Mirek’s Photo Blog

Adobe Labs released Lens Profile Creator utility that enables creation of custom lens profiles for Photoshop CS5 (ACR 6.1) and Lightroom 3.

This page contains profiles for some Zeiss ZE lenses for Canon EOS 5D II that I created using that tool. The profiles are free to use.

Currently there are profiles for the following lenses:

• Distagon 21mm ZE
• Distagon 28mm ZE
• Makro-Planar 50mm ZE
• Makro-Planar 100mm ZE

Click here to download

Examples

Chromatic aberration

Extreme corner shot with the 28/2, uncorrected.

Extreme corner shot with the 28/2, corrected with ACR 6.1

Distortion and vignetting

Distagon 21mm, uncorrected (click the image for larger picture)

Distagon 21mm, corrected in ACR 6.1. Exact same settings, but lens correction on.

Distagon 21 – distortion correction

Sharpening and micro contrast enhancement has been frequently used in film photography. This article briefly introduces similar digital and analog techniques.

Digital Unsharp Mask

Unsharp mask selectively increases contrast of edges and small objects by making dark areas of edges darker and light areas of edges lighter. Since this creates a sharpening effect, unsharp mask is frequently used to improve apparent sharpness in digital imaging. The following image shows enlarged detail of an edge with and without unsharp mask. The image consists of two gray boxes; the upper part is with USM applied.

If USM is used with narrower radius settings, it can effectively increase apparent sharpness. The following image shows effect of the unsharp mask on an image. The first character R is normal, the second is a copy of the first one with gaussian blur applied and the third one is a copy of the blurred one with unsharp mask applied.

If USM is used with wider radius settings, it can increase micro contrast.

Eberhard Effect

Exposed molecules of silver halides react during development with developing agent and convert the white silver halide into black silver and creates visible photographic image. This chemical reaction breaks up the developing agent and creates certain byproducts that are capable of retarding further chemical reaction – development. Eberhard found that if developer is left absolutely still, smaller details and edges are developed more rapidly than larger areas, because smaller areas and edges can be easily surrounded by fresh developer, while silver halide particles inside large areas are surrounded by exhausted developer that does not have the vigor of the fresh one. Eberhard effect can be observed in still developers and diluted developers, but exists in conventional development as well and is frequently used by photographers to increase micro contrast and apparent sharpness. The effect can be replicated by digital unsharp mask. An extreme detail boosting technique, which turns photograph into graphics is based on copying or enlarging the photographic image to a large format high contrast film and developing with emulsion soaked in developer pressed on a glass sheet so that there is only thin layer of developer that can’t move. The film is copied and developed this way several times. I applied USM  on black and white image in Photoshop four times to achieve visually identical effect:

As you can see, tonality is gone and the image is created only by local contrast.

Micro Contrast in Lenses

Camera lenses also exhibit micro contrast – or lack thereof. Lenses with good micro contrast  have good MTF at frequencies around 10lp/mm. There is some consensus that Zeiss lenses consistently excel in micro contrast.

The following two images are taken at the same day and hour, identically processed, but with two different lenses. First was taken with Canon EF 35/1.4L and second with Zeiss Distagon 28/2 ZE. Both are sharp lenses, but one has bad micro contrast and one has good. Click on the images to see larger version.

Useful links:

http://www.luminous-landscape.com/tutorials/understanding-series/lens-contrast.shtml

http://www.zeiss.com/C12567A8003B8B6F/EmbedTitelIntern/CLN_30_MTF_en/$File/CLN_MTF_Kurven_EN.pdf

There are many web sites where photographer can find tests and reviews of lenses. Some of these sites are listed below and this is no attempt to provide information about general performance of lenses. Instead, I am focusing on suitability of my lenses for baby photography.

General Thoughts

How I Work

I’m using full frame camera and the suitability of lenses described below is related to full frame cameras only. I never use flash and rely on fast lenses and high ISO performance of my camera.

Age

The requirements and usability of lenses change with the ability of the baby to move. When they start walking, the AF speed becomes more important, especially during the age when they can’t understand everything you say and any attempt to stop them or attract their attention only causes that they start running to you. This has been a challenge especially with my 85mm lens and to some extent with 50mm.

Focal Length

Most of my photographs are from indoors, a situation which usually calls for shorter fast lenses. For baby photography I have been using mostly 85mm, 50mm and 35mm focal lengths, in that particular order.

EF 17-40/4L

Suitable for snapshots where it is necessary to get very close or where you need to include lots of background. The minimum aperture of 4 is limiting for indoors shots, but ok during daylight. You will often get close enough for the child to be able to reach the front element of the lens. Since the lens hood is shallow and won’t prevent the baby from actually reaching the glass, an UV filter is handy. If you are considering purchasing a wide angle zoom lens and plan to use it for baby photography, think about 16-35/2.8 II, which is more practical for low light situations.

EF 35/1.4L

This 35mm lens is considered by some the best wide angle lens from Canon, some reviewers rate it higher than 35mm Summilux R from Leica, but I found achieving sharp looking shots more complicated than with other lenses. Wide open it has shallow depth of field, but does not isolate the main subject the same way as longer fast lenses. This may create tricky situations on pictures with multiple isolated subjects, where some of the photographed subjects are not perfectly sharp and the viewer percepts this as bad focus. This may and may not be a problem for your type of photography. This lens can be hand hold at 1/30s and at wide open allows taking pictures with virtually no light. This is great especially during the period when baby works in 3 hours cycle and lots of interesting photo opportunities (feeding) happen in the middle of the night.

EF 50/1.4

I found this lens practical especially in the hospital. It is relatively unobtrusive, light and universal enough to cover all the needs. This lens is fast, makes nice background blur and I would recommend this lens for baby photography to anyone on budget.

EF 85/1.2L II

This is my most used lens. 85/1.2 closest focusing distance works great for adult portraits, but babies are smaller and you won’t be able to get close enough to fill the frame with head and shoulders of several months old baby. When they grow and start to move, this lens’ AF system is not fast enough to track their movement. Nevertheless, the fast aperture and background blur made it my favorite baby lens. The 85/1.8 is faster and focuses from closer distance and should be a great alternative.

EF 135/2L

Greater shooting distance makes this focal length great for shooting unnoticed. Also great for details with blurred background. This lens is sharp wide open and despite the longer focal length can be used indoors. It’s sharpness and beautiful background blur make it tempting to use it wide open, but babies, especially those that can walk, may easily move after locking the focus and before the exposure and easily get out of the shallow depth of field. Using AI-servo, checking the shots for sharpness,multiple exposures and stopping down the lens will help mitigate the problem.

Other lenses

The only other lens from my collection that I used for baby photography was 180/3.5 macro. I used it for some detail shots but also for shots where the 135mm was short. This is a great lens and the pictures are fine, but as one would expect, this lens is absolutely impractical for macro shots of moving subjects. Shorter macro lens would serve that purpose much better.

Useful Links

http://www.fredmiranda.com/reviews/index.php?cat=45

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx

http://www.16-9.net/lens_tests

 

Update – Two Years Old

At two years, ability to track the child and quickly frame became more important than anything else.

EF 16-35/2.8 II

This lens proves to be very practical for action shots. It allows taking reasonable pictures even if the kids get very close to you – and this happens a lot, especially if you talk to them. It also allows reasonable framing without cropping in post processing when the kid or kids quickly move back and forth.

 

 

EF 35/1.4L

Still one of the favorite lenses for that purpose, especially in low light, like these pictures taken when my daughter befriended with our waitress in a restaurant and the other one from a hotel room.

 

 

EF 85/1.2L

Great for portraits. While lenses like the 16-35 allow the photographer to become part of the action, the 85 gives you certain distance and isolation allowing you to capture totally different moments.

 

B+W polarizers have great reputation and I always thought about them as the automatic selection for my lenses. However, I needed new slim  polarizer for my UWA lens and B+W did not work well for me this time, because the slim version was needed and slim B+W filters don’t have front thread, which I find annoying. I looked at competition and decided for Hoya HD. Now I have several brands of CPL filters, which is a great opportunity for some comparison testing.

The Contenders

• Hoya HD CIR-PL 82mm
• B+W Kaesemann Circular Polarizer 77mm
• B+W Kaesemann Circular Polarizer 72mm
• Soligor C-P.L. 58mm. This is a circular polarizer in aluminum ring, apparently with no coatings (not anymore, anyways). It is 15, maybe 20 years old and I could not find any specs

The Test Objective

This is not attempting to be a scientific test and is not conducted in laboratory conditions. Tested filters have different sizes, have different age, everything is tested on a camera with 50/1.4 lens with larger filters fitted with an adapter ring. No lens hoods were used. Transparency/density test was repeated twice, other tests were run only once. Areas of interest were:

• Density. How much light the different polarizers block
• Tint. What kind of tint the filter have, if any
• Strength of the the polarizing effect. Is weak polarizing effect the price for better transparency?
• Look of the green and blue polarized areas. Do the filters create equally pleasing greens and blues?
• Contrast. See how different filters impact contrast f no lens hood is used. Lens hood could not be used due to different sizes of filters used in the test

Testing And Results

Density

Density was evaluated by two methods:

1. With filters placed on a transparency light table, every filter placed at the same location of the table to eliminate unevenness. Power voltage fluctuation was around 0.5% during the test. The filter was shot with a macro lens with camera in manual setting. The images were brought to Photoshop and color balance was set to 5400k, which is the “paper” color temperature of the light table. Center of each filter image was then cut in Photoshop and placed in a composite image for visual comparison
2. Diffuse white wall was photographed with the filters on. Each filter was turned to make sure there is no influence of the polarizing angle. Two different walls were used and exposure readouts collected and deltas between no filter and filter on averaged.

Outcome of the first method. Comparison of filter density/transmittance. Four identical manual  exposures of 4 different filters on a light table. White balanced to the color temperature of the light table.

Outcome of the second method. Number of f-stops needed to achieve correct exposure  compared to lens with no filter.

Filter	Stops (approx.)
No filter	0
Hoya HD	1
B+W Kaesemann 1	1 3/4
B+W Kaesemann 2	1 3/4
Soligor	1 1/2

 

Transmittance calculated from method 2.

Tint

The outcome of the Transparency 1 test served as an evaluation target for tint. The following image is the same as above, but with increased saturation to show the color of the shift.

 

Strength Of The Polarizing Effect

Strength of the polarizing effect was evaluated from pictures of a plant with waxy leaves. The pictures show no significant difference between the images. 

Color of Polarized Sky And Foliage

Sky

This is a crop from the darkest corner taken with different filters. Camera exposure was set to auto, no individual corrections were made on the raws – all processed the same way.

The strength of all filters looks very similar. The yellowish tint of Hoya is clearly visible in the sky  (and clouds).

Green Foliage

This is another crop from the same field, showing green foliage. The same processing conditions as above.

The yellow tint does not seem to bother here.

The following composite shows the whole picture from which the above crops were taken. The white balance is equalized, this time.

 

Contrast

Contrast was evaluated on pictures taken without lens hood. The differences are subtle, Hoya HD seems to give the same or better contrast as the Kaesemanns, the Soligor shows discernibly lower contrast than the others. The Soligor has apparently  no coating and therefore this behavior is not a surprise.

Conclusion

The above mini test shows behavior of three different brands of filters and sample variation between two B+W polarizers. Let’s sumarize the major findings:

• The Soligor that has no coating shows discernibly lower contrast when used without lens hood. On the other hand, the other qualities are similar to much more expensive filters
• The B+W Kaesemann shows some sample variation, more than I would expect with premium filter. The polarizing effect and resulting color is about the same as the other filters
• The Hoya HD passes 3/4 EV more light than the B+W, which is nice. It is very thin, but still has a front thread, which is very nice. The polarizing effect and contrast seem to be on par or perhaps better than the B+W. The yellowish tint may require one extra step in post processing

The test does not say anything about sharpness, flare, quality of the mount, ease of maintenance or durability. Without these parameters I’ll avoid rating these filters. 

Canon 100/2.8L Macro is a lens with primary use in macro photography. This article is looking at how is it doing at longer distances to help decide suitability for landscape work.

Sharpness

Test Conditions

Test images were taken with Canon EOS 5D2, from tripod, using remote control and Live View (no mirror vibrations) on a rainy day. Contrast should not be directly compared due to possible variability of rain intensity. All shots were focused manually and individually with Live View at 10x magnification. Each f-stop was shot several times and sharpest sample selected for the final mosaics. All shots were taken from the same place, therefore shots from 85mm lens appear smallest and those from 135mm largest, sorry.

Postprocessing

The images are processed with ACR with default settings, i.e. no CA reduction, sharpening 25.

Center Sharpness

Click on the picture to see actual pixels.

The difference in contrast between the 100L and the other two lenses is most likely caused by heavier rain. The 135L is probably sharpest and the 100L least sharp, however the difference across the lenses does not seem to be significant, especially between the 100L and 85L. Note that the 100L has 4-stop image stabilizer, which will most likely reverse the order for handheld shots under 1/200s.

Corner Sharpness

Click on the picture to see actual pixels.

The 135L seems to be sharpest again, especially at 2.8 and the difference is bigger than in the center. The 100L suffers from chromatic aberration. The sharpness and CA can be easily fixed in ACR or Photoshop though. The following picture shows corner shot from the 100L @f2.8 with and without CA correction and some additional sharpening.

Samples

The following three samples were taken with 1/60s and image stabilizer. Click on the picture to see larger size.

The following image is a center crop with actual pixels.

Some images showing background and foreground blur.

To be continued…

Problem

There has been some discussion about noise level of underexposed shadows from Canon EOS 5D MkII at ISO 100 compared to other FF cameras. I usually don’t have to push the exposure and find the problem theoretical, but I found the issue intriguing and did some tests.

• I do not think it is a good idea to underexpose like that, I am merely looking at the behavior of the camera or the raw converter
• The noisy samples are 100% crops from 21MP sensor and I the impact on actual prints will be minimal
• It is possible that Canon or Adobe will be able to mitigate the problem in next update of Canon firmware or ACR update

Samples

Image straight out of the camera with default ACR settings, no corrections applied.

 

Here is the actual ACR settings from the Basic tab. In addition to that, I tweaked the Luminosity and Color Noise sliders so that the noise is minimized, but not so much as to introduce the compact camera look. I use custom camera profile based on the GMB checker. I also tried the canned Adobe profiles as well, they produce different levels of noise, some are higher than the examples here and some are lower.

 

100% crop from the above image. The image has been pushed by 1.5 EV. Given the heavy exposure correction, it looks decent to me.

 

Another 100% crop from the above image. The image has been pushed by 1.5EV. It is possible to tweak the curves and noise slider so that there is no noise on the roof or in the sky, but other areas of the image then start looking like a compact camera picture with removed noise.

 

Another 100% crop from the above image. Again, pushed by 1.5EV.  If more noise removal is performed, the rendering of the branches and edges looks like a compact camera shot.

I have been using the 17-40 for several years. I liked this lens a lot, especially its color and contrast. Unfortunately, I do not use flash and even for hand hold nature shots I frequently use polarizer and therefore I prefer faster lenses. As a result, the 17-40 usually stayed at home. I recently got the 16-35 and kept them both for few weeks to compare.

Some data from the Canon web site:

Focal Length & Maximum Aperture	17 – 40mm; 1:4	16-35mm f/2.8
Lens Construction	12 elements in 9 groups	16 elements in 12 groups
Diagonal Angle of View	104° – 57° 30′	108°10′-63°
Closest Focusing Distance	0.28m / 0.9 ft.	0.92 ft./0.28m
Filter Size	77mm	82mm, P=0.75mm/1 filter
Max. Diameter x Length, Weight	3.3″ x 3.8″, 1.1 lb. / 83.5 x 96.8mm, 500g	3.5 in. x 4.4 in./88.5mm x 111.6mm, 640g

The Testing

I compared both lenses at on 35, 24 and 17mm at f/4, the widest opening of the 17-40 and f/8, which should be close to optimal on both lenses. I also looked at performance of the 16-35 at f/2.8 and briefly compared both lenses to the EF 35/1.4L. I shot everything in raw on Canon 5D and processed in Adobe Camera Raw with sharpening turned off.

The Results

Color and Contrast

The color and contrast are excellent with both lenses. The 17-40 shows little bit more vignetting.

Sharpness – Pixel Peeping at 100%

Focal length	Center	Corners
17mm	The performance of both lenses is about the same	16-35 is significantly better
24mm	16-35 is slightly better	16-35 is slightly better
35mm	16-35 is significantly better. The difference disappears at about f/8	The performance of both lenses is about the same

 

Looking at the center at 100% magnification in Photoshop, the 16-35 performs better. The difference is most significant at 35mm, still visible at 24mm and almost non-existent at 17mm. The 16-35 performed almost as well as EF 35/1.4 at equal apertures.

Sharpness – Prints

Prints – this is where the rubber meets the road. With naked eye and 10″x15″ prints, the difference is only clearly visible at 35mm with f/4. At other f-stops and focal lengths the difference is negligible. 

Click on the following image to see full size image taken with the 16-35 @ f/5.6 and 26mm. Warning – very large file.

Further Info

http://www.fredmiranda.com/reviews/

http://www.16-9.net/lens_tests/

http://www.the-digital-picture.com/Reviews/Canon-Zoom-Lens-Reviews.aspx

Many articles and posts on the Internet claim or disprove advantage of small and medium film format over digital. The two media are very different and it is not completely fair to look at digital with having film as an etalon of the best achievable quality. Here is in a nutshell what I have learned in my almost three decades of film photography and several years of work with DSLR.

 

Digital gives more faithful reproduction of reality

Some color films are optimized for portraits, some are good for landscape and all have some characteristic presentation. Digital cameras don’t do that. Owner of the GMB Color Checker can see that in a simple test. With a calibrated camera, the colors from the digital system will be spot on, but on film some patches will be noticeably off, no matter how good is the process.

Images from digital cameras are easier to post-process

This allows more flexibility with digital images. Compared to film, digital files straight out of the camera can be flat looking, but they can be easily manipulated without any damage to the data. From one shot one can create Velvia looking picture as well as Portra looking one. Processing actual Velvia shots into good portraits is not an easy task.

Digital photography should be compared with color film photography and not B&W film photography

Some photographers like to argue that digital printing does not allow to create that "Baryta look". Some people say this is not true anymore, but  my point is that you should only compare digital with color film.

Resolution on digital cameras works differently than film

With film, the quality of the print slowly deteriorates with enlargement. The best print quality is achieved with contact print. The bigger the enlargement, the lesser the detail and sharpness. To the contrary, with very high quality lens, quality of digital print does not exhibit any deterioration with enlargement as long as the original image has enough pixels to print at around 300 dpi. The digital image will however start breaking up when the size reaches the limit when there is not enough pixels to print at 300 dpi. When the photographer starts "upsampling" the image, the quality goes down rapidly.

As a result, when  we compare a 300 dpi print from digital camera with medium film print (where the 300 dots per inch come from the camera and not from upsampling), the level of detail and sharpness will look very similar, but if we compare 20x enlargement from film with digital print of the same size, the film will look more detailed and more natural. This is apparently the main reason for disagreements whether film is better than digital or not.

Sharpening and micro contrast enhancement has been frequently used in film photography. This article briefly introduces comparable digital and analog techniques.

Digital Unsharp Mask

Unsharp mask selectively increases contrast of edges and small objects by making dark areas of edges darker and light areas of edges lighter. Since this creates a sharpening effect, unsharp mask is frequently used to improve apparent sharpness in digital imaging. The following image shows enlarged detail of an edge with and without unsharp mask. The image consists of two gray boxes; the upper part is with USM applied.

If USM is used with narrower radius settings, it can effectively increase apparent sharpness. The following image shows effect of the unsharp mask on an image. The first character R is normal, the second is a copy of the first one with gaussian blur applied and the third one is a copy of the blurred one with unsharp mask applied.

If USM is used with narrower radius settings, it can be used to increase micro contrast.

Eberhard Effect

Exposed molecules of silver halides react during development with developing agent and convert the white silver halide into black silver and creates visible photographic image. This chemical reaction breaks up the developing agent and creates certain byproducts that are capable of retarding further chemical reaction – development. Eberhard found that if developer is left absolutely still, smaller details and edges are developed more rapidly than larger areas, because smaller areas and edges can be easily surrounded by fresh developer, while silver halide particles inside large areas are surrounded by exhausted developer that does not have the vigor of the fresh one. Eberhard effect can be observed in still developers and diluted developers, but exists in conventional development as well and is frequently used by photographers to increase micro contrast and apparent sharpness.