Tuesday, 26 February 2013

Using protective filters on lenses

One of item of controversy, is whether to use protective glass filters on lenses or not. Typically, people fall into one of two categories:

1) Those who think that having a clear glass filter on the front of the lens is crucial for protection of the front lens element: If you accidentally bang the lens into something, the filter will take the blow, protecting the much more expensive lens underneath it.

2) On the other hand, there are those who think that the filter does little to protect the front lens element: It is much less strong, since it is thinner than the front lens element of the lens is. Also, if it does break, the shards from the broken filter could further damage the lens. The naked lens might not have been damaged at all by a similar blow, according to this group. This group also tends to propose using a lens hood for protection, and to say that any filter put outside the lens will degrade the image quality.

Now, it is hard to say which group is correct. I'm certainly not going to bang my lenses into a sharp object to see what kind of damage it takes.

But one thing that can be tested is if the filter degrades the image quality. I tried to use a cheap filter previously, and found that it did degrade the image quality significantly in high contrast situations, at night. This time, I use a somewhat more expensive filter, which claims to be multicoated:

Hoya HMC multicoated slim frame 58mm

I used it on the Lumix X 12-35mm f/2.8 lens. Here is a picture taken at 35mm at night, both without and with the protective filter:

Without filterWith filter

The only difference I can see here, is that the full moon has moved between the shots. It appears that the filter does not add any extra flare, unlike my previous test with a cheap filter.

I can compare the sharpness by looking at 100% crops from both images:

This comparison shows that there is in fact some loss of sharpness when using the filter. The image is not as crisp on pixel level. However, you can ask yourself if you really need that much sharpness on pixel level.

Here is another test with the lens at 12mm:

Without filterWith filter

In this case, there is a very subtle difference in the flare: Slightly more flare when using the filter:

But beyond this rather subtle difference, there is no significant difference between the image taken with and without the filter.

In daylight, when the contrast is smaller, there tends to be even less impact of using filter.


Based on this and my previous test, I can conclude that using a cheap, non-coated filter can seriously degrade your images, especially in high contrast. However, a reasonably priced multicoated filter can be used with little risk of degrading the image quality. If using a filter like this gives you peace of mind in terms of avoiding damage, then you don't risk damaging the images significantly either.

You can buy a filter like this on Amazon.

Appendix: Technical details

I took the images with a Panasonic GH2. I would have preferred to use my GH3, however, it is still with Panasonic for a repair. After some weeks of use, the automatic switching between the LCD and EVF stopped working.

The camera was set on a tripod, with OIS turned off, the ISO set to 160 (base value). I refocused for each shot, also after mounting the filter for the second otherwise identical shot. I used the self timer to avoid camera shake. The aperture was set to f/4, and the shutter speeds were around 1-2s.

Wednesday, 20 February 2013

Sigma 19mm f/2.8 EX DN review

The Sigma 19mm f/2.8 EX DN was released at the same time as the Sigma 30mm f/2.8 EX DN. Together, these were the first Sigma lenses designed specifically for mirrorless camera systems, and were released for the Sony E-mount and Micro Four Thirds mount in 2012.

Both of these lenses have somewhat odd focal length for the Micro Four Thirds format, compared with conventional prime lenses. The 19mm lens has the same field of view as a 38mm lens on a traditional film camera, while the 30mm lens corresponds to 60mm. Both of these focal length equivalents are unusual.

However, the answer to these odd focal lengths lie in the fact that the lenses were designed for the APS-C format. With a smaller crop factor of 1.5x, the lenses correspond to the 28mm and 45mm, i.e., the classic wide angle and normal lens, respectively.

Also, keep in mind that while 19mm is an odd focal length for Micro Four Thirds, it is very close to 20mm, and the Lumix G 20mm f/1.7 is a very successful lens. They are both shown below:

Wednesday, 13 February 2013

Lens disappointment, Lumix G HD 14-140mm

Normally, I write the most about the lenses I like the most. I guess this is partly natural, to write about what is positive, and also reflecting my nature. But, this time, for a change, I write about a lens that has disappointed me: The Lumix G HD 14-140mm f/4-5.8.

To put it shortly, I think this lens has failed to live up to the promise: It is not very sharp, with the exception of the middle of the zoom range. Also, it does not focus faster than the other, and much cheaper, zoom lenses, nor does it operate less loudly, both of which were the claims of the "HD" branding.

The lens was originally launched in 2009, and for the first period only sold in a kit with the Panasonic GH1. It was marketed as a video optimized superzoom lens, and said to have a very quiet, accurate aperture mechanism, capable of near stepless changes. Some have said that the aperture can change in 1/6th stop steps.

It is fairly well known and agreed that the lens is dull in the long end. To further make this claim more credible, here is a comparion of the sharpness at 100mm, where significantly less expensive lenses come out much better. In this test, I am comparing the lens to the Lumix G 14-42mm f/3.5-5.6 and the Lumix G 14mm f/2.5 pancake lenses at 14mm, and conclude that both of the latter are significantly better. So in my experience, the 14-140mm lens only really shines in the middle of the zoom range.

When measuring the autofocus noise, the lens comes out at a fairly average level, comparable with all the other zoom lenses. Also, the aperture change noise was measured to be at a higher level than the other zoom lenses, not really any impressive. This suggests that the HD designation of the lens is rather worthless, as non-HD lenses perform as well, or even better.

The aperture range of the lens is also not so impressive, f/4-5.8. Competing superzoom lenses tend to have the aperture range f/3.5-5.6. Also, when looking at the aperture as a function of the focal length range, you'll see that it very quickly goes up to f/5.8:

Given these experiences, it is perhaps only fair that this lens is now rumoured to get a makeover. A picture of a new version of the lens has been seen. The new lens retains the same basic specifications, but is smaller, with a 58mm front lens thread. The original lens has a 62mm lens thread. The 58mm front lens thread is the same as is used on the Lumix X 12-35mm f/2.8 and Lumix X 35-100mm f/2.8 pro spec zooms.

Since making the original 14-140mm lens, Panasonic has come a long way. The new X branded f/2.8 zooms are very good, and I am sure that with this experience, Panasonic can design a much better 14-140mm lens.

Since writing this article, the new lens has been announced, and it has the more attractive aperture range of f/3.5-5.6, in additon to being smaller and lighter. I would definitively go for the new version, if I had the choice now.

Sunday, 10 February 2013

One picture, different times

I had the possibility to photograph a nice view early in the morning, and took two images spaced by 29 minutes, to see the difference. I used the Panasonic GH2, since my GH3 is still in for a warranty repair. I did not have a tripod, and handheld the camera with the Olympus 45mm f/1.8 lens.

The first picture was taken at 7:41 in the morning, 35 minutes before sunrise, at ISO 1600, f/2, 1/25s:

The second was taken at 8:10 in the morning, six minutes before sunrise, at ISO 160, f/2.5, 1/45s:

In practice, I took a series of four images in both cases, and then later chose the least blurry of them. That's how I handled the slow shutter speeds without a tripod.

I think most would agree that the first image is the most striking, at least at first glance. It has the deep blue tint associated with the early morning, and there is the contrast with the yellowish artificial lights.

To be able to photograph the first one without a tripod, I needed to push the ISO rather high, at ISO 1600. This is at the border of what the GH2 can handle, and the image quality suffers. Here is a comparison at pixel level between the images shot at ISO 1600 and ISO 160:

Note, though, that this difference is not purely due to ISO differences. The first image also has a slower shutter speed (1/25s versus 1/45s) and a larger aperture (f/2 versus f/2.5), both potentially making the image less sharp. The Olympus 45mm f/1.8 lens is quite sharp, but you'll usually notice that the sharpness increases quite a bit when stopping down from f/1.8 to, say, f/2.8. Despite this, I think the first image could well be printed fairly large without any problems.

The pictures above are from the out of camera (OOC) JPEGs. One could say that the first of the images is rather blue. However, this is not the GH2 auto white balance doing anything wrong, it was quite simply very blue this early in the morning. Opening the RAW image in Lightroom 4, it could be edited to something like this:

I think this edited version is probably a bit more striking.


After having looked at these images, I think it is quite impressive that it is possible to take a handheld image half an hour before sunrise, and even with a tele lens to boot. This shows how far the technological development has come.

Still, the image quality could be improved by using a tripod. If I had a tripod, I could have set the shutter speed to, e.g., 1 second, and used a significantly lower ISO, and a larger aperture for the best image quality.

Thursday, 7 February 2013

Bokeh comparison @ 20mm and 19mm

The Sigma 19mm f/2.8 EX DN lens can be an interesting alternative to the Lumix G 20mm f/1.7. It focuses significantly faster for both stills and movies, and while it is not as sharp as the 20mm lens, it is still sharp enough. It also comes at a much lower price.

In this article, I'm looking at the out of focus rendering of the lenses, the bokeh. I also threw the Lumix X 12-35mm f/2.8 into the comparison.

The comparison was done by focusing on a close object, the box at about 50cm distance in the left foreground:

Sigma 19mm @ f/2.8 Lumix G 20mm @ f/2.8 Lumix X 12-35mm @ f/2.8

Here are some 100% crops from the centre:

And from the extreme right border:

The Lumix G 20mm f/1.7 features asymmetrical out-of-focus rendering outside of the centre of the frame: The further from the centre, the more elliptical the bokeh highlights become. You can see the same phenomenon in this study. This property is also shared with the Lumix-Leica 45mm f/2.8 1:1 macro lens.

The Lumix X 12-35mm f/2.8 also has some minor disturbance in the border: The out-of-focus highlights are rendered somewhat non-circular. When stopping down, this becomes less of a problem. However, this is a very minor issue.

Overall, I think the Sigma 19mm f/2.8 EX DN probably has the most pleasing out of focus rendering in this high contrast scene.