Tamron is very lightweight and it’s easy to hold while doing street photography or portrait work. The lens is not a beauty by any means, and feels kind of cheap in use mainly due to the almost all-plastic build and the odd focus clutch mechanism. The front element is deeply recessed from the front of the lens, about 20mm, which makes it hard to reach for cleaning, and also because of this I never used the lens hood.
In macro photography and manually focusing, the lens provides a focus ring with a long throw for careful focusing, which is always nice to have, but the lens extension is so long that it can be very cumbersome when shooting insects from very close distances. This aspect of the lens is what makes it feel so cheap in use and separates it completely from the likes of the Micro-Nikkor AF-S 105mm f/2.8 G ED VR. On the other hand, there are differences concering autofocus as well: although the Tamron can’t be said to be slow focusing, it hunts a lot in low light and even under good light if the contrast is a little low. Switching the AF limit on can speed up the back-and-forth hunt but not by as much as one could expect. Another issue with the Tamron is the permanent out of focus when trying to focusing on targets at similar distances, but only at medium to high distance targets. This problem arises when I use the lens to shoot portraits or street photography at maximum aperture, whenever I do a sequence of shots on the same target, and the result is: the first shot generally is in focus, but all the next shots will be out of focus, and the only way to work around this is to stop down to f/4 at least. Mind you that the problem exists only at maximum aperture, so I’m not sure if it’s only with my copy or a known issue with this lens. Because of these observations, and not surprisingly, the Micro-Nikkor trumps the Tamron in the AF department in every way possible. It’s simply in another class of its own, as the price suggests.
In summary, handling this lens is not really very pleasant, and unfortunately it may be not the ideal tool for some areas of macro photography either.


Resolution

For the resolution test I shot the 5 Euro bill as usual. Focus was achieved using Live View to avoid auto-focus imprecisions and to compensate for possible field curvature issues.
The first column shows a crop of the image center, the DX corner crop is on the second column and the third column shows a crop of the FX extreme corner. Each row represents an aperture setting, from maximum to f/22 in full stops. I opted to show crops only until f/22, since closing the aperture even more results in more and more diffraction. Also, the target was at a such distance from the camera that the maximum effective aperture was f/3.0 instead of f/2.8 – the camera always reports the effective aperture. Here are the results:

It shows the incredible amount of detail that the lens is capable to deliver in short distance department, right from the maximum aperture from corner to corner. It appears to be even a hair sharper than the 105mm VR in the FX corner, which loses a bit of “bite” there at maximum aperture.
The Tamron is a winner here, and for the price it’s fantastic! But what about long distance targets? Let’s see below:

There’s a loss of sharpness at f/2.8 in the FX corner, just like happened with the 105mm VR, and there’s also some vignetting affecting the exposure. At f/4 the problem is solved.
This indicates that the Tamron could be a superb lens for every kind of shooting, and I can only fault it when it misses focus wide open when using AF.


Distortion

I tested distortion with the usual brick wall shot:

Distortion is extremely low but has a wavy characteristic that it’s a little hard to describe and depict here in this image, but it’s virtually never a problem in the field and barely visible even in this kind of targets.


Vignetting

In this test, I shot a white wall at home using tungsten white balance and set exposure manually:

There’s a lot of vignetting at the widest aperture that affects the entire picture, but decreases substantially by stopping down. At f/5.6 it’s completely gone.


Chromatic aberrations

For this test I shot a car from above on a bright sunny day early in the afternoon:

This lens is Di (Digitally Integrated) which means it should take care of chromatic aberrations better than its antecessors, but although this test doesn’t show any vestiges, I know from experience that it shows a some wide open especially in metal surfaces under bright light. But at f/4 they’re completely gone.


Coma

Coma was tested using a LED source, at home in a dark room. I put the light source at the center (first column), corner and extreme corner of the frame (second and third columns, respectively), at maximum aperture and stopped down.

Coma is visible at full aperture already in the DX corner in these type of test shots, but it’s rarely seen in practice. The points of light are always pretty much circular everywhere and only the small less bright areas around the center are kind of oval, but hardly distracting. By f/4 the “problem” is not visible anymore. All in all, it’s a more than good performance here.


Flare

I shot some foliage in my garden against the sun, to see if I could see any flare vestiges. I shot directly against the sun, then placed the sun at one corner and finally made some shots with the sun just outside the frame.

Shot directly against the sun.
Shot with the sun placed at one corner of the frame.
Shot with the sun just outside the frame.

The lack of special elements in the design of the lens also shows here in this test. Shooting against the sun doesn’t result in a real loss of contrast, but the problem is the high amount of internal reflections when the sun is placed in the corner of the frame or just outside. Of course, macro photography and sun in the frame usually don’t go together, but it’s a real problem when shooting landscapes of any things other than macro with the sun there or nearby. It’s also susceptive to flare at night with street lamps, so watch out.


Bokeh

The lens has 9 straight aperture blades, and therefore it’s expected to see circular out of focus highlights at maximum aperture and polygonal shapes when stopping down. I took a defocused picture at the widest aperture of the city lights and got crops of the center, corner and extreme corners. The test was repeated for the subsequent two stops. Due to the focus distance I selected for this test, the maximum effective aperture was reduced to f/3.5.

As expected, bokeh at the widest aperture is generally pleasant but not as creamy as with the 105mm VR because of the visible edges. This is especially noticeable when shooting against foliage and the sun, where the multiple out of focus highlights and their edges can be a little distracting. At f/4 the difference is the begin of the polygonalization of the highlights, which are clearly seen at f/5.6.
Of course, overall it’s still pretty good.


Macro/Close-up

With the focus limiter switched on, the Nikkor was capable to focus as close as 43.2 centimeters from the sensor plane, or 27 centimeters from the front element. When the focus limiter is switched off, the focus distance drops down to 29 centimeters from the sensor plane or 9.5 centimeters from the front element, to a maximum magnification ratio of a real 1:1.
I shot an 1 Euro coin and this is what to expect at the minimum focus distance:


Focus limiter on


Focus limiter off


Summary

Build quality 6 Almost entirely plastic but the build is tight
Handling 4 Lens extension while close focusing and clutch mechanism may bother some, repetitive misfocus wide open is bad
Resolution 10 Superb sharpness across the frame at all apertures, it hardly gets better than this
Distortion 10 Absent in all practical shooting
Vignetting 9 Very strong at full aperture, but improves a lot by stopping down
Chromatic aberrations 8 Easily seen in metal surfaces in bright light at maximum aperture
Coma 10 Pretty much unnoticeable in practice
Flare 4 Contrast almost doesn’t suffer but may produce harsh internal reflections
Bokeh 6 Very soft but highlight edges can be distracting at times, gets polygonal too early by stopping down
Overall 78% A good lens for getting started with macro photography, and delivers beautiful results in many other areas


Samples

Here are some samples of pictures I made with this lens. Settings: native JPEG, picture control set to Landscape mode, no post-processing applied except reducing to 600 pixel width.

90mm, f/11, 1/125s, ISO 400, external flash
90mm, f/8, 1/200s, ISO 200
90mm, f/2.8, 1/200s, ISO 200
90mm, f/20, 1/80s, ISO 500, external flash
90mm, f/4.5, 1/10s, ISO 200, tripod
90mm, f/3.0, 1/500s, ISO 200

Voigtländer Color-Skopar 28mm f/2.8 SL II Aspherical : Review


Introduction

I’m back again with my newest lens review, this time of the Voigtländer Color-Skopar 28mm f/2.8 SL II Aspherical lens for the Nikon mount, tested as usual on a Nikon D700 full-frame DSLR. Just like its twin sister lens, the Voigtländer Color-Skopar 20mm f/3.5 SL II Aspherical, this is a pancake lens with a slim, elegant design, and makes a very portable and lightweight combo with the D700 as shown below:

Just like its twin, the retail price of €449 doesn’t seem cheap at first, but besides being an all-metal lens it has a relatively fast maximum aperture, especially for a pancake design. Pancake lenses are particulary rare in Nikon mount, and this lens may be like a God’s send for those who appreciate shooting with this type of lenses. This lens looks so similar to the 20mm f/3.5 that I will base my description of the 28mm on that review I wrote some time ago.
The lens includes an aspherical element to reduce aberrations and distortions. Being a manual focus lens, it can be a bit limiting for you depending on your shooting style, but that’s also true for every manual focus lens such as Nikkor AI or Zeiss ZF lenses. The fact that the lens has a wide focal length reduces this limitation a bit, being very easy to focus anywhere you want.
I’ve been shooting with this lens for some months now and I have better feelings with it than those I had with the 20mm which performed badly on the full-frame sensor, and now I trust this lens for every occasion where I need to travel light. The 20mm was a sweet lens on DX, it had the perfect focal length to cover a great range of scenarios, and I used it mainly to shoot street scenes, landscapes, monuments, inside museums and churches, cafés, pretty much everything I wanted. But when I migrated to FX I wasn’t that enthusiastic with the focal length, and worse, it performed badly in the corners. So I looked forward to Cosina to release a new pancake lens with a equivalent focal length on FX, and voilà, it seemed that Cosina was listening to me and didn’t took much time for this 28mm to arrive. I was also thrilled that they increased the maximum aperture to f/2.8.
Last time I used this lens was on my summer vacations in southern Spain and I loved it, it performed so much better than the 20mm, that I was shooting at f/4 all the time in churches and museums with no regrets. The maximum aperture was also excellent when I needed to do a close up or a quick portrait, since the quality at the center of the frame is very high. It looked like I was shooting again with the 20mm on my old D300, but seemed even better! This lens was also good to shoot unnoticed among the multitude of tourists that visited the region at that time of the year, but also got a few smiles from other photographers that were curious about its diminute size. But I will describe its performance in detail below.


Technical Specifications

Focal length 28 mm
Biggest aperture f/2.8
Smallest aperture f/22
Field of vision 74.8 degrees (on FX)
Weight 180 g
Dimensions 25 x 62 mm
Optical construction 7 elements in 6 groups, 1 aspherical element
Aperture blades 9, straight
Filter diameter 52 mm
Minimal focus distance 22 cm
Hood LH-28N (optional)
Mount AI-S, CPU integrated


Mechanical Characteristics

Zoom ring n/a
Focus ring All-metal, with infinity stop
Focus throw 160 degrees
Focus motor n/a
Optical stabilizer n/a
Front element rotation while zooming n/a
Front element rotation while focusing No
Lens extension while zooming n/a
Lens extension while focusing Yes, 6 mm
Internal focusing No


Handling

Handling this lens is every bit the same as handling the 20mm, so I will basically copy-paste the same description.
The Voigtländer is a very compact lens as you can see above in the pictures, and being all-metal it’s very well built and surprisingly heavy for such size. Being a manual focus lens, probably the most important aspect is how the focus ring feels on your fingers, and it does indeed feel great. The focus ring is perfectly damped and focusing is very smooth. I have been shooting with some manual focus lenses (AI-S and Samyang) but focusing with this lens feels even better than those. It feels smooth and has a long throw of about 160 degrees for precise focusing. This is mostly useful in near distance, of course, because from about 10 meters to infinity there’s almost no need to move the focus ring again. The lens stops focusing on infinity which is great news for astrophotographers.
The less positive things about this lens is that it doesn’t have a lock on the aperture ring, therefore watch out and keep the aperture ring at f/22, otherwise you’ll get a fEE error on your LCD display. Because the lens is so tiny and there’s almost no space between the focus and aperture rings, attaching and dettaching the lens from the camera can be a bit difficult, but with a firm grip on both focus and aperture ring, along with the fixed 3mm spacer in between, it can be done but be warned that the lens gets almost glued to the camera.
Anyway, my only thumbs down goes (once again) to the front cap which gets off so easily with just a small touch. I instantly replaced that cap with a (heresy!) Canon one.


Resolution

For the resolution test I shot a near distance object and a far away building to find out if there were any visible differences in image quality between those distances. As usual, I focused with Live View and re-focused when moving the target to the corners.
The first column shows a crop of the image center, the corner crop is on the second column and the third column shows a crop of the extreme corner. Each row represents an aperture setting, from maximum to minimum in full stops. Here are the results:

The resolution at the center is already excellent straight from the maximum aperture, showing high contrast and clarity. This is always true until f/16, from which diffraction starts affecting the image and there’s a loss of contrast. The DX corner resolution is not so good wide open but is more than acceptable at f/4 and peaks at f/5.6, and keeps on high values until f/16. The extreme corners are a bit weaker at f/2.8 and appear affected greatly by vignetting, but after quick correction in post-processing it is visible that the problem is not that much the lack of sharpness. If DX corner sharpness improves greatly at f/4, the extreme corners will take one more stop to reach similar levels. But still, for very good sharpness across the frame it’s better to stop down to f/8, or f/11 if critical sharpness is needed on the extreme corners of FX.
This is great performance on very close targets, but what about “real world” where most of the time the target is at at least several meters away?

Although this target seems a bit flat in contrast, careful examination shows that there are no meaningful differences in resolution, being already excellent at f/2.8 in the center and very good in the DX corner, but not so good in the FX corner and still affected by heavy vignetting. The aperture of f/8 delivers very good resolution across the frame, just like we saw in the previous test.
In conclusion, this is a much more versatile lens than the 20mm with a broader range of useful apertures, which was to be expected from the start since 20mm is more difficult to produce than a 28mm lens, especially when we’re talking about pancake lenses, and also because the maximum aperture is higher.


Distortion

For the distortion test I shot some tiles:

The lens distortion is well controlled. There’s a mild amount of barrel distortion but generally it’s not noticeable on most shots, and it’s easy to correct in post on architecture photos if needed.


Vignetting

In this test I shot a white wall at home using tungsten white balance.

Just like it was detected previously on the resolution test, vignetting is very heavy wide open and improves significantly by f/4. From f/8 and down it generally doesn’t show up in images. Notice that this is mostly visible on shots of white walls like this. As a matter of fact, f/5.6 is already pretty much free from vignetting.


Chromatic aberrations

It was a clear bright sunny day with the sunlight reflecting on my car, and I took pictures of a area of high contrast:

The use of an aspherical element is an important feature in this lens for taking care of distortions and aberrations, and once again Cosina made an excellent job to keep aberration levels at negligible levels. From all the shots I made with this lens there wasn’t a single day I noticed any fringing issues. The lens deserves the maximum score here.


Coma

Coma is an important requirement in astrophotography and usually affects the corners of most lenses, especially wide-angle lenses and particularly old designs. I tested coma using a LED source of light in a dark room.
I put the light source at the center (first column), corner and extreme corner of the frame (second and third columns, respectively), at f/2.8 and f/4 (one stop down).

At f/2.8 the lens produces high amounts of coma even in the DX corner and much more so at the extreme corners, and doesn’t really get much better by stopping down. This is very unnattractive for astrophotographers, and if you are thinking of going out shoot the stars wide open, you better think twice or you might not like the results. This lens is awful in this regard, just like the 20mm was.


Flare

I shot several pictures on my backyard trying to get flare vestiges, and the result is as follows. I started to shoot directly against the sun, then placed the sun at one corner and then shot with the sun just outside the frame.

Shot directly against the sun
Shot with the sun placed at one corner of the frame
Shot with the sun just outside the frame

The Voigtländer has very good quality coatings inside and the results put it along with the best lenses I’ve used, with great flare resistance and keeping the overall contrast on high levels. There’s just a bit of veiling flare around the sun but no ghosting was detected on other parts of the frame. You just need to be a little cautious when shooting with the sun just outside the frame, and for this it might be a good idea to use the LH-28N lens hood available separately. Overall this is a very good result.


Bokeh

This is a wide angle lens and therefore the quality of the out of focus area of an image is not a primary aspect to consider, but still, since the lens has a relatively fast maximum aperture and 9 diafragm blades, one might expect acceptable renderings of out of focus highlighs. So lets see how the lens fares in this matter.
For this test I took a defocused picture at f/2.8 of the city lights and got crops of the center, corner and extreme corners. The test was repeated for f/4 and f/5.6.

The out of focus highlights are pretty circular at f/2.8 thanks to the 9 aperture blades, as expected, although the corners reveal the cats-eye distortion due to vignetting. Although circular, the highlights show pronounced edges and nervous interiors, but nevertheless it’s not distracting at this aperture when looking at the entire picture on my monitor. Closing down to f/4 reveals the outcome of the straight aperture blades and stays like this when closing down further. Generally, with this lens it’s better to shoot close-up subjects with the maximum aperture, where the center sharpness is already very good, to get the best out of focus renderings.


Macro/Close-up

The Voigtländer was capable to focus as close as 22 centimeters from the sensor plane, or about 15 centimeters from the front element. I shot an 1 Euro coin and this is what to expect at the minimum focus distance:


Summary

Build quality 9 A beautifully well crafted lens, but the front cap is rubbish
Handling 10 Buttery smooth focus ring and very solid, compact lens, with CPU contacts for the automated exposure modes
Resolution 8 Excellent center sharpness, very good corners and fairly good extreme corners at moderate apertures
Distortion 8 Well controlled barrel distortion for a wide-angle lens and easily correctable if needed
Vignetting 7 Very strong wide open but improves greatly stopped down
Chromatic aberrations 10 A winner just like its 20mm sister lens
Coma 6 Very strong in the corners wide open, gets just a little better by stopping down
Flare 9 Great performance, but be most careful when the sun is just outside the frame
Bokeh 5 Circular only at f/2.8 but very nervous and pronounced edges, and gets polygonal at smaller apertures
Overall 79% A good pancake lens with very much expanded range of useful apertures for FX


Samples

Here are a few pictures I took during my past holidays and in my backyard. Settings: native JPEG, picture control set to Landscape mode, no post-processing applied except reducing to 600 pixel width.

f/4, 1/25s, ISO 800
f/8, 1/400s, ISO 200
f/5.6, 1/160s, ISO 200
f/8, 1/80s, ISO 400
f/4, 1/640s, ISO 200
f/8, 1/320s, ISO 200