Tuesday, 29 January 2013

New Panasonic kit zoom lens

Today, a new Panasonic Lumix kit zoom lens was announced. This doesn't look like a very interesting lens for seasoned Micro Four Thirds users, as it has the rather boring specifications 14-42mm f/3.5-5.6. Panasonic has already released three lenses with similar specifications, so why make one more?

Lumix G 14-42mm f/3.5-5.6 released in 2013

Here are two of the previous lenses:

Lumix LensG 14-45mmG 14-42mmX PZ 14-42mmG 14-42 II
Year released2008201020112013
Filter thread52mm52mm37mm46mm
Front lens diameter*45mm30mm21mm25mm

*) The front lens diameter is not stated in the lens specifications, but is based on my own measurements, supplemented with inspections of product photos.

I'm guessing that the new lens will replace the 2010 Lumix G 14-42mm f/3.5-5.6 kit zoom lens. It makes no sense to replace the Lumix X PZ 14-42mm f/3.5-5.6 pancake lens, since Panasonic needs this one to compete with a similar lens from the Sony NEX system, the Sony 16-50mm f/3.5-5.6 pancake zoom lens.

When looking at this table, it is quite clear why Panasonic is replacing the 2010 zoom lens. The new lens is smaller, and lighter. It probably uses a plastic lens mount, which I see as no problem for such a small lens.

The compactness is an important selling argument for Micro Four Thirds lenses. With smaller and smaller competitors, Panasonic must step up and make even smaller stuff, to stay competitive. The old kit zoom lens was not as compact, and hence, not attractive enough for the market.

The new lens appears to be the type which is the most compact at the centre of the zoom range, unlike the older Lumix G 14-42mm, but like APS-C tele zoom lenses.

Another observation is that the front lens element is smaller than that of the 2010 14-42mm lens, but larger than that of the pancake powerzoom 14-42mm lens. The pancake zoom lens achieves the small front lens element by sacrificing aperture speed in the middle of the zoom focal range. Here is a diagram illustrating the aperture as a function of the focal length for the two lenses:

The Lumix X PZ 45-175mm also achieves a smaller front lens element in the same way. Read about it here. With this in mind, I'm guessing the new lens will have an aperture range somewhere between the two lenses above. One could say that this is a form of cheating: Panasonic achieves a more compact lens by lessening the specifications in a way which is not visible in the lens name: The aperture end points are still f/3.5-5.6.

As for the finish of the new lens, it has the new glossy appearance. I'm not too happy with this, as I prefer lenses to be matte and non-obtrusive. But it appears that glossy is what the market demands at this time.

Other lens news

Incidentally, other lens news also fall into the same category: Remakes of existing lenses.

Olympus has launched an updated version of the Olympus M.Zuiko ED 75-300mm f4.8-6.7. The new version has a new design, better matching the Olympus OM-D E-M5 camera, and also a new lens coating. But the lens is essentially the same. The good news, though, is that the price has decreased significantly.

Olympus M.Zuiko ED 75-300mm f4.8-6.7 II Lens

Sigma has updated their Micro Four Thirds prime lenses. The new lenses appear to have exactly the same optical design, but have a new lens barrel design. In what appears to be a strange design choice, the ribbed plastic focus ring of the original lenses has been replaced with smooth metal rings. From an ergonomic point of view, it is hard to see how this is good news. I'm guessing this was done to make the lenses look more expensive, and to motivate a higher retail price. They are part of Sigma's "Art" line of lenses.

Sigma 19mm f/2.8 EX DN updated

Sigma 30mm f/2.8 EX DN updated

Sigma also launched a new lens, the Sigma 60mm f/2.8 EX DN. This lens is a completely new design, and unlike the Olympus lens with similar focal length and aperture, this one is not a macro lens:

Sigma 60mm f/2.8 EX DN

Sunday, 27 January 2013

Sigma 19mm autofocus comparison

The Sigma 19mm f/2.8 EX DN is perhaps a strange lens. In terms of specifications, it is quite close to the Lumix G 20mm f/1.7 pancake lens, except that it is slower, and less compact. So, why would anyone be interested in it?

Lately, it sells at a reasonable price, and besides, it features the normal internal focus system, while the Lumix G 20mm has an old style focus mechanism which moves all the lens elements, being slow and noisy.

I have found that the Sigma 19mm lens is slightly less sharp than the Lumix 20mm lens, at similar apertures. This is consistent with other people's findings.

With still images

What about the autofocus performance, then? Is it faster, as one could guess? Here is a comparison where I put both lenses on the Panasonic GH2, and focused down to around 0.3m:

The Sigma 19mm lens spends 0.28s focusing down to close to the minimum focus distance, while the Lumix 20mm lens needs 0.64s. So the Sigma lens is indeed faster, as we had expected in advance.

During video recording

Another important aspect is continuous autofocus during video. This is still the achilles' heel of the Micro Four Thirds system, and not even the most recent cameras do this well.

To test it, I mounted the Panasonic GH2 with the Lumix G 20mm f/1.7 lens to a wooden plank, and next to it, mounted the Panasonic GF3 camera with the Sigma 19mm f/2.8 lens. I used Manfrotto Superclamps to mount the cameras, just like when I tested the rolling shutter properties of the GH2 and GH3. I set both lenses to f/2.8 for similar depth of focus properties.

Of course, the 19mm lens has a wider field of view. But on the other hand, the GH2 has the multi aspect sensor, oversized sensor property, which gives relatively wider field of view in video compared with the non-multi aspect sensor GF3 camera. So the field of view is probably quite similar for the two cameras.

Here are the two video footages, combined to one clip. To see the focus performance, it is best to view this in 1080p, click on the youtube icon to access this possibility.

The results are not completely consistent, but I think the Sigma 19mm lens and GF3 combo keeps the focus better during the video. One could speculate that the GH2 probably has the better image processing technology, and hence, should have the advantage in terms of handling continuous autofocus. Despite the handicap of being combined with the basic GF3, the Sigma lens performs better.

It would have been better to mount the lenses to the same camera, of course, but I don't have two of the same camera model. Also, I would normally have used the newer GH3 camera, but my camera is sent back for repair. I hope I get it back soon, and that I don't have to wait for three months, like I had to when sending back the Lumix G 14-42mm lens for misaligned aperture diaphragm blades.


What this shows, is that the Sigma 19mm f/2.8 EX DN lens is better for video use, as long as you can live with the f/2.8 aperture. While it is not as sharp as the Lumix G 20mm lens, it is still very sharp, and certainly more than sharp enough for video use. The autofocus is also less audible.

Even if the Sigma 19mm lens is larger, it is still quite non-obtrusive, with a matte black finish.

The Lumix G 20mm f/1.7 pancake lens is a classic Micro Four Thirds lens, praised for the very good sharpness. However, people also find it annoying for the slow and noisy autofocus performance. People have also reported that the focus mechanism can be clogged with dust, due to the moving front elements. The Sigma lens has no moving lens elements on the outside (only internally), and is probably more solid in that way.

Given the low price the Sigma 19mm f/2.8 EX DN sells at, it can be considered a good value for money, and an interesting alternative to the Lumix G 20mm f/1.7.

I haven't tested the "brother" lens, the Sigma 30mm f/2.8 EX DN in the same way, but I would guess that it, too, performs well in terms of autofocus speed.

Friday, 25 January 2013

GH3 rolling shutter examples

The Panasonic GH3 has a very useful feature, the electronic shutter. One disadvantage of the function is that it suffers from quite severe rolling shutter artefacts. Read more about it here.

Here are some example images, taken with the electronic shutter, the mechanical shutter, and a still image from the video output:

Electronic shutter (f/1.7, 1/250s, ISO 200):

Mechanical shutter (f/1.7, 1/250s, ISO 200):

Video still image:

We see that the electronic shutter has a slower sequential readout than the video mode. This is unlike the Panasonic GH2, which used the video output for the electronic shutter mode, and they shared the same rolling shutter properties. But then again, you only got a 4MP image with the GH2 electronic shutter mode, and it was quite unsharp, probably scaled up from 1080 lines or thereabouts.

This will be my last Panasonic GH3 related article for a while, and the review I am writing will be delayed. The reason is that I have delivered the camera for a warranty repair: The automatic switching between the LCD and EVF suddenly stopped working.

My GH1 and GH2 have operated flawlessly for years. It is perhaps ironic that the GH3 was my first Panasonic camera to require a repair, since it is marketed as being extra solid and weather proof.

The last time I had a Panasonic item for repair, due to uneven aperture diaphragm blades, they kept the lens for three months. And when I did hear from them, they just gave me a brand new lens, which turned out to exhibit exactly the same problem.

I'm hoping they don't keep my GH3 for as long.

Wednesday, 23 January 2013

Sharpness comparison: 19mm vs 20mm

Sigma recently launched to Micro Four Thirds lenses, the Sigma 19mm f/2.8 EX DN, and the Sigma 30mm f/2.8 EX DN.

I have previously been quite happy with the sharpness of the 30mm lens, so I am curious to see how the 19mm lens performs. It makes sense to compare it with the Lumix G 20mm f/1.7, with a similar focal length.

The test images were taken with the GH3 camera at ISO 200, on a tripod, and with self timer. The focus was set on the centre of the frame.

Lumix G 20mm @ f/2.8Sigma 19mm @ f/2.8

Here are 100% crops from the centre of the image frame at various apertures:

And from the lower left corner:


We see that the Lumix G 20mm f/1.7 has an impressive level of sharpness, even wide open. The Sigma 19mm, on the other hand, needs a bit of stopping down before reaching the same level of sharpness. At f/2.8, the 19mm lens is a tad bit dull, even in the centre of the image frame.

This finding is consistent with other tests I have seen. Generally, it is observed that the Sigma 19mm lens is not the sharpest at f/2.8, and improves when stopped down to f/4 and f/5.6. Stopping down beyond f/5.6 does generally not add anything to the overall sharpness, but does give you more depth of focus (DoF). If you need a deep DoF, it may still be wise to stop down to f/8 or even further, but this will give you slightly worse sharpness at pixel level.

Still, it makes sense to say that both lenses are sharp. Comparing the Sigma lens with the Lumix G 20mm f/1.7 is perhaps a bit unfair, since the latter is known to be very sharp. For most uses, the Sigma lens is certainly sharp enough.

On the positive side, the Sigma 19mm f/2.8 EX DN appears to handle flare a tad bit better. There is less flare effects around the strong highlights.

My experience also confirms that the Sigma 19mm f/2.8 EX DN has a much better autofocus performance. The focus operation is also less noisy.

Sunday, 20 January 2013

Lumix 20mm lens and banding with the GH3

Recently, there has been a lot of talk about banding at high ISO with the Olympus OM-D E-M5 and the Panasonic GH3 in combination with one single lens, the Lumix G 20mm f/1.7. This is in fact also mentioned in the GH3 manual, "Stripes may appear in High ISO sensitivity [with the Lumix G 20mm]."

To test this, I took the same image using the Lumix G 20mm f/1.7 and Lumix X PZ 14-42mm f/3.5-5.6 @ 20mm. Both lenses were used on the GH3 camera, at f/5.6, with ISO ranging from 200 to 12800.

Lumix G 20mm f/1.7 pancakeLumix X PZ 14-42mm @ 20mm

To investigate the banding problem, I have made 100% crops from the images at the ISO values 200-12800. Click to enlarge the image:

Based on these images, it is not easy to find any systematic difference between the Lumix G 20mm and Lumix X PZ 14-42mm lenses.

According to other people's findings, this problem can be seen in the shadows at high ISO. So, here are two more crops at ISO 12800, f/5.6, 1/80s:

Perhaps one can say that there is some more systematic banding with the 20mm lens, but it is certainly very hard to see. I don't think there are any significant stripes in the left image.

I tried a second shoot at ISO 6400, this time with the Lumix G 14-42mm f/3.5-5.6 as the alternative lens. The exposure parameters are: 1/60s, f/5.6.

Lumix G 20mm f/1.7 pancakeLumix G 14-42mm @ 20mm

And a comparison at 100%:

Again, I don't see any problems with banding when using the Lumix G 20mm f/1.7 lens.

Finally, here is a comparison with the Sigma 19mm f/2.8 EX DN lens, with a similar focal length:

Lumix G 20mm f/1.7Sigma 19mm f/2.8 EX DN
@ ISO 6400, f/5.6, 1/6s@ ISO 6400, f/5.6, 1/6s

And here are 100% crops at both ISO 6400 and 12800:

The images are a bit dark, so I tried to use the curves tool to bring out more details:

In terms of banding or stripes, I am not able to see any significant difference between the two lenses. I can't see any significant banding when using the Lumix G 20mm f/1.7 lens.


Here is an example video using the Lumix G 20mm lens on the GH3 at ISO 3200. I don't see any banding issues here:

Source of the problem

There has been some discussion about why banding problems were experienced with the Lumix G 20mm f/1.7 lens exclusively. Some have suggested it may be due to the focus construction of the lens. It is one of the very few Micro Four Thirds lenses which has a traditional focus mechanism, in which all the lenses move back and forth. This takes a stronger motor, and is slower and more noisy than internal focusing. The lens also has a large spiral spring in the focus assembly, and a lot of people have suggested that this might be the source of the problem.

Since the banding issue has been seen on the Olympus OM-D EM-5 and Panasonic GH3, some think this is a sign that their sensors are related, and share vital parts.

Also, it should be noted that if you use the electronic shutter option with the Panasonic GH3 in combination with fluorescent light (indoor), you may get banding with certain shutter speeds. There is an easy to understand explanation for this: Fluorescent light flickers with a frequency of 100Hz or 120Hz (depending on the country), and the electronic shutter reads out the imaging sensor sequentially during about 1/10s. Hence, a shutter speed faster than these shutter speed will give you significant banding. Read more about this here.


In my tests, I have not been able to find any significant negative effects due to banding with the Lumix G 20mm f/1.7 lens. My conclusion is that it is safe to use this lens, also at high ISO.

On the other hand, I have seen people posting example images where the banding problem can be seen, so I don't doubt that it can be a problem.

Another conclusion is that the ISO 12800 image quality of the GH3 is quite good. There is a lot of noise, of course, but still an amazing level of details. For web use, the images are usable.

Sunday, 13 January 2013

Sales statistics from Japan

BCN Ranking publishes sales statistics for various categories of cameras in Japan. Last year, I made a summary of the system camera category, which includes both mirrorless sytems and DSLRs. We saw that Micro Four Thirds had reached a healthy 19% of this category by the end of 2011:

For the 2012 statistics, BCN Ranking has split these into two categories: DSLRs and mirrorless. This makes sense, of course, but the downside is that I cannot continue my time series of statistics. BCN only give the percentage of sales, not the absolute number, hence I cannot recombine the two categories.

Anyway, let's see what the 2012 statistics for the mirrorless category looks like:

And when summing up per camera system:

Just like the previous statistics, these are based on the twenty most sold camera models. This means that you cannot expect the percentages to add up to 100%. There is still another 9% missing. In the missing category, I would expect that you find cameras like the Panasonic G5 and GH2, among others.

Micro Four Thirds has a very healthy lead in the mirrorless camera category, at 50% market share. I would attribute this to the fact that they launched their system early, and have a lot of quality lenses out.

One could expect to find the Samsung NX system on the list. On the other hand, it is fair to say that there is some rivalry between Samsung, which is a Korean brand, and the other manufacturers on the list, which are Japanese. While Samsung does not sell well in Japan, I think they do better in other markets, like Europe.

The Nikon 1 system is already at a good market share of 13%. They have a very strong brand in Japan, and have invested heavily in marketing. The volume models, the J series, look cute, with various colour schemes, and a clean design with geometric shapes.

The Pentax Q system er quite odd, in my opinion, with a very small sensor size. At this sensor size, I think most people are better off buying a premium compact camera, like the Panasonic LX7 or Olympus XZ-2.

Pentax K-01

The Pentax K-01 is also an odd machine. It shares the lens mount with the Pentax DSLR cameras. Hence, it is a mirrorless camera with the register distance of an SLR. While sharing the lens mount has some advantages, it also means that there is little size advantage to the mirrorless system. Because the camera needs to be as large (except for the pentaprism), and the lenses cannot benefit from a shorter register distance.

Monday, 7 January 2013

From the competition

While I use the Micro Four Thirds system myself, there is no denying that I watch what is happening in the competition. Here are some of my personal reflections on the market.

Premium compacts

Back in the mid 2000's, Panasonic ruled the premium compact segment with their LX series, along with the Leica branded versions of the same cameras. Other brands also wanted a part of this cake. Canon re-launched their S-series, with S90, S95, and so on.

Recently, Samsung broke the f/1.4 barrier with their EX2F premium compact with a 1/1.7" sensor. However, it was not long until Panasonic trumped this with LX7, with an even longer zoom range, and a larger aperture in the long end to boot. How could Panasonic pull this off, given that both use a 1/1.7" sensor?

Samsung EX2FPanasonic LX7

One explanation is that while both the Samsung EX2F and the Panasonic LX7 uses the same sensor size, their lenses do not project the same image circles. The LX7 has a multi aspect feature, which means that the corners of the sensor are never used: The lens projects an image circle smaller than the sensor diagonal. And with a smaller image circle, it is easier to design a lens with a longer zoom range, and a larger aperture.

Recently, Pentax announced the MX-1, which is similarly specified as the EX2F and the LX7, although the lens starts at 28mm equivalent at f/1.8, both less impressive. So how can it compete against the Panasonic LX7 and Sony RX100, the latter with an 1" sensor? Do Pentax expect that the nostalgic retro design will do it?

The Pentax MX-1 also shares the lens and sensor specifications with Olympus XZ-2, and they are probably based on the same components, but with different body styling.

Pentax MX-1Sony RX100


When the Panasonic GH1 was introduced back in 2009, it was the first consumer interchangeable lens system camera which could autofocus during video recording. Canon has noted the importance of being able to autofocus effectively during video recording, and launched a new series of lenses with STM motors, optimized for silent focusing during video.

Also, the new consumer DSLR, the 650D features an image sensor including photosites capable of phase detection autofocus (PDAF), which has the potential of improving the autofocus performance when the mirror is up, e.g., for video recording.

In terms of mirrorless cameras, Canon first launched the G1X, a large sensor compact with a non-interchange zoom lens. Based on the sensor specifications, one can speculate that it has the same sensor as the Panasonic GH1.

Now that Sony launched the RX100, with a slightly smaller sensor, but a much faster lens is a far more compact package, I'd say the Canon G1X is dead in the water. The Sony RX100 appears to be a more sensible alternative.

When Canon did launch the Canon EOS M camera, their interchangeable lens mirrorless camera system, it was surprising to see that they went for the APS-C sensor format, since the G1X had a smaller sensor.

Canon EOS M

The Canon EOS M also uses the on-sensor PDAF and STM lens technology. However, when testing it, I was not very impressed with the autofocus speed.

One advantage with this camera, over similar mirrorless cameras from other manufacturers, is that you can use any Canon EF lens with the "Canon Mount adapter EF-EOS M". Autofocus, aperture handling, and image stabilization works just fine when you put the adapter between the lens and the camera, even for old Canon EF lenses from the 1980's. You also get EXIF data in your images. The autofocus cannot be expected to be as fast as when using the Canon EF lenses on a Canon DSLR camera, though.


Sony was rather early into the mirrorless camera market, with their NEX series. The first batch of cameras were rather strange: With a glossy metal surface, no built in flash, a minimum of buttons, and a weird flash connector. One could almost think that they were designed to cause confusion in the market, rather than be actual products.

But the recent Sony NEX 6 changes the picture for me: Here is a camera that looks and feels like a camera, with a good grip, more buttons and wheels, a proper flash connector socket. And there is even a viewfinder in the corner, to keep those who like the rangefinder style happy:

Sony NEX 6

With a collapsible lens as well, this is a very powerful package.


Just like Sony, Samsung were quite early in the mirrorless camera market. So far, I think their cameras have appeared a bit immature. However, they have churned out an impressive variety of lenses. Their most recent is a 3D capable 45mm f/1.8:

Samsung 45mm f/1.8 3D

Contrary to the Lumix 12.5mm f/12 3D lens, the Samsung 45mm only has one lens unit. It achieves 3D recording during video by masking off the left/right half of the aperture for alternate image frames. This making the camera see the subject from slightly different perspectives on alternate frames, suitable for composing a 3D video.

Compared with the Lumix 12.5mm f/12 3D lens, it has a proper aperture mechanism, and an actual focus mechanism. This makes the lens much more useful. However, due to the strange but innovative 3D system, the 3D base is still very small, probably around 1cm, just like the Lumix lens. This means that the 3D effect will still be fairly limited. Unlike the Lumix lens, though, this lens also makes good sense for ordinary 2D images. The 3D part may be mostly a gimmick, but this is still quite exciting.


Sigma pioneered the large sensor compact camera segment with the Sigma DP series with Foveon sensors. While this was an innovative piece of technology, the cameras never seemed to deliver in terms of handling, image quality, or value for money.

The Fujifilm X100 got this segment rolling when released in 2010. While the responsiveness and autofocus speed was not stellar, the retro design made up for it, and it brought the large sensor compact to the masses. Well, to a select part of the masses, anyway.

In the very start of 2013, the updated version Fujifilm X100s gets announced. While the exterior is similar, it updates the sensor resolution, and not least, it adds PDAF to the imaging sensor, for better autofocus performance.

Fujifilm X100s

Given that Fujifilm were early users of the PDAF on imaging sensor technology, one could say that it was about time they put this to use in their cash cow camera system.


Nikon, together with Canon, forms the Big Two group, in terms of serious DSLR systems. They launched their Nikon 1 mirrorless system fairly early. In the beginning, it did not look like much, with a handful of standard zoom lenses, and a slow wide angle prime.

At this time, though, they have added a wide angle zoom lens 6.7-13mm, and a 18.5mm f/1.8 fast portrait lens, among others. This seems like a good start for the lens lineup.

Their cameras have improves as well. The first wave of cameras did not look convincing in terms of ergonomics. But the Nikon 1 V2 looks better, with a sensible grip and viewfinder

Nikon 1 V2

Right from the start, the Nikon 1 cameras had an electronic shutter feature. In fact, only the V1 and V2 have a mechanical shutter at all. I have not heard much about rolling shutter artefacts, so I am guessing that Nikon have implemented a faster rolling shutter sequential sensor readout than the Panasonic GH3, which suffers from severe rolling shutter with electronic shutter.

The Nikon lineup also contains a compact standard zoom lens at 30-110mm, which receives good reviews. When combined with the PDAF capable cameras, I'm sure this is an interesting combination to bring for sports photography. The lack of PDAF makes photographing moving objects quite challenging with Micro Four Thirds.

Nikon had a big volume success with the D40 in 2006. It was a small, light DSLR. Given the market standard at the time, it came with a fairly low 6 mega pixel count, and was renowned for the good high ISO image quality. One could say that the first wave of Nikon 1 cameras copy this strategy. Again, we are seeing a lower than standard mega pixel count. While most of the competition are showing 16MP and more, the Nikon 1 J1 and Nikon 1 V1 only came with 10MP.


With the Leica M9, they have a monopoly: An interchangeable compact camera with a fullformat 36mm x 24mm sensor. No other manufacturer can match that, at the moment.

Still, the competition is getting closer. Sony RX1 features a fullframe sensor, but with a fixed 35mm f/2 lens. It's probably just a matter of time before somebody launch a mirrorless system with a fullframe sensor and autofocus.

Anticipating this competition, Leica have launched their M-E camera, a lower cost alternative. And for the high end, the Leica M has got a 24MP sensor, and even live view and video, a first for the Leica M series.

Leica M

Beyond making their own cameras, Leica also lend their name out to Panasonic. Many Panasonic cameras have a Leica-branded lens, and also the Panasonic-Leica 45mm f/2.8 1:1 macro lens and Panasonic-Leica 25mm f/1.4 Micro Four Thirds lenses have the Leica co-branding.

One could guess that this is a mutually benefiting relationship between them: Panasonic gets the credibility and marketing benefits from using the Leica name, and Leica gets the additional volume and diversity from rebranding Panasonic cameras like the LX series of high end compact cameras.

Nikon and Canon don't need this type of rebranding. Their names are credible enough even without them. On the other hand, we have Sony using the Zeiss name, and Samsung using the Schneider-Kreuznach name on their lenses.

Panasonic started this back in the Four Thirds day, with their premium lenses being Lecia branded. Now, though, their self confidence appears to have increased, as they have released premium lenses without the Leica name, e.g., the Lumix X 12-35mm f/2.8 and Lumix X 35-100mm f/2.8.

While this strategy brings cash to Leica, no doubt, it is also a dangerous route. In the longer run, it risks lowering the perceived value of the Leica brand. Other premium brands have done the same, e.g., Ferrari, which have sold their name for use on toys, clothing, and even computer accessories. Rolex, on the other hand, have never licensed their name for use on other objects than their own watches.

Sunday, 6 January 2013

GH2 vs GH3 rolling shutter evaluation

The Panasonic GH1 and GH2 had pretty much the same rolling shutter properties. I have previously examined the rolling shutter artefacts of the GH3, compared with the GH2, and found that the GH3 has somewhat less artefacts. But my test was based on a rotating propeller setup, which is not so realistic.

Rolling shutter artefacts are typically identified when panning quickly during video recording. This can lead to "wobbly" effects, square objects can be seen to lean towards one side.

To compare these artefacts again between the GH2 and GH3, I mounted both cameras to a piece of wood:

I used Manfrotto Superclamps with ball heads to fix the cameras to the piece of wood. These are quite useful.

On the GH3, I put the Lumix G 14mm f/2.5, and on the GH2, I put the Lumix G 14-42mm f/3.5-5.6 kit zoom lens.

The latter zoom lens, used on the GH2, was not set to 14mm, as you might expect, but rather to 16mm. The reason for this is the lack of the multi aspect sensor on the GH3. So to match the video field of view of the GH3 at 14mm, the GH2 zoom lens must be set to around 15-16mm. Read more about the multi aspect sensor here, and also how the lack of this feature affects the GH3.

After having mounted both cameras to the plank, I set them to ISO 400, f/5, and recorded video as I panned them simultaneously in front of some buildings. Here is the result, where I have used the left part of the frame from both cameras. Note that due to the wide angle used, there is a bit of perspective skewness as well to the video:

We see that when panning quickly, the buildings appear to lean a bit to the side. Here are some screen shots from the video:

When stationary:

When panning left:

When panning right:


The differences here are quite subtle. There are visible rolling shutter artefacts when using both cameras. In general, I think the footage coming from the GH3 shows somewhat more similarity between the stationary footage, and the panning footage. But in terms of rolling shutter artefacts, there is not too much difference.

Keep in mind that the panning here was quite extreme. You would rarely pan this fast in reality. So for most uses, the rolling shutter artefacts using any of the cameras is no problem at all.

When using the electronic shutter mode of the Panasonic GH3, though, there is significantly more artefacts coming from the slow sensor readout.

Note that this was not done to compare the video quality between the cameras. I selected quite low bitrate on both, to avoid bogging down my hard drive with large files. That said, I think the GH3 does better in terms of auto white balance.

Wednesday, 2 January 2013

Lumix X 12-35mm f/2.8 review

The Panasonic Lumix X 12-35mm f/2.8 is a premium standard zoom lens. It was released in 2012, but obviously intended as the kit lens for the high end GH3 camera released later the same year. The lens is show below, together the with basic Lumix G 14-42mm f/3.5-5.6 kit zoom lens:

LensLumix G 14-42mm f/3.5-5.6Lumix X 12-35mm f/2.8
AnnouncedMarch 7th, 2010May 21st, 2012
Lens elements/groups12/914/9
Minimum focus0.30m0.25m
Filter thread52mm58mm
Hood suppliedYes, H-FS014042EYes
Lens mountPlasticMetal
Equivalent focal length28-84mm24-70mm


The Panasonic lenses now come in four distinct looks: There are the basic lenses, with the matte appearance and grey ring close to the camera. Then, we have the Leica-branded lenses, which are matte black, without the grey ring.

The two Lumix X branded powerzoom lenses, Lumix X PZ 14-42mm and Lumix X PZ 45-175mm, which have a glossy black finish.

Finally, in 2012, we got the two f/2.8 zoom lenses, with the premium metal finish. The lens barrel is made of anodized metal with a purple-ish colour. This leaves me a bit unhappy. The lens barrel has two functions, in my opinion: To be solid, and to provide a good grip. With the latter in mind, why make it out of glossy metal? Other manufacturers go for a matte crinkle finish, which I think is better.

While the two kit zoom lenses above look similar, they are in fact very different. The Lumix G 14-42mm basic kit lens has a plastic lens mount, and is a very light lens. The Lumix X 12-35mm f/2.8 lens has an OIS switch on the barrel, and is much more heavy.

In use

The zoom ring on the Lumix G 14-42mm basic kit lens is plastic, and feels a bit sticky when I rotate it. The Lumix X 12-35mm lens, on the other hand, has a rubberised zoom ring, which is very smooth to rotate, and well dampened. Even without a power zoom function, it is possible to zoom very smoothly during video using this lens, due to the well dampened zoom ring.

The lens comes with a hood in the box. The hood is well designed, and I certainly recommend using it. The only downside is that the ends are very rounded, meaning that you cannot safely put it upside down on a table.

The focus ring is not rubberised, which is perhaps a bit of a disappointment. But using the focus ring still works quite well. Just like the zoom ring, it has a very smooth and nice damping, even if it doesn't directly control the focus. Just like the majority of the Micro Four Thirds lenses, the manual focus is "by wire".

Autofocus speed

I've tested the two kit zoom lenses on the Panasonic GH3 in the same conditions:

With both lenses at 35mm, the camera spent 0.225s focusing with the Lumix X 12-35mm lens, and 0.298s with the Lumix G 14-42mm lens. Here is another test, showing that the focus speed is even quicker at 12mm, which is quite common for the kit zoom lenses.

In general, this lens focuses very quickly and virtually inaudibly. This is what one would expect from a premium lens like this.


I made a sharpness comparison at close focus distance, comparing the lens with the Lumix G 14mm f/2.5 pancake, and the Sigma 30mm f/2.8 EX DN. The tests were not entirely perfectly made, but they still show that the Lumix X 12-35mm f/2.8 has very good results in terms of sharpness. It is more prone to flare with strong light sources in the frame, though.

Here is another sharpness comparison with the Lumix G 14-42mm lens at infinity focus distance:

At 14mm

Here are the full images:

Lumix G 14-42mm @ 14mm f/3.5 Lumix X 12-35mm @ 14mm f/2.8

To evaluate the image quality, here are 100% crops from the top left corner:

Here, we see that the Lumix X 12-35mm lens is the sharpest, even at f/2.8, wide open. The corner sharpness is very good.

At 35mm

Here are the full images, taken at wide open:

Lumix G 14-42mm @ 35mm f/5.2 Lumix X 12-35mm @ 35mm f/2.8

From the top left corner, I took these 100% crops:

Here we see that the Lumix X 12-35mm lens is a tad bit soft at f/2.8, but becomes very sharp at f/4.


You can find some out of focus rendering (bokeh) for the Lumix X 12-35mm lens here.

Here are some example images taken at close focus distance, so you can see how the out of focus rendering is. Click on the images to enlarge them:

Lumix G 14mm f/2.5 @ f/2.8 Lumix X 12-35mm @ 14mm f/2.8

Here are some 100% crops from the centre using both lenses:

We see that the Lumix G 14mm f/2.5 has somewhat non-circular bokeh outside of the centre of the image frame. The Lumix X 12-35mm lens gives you more smooth out of focus rendering here.

And at 30mm:

Sigma 30mm f/2.8 EX DN @ f/2.8 Lumix X 12-35mm @ 30mm f/2.8

And centre crops from both lenses:

In this comparison, we again see that the Lumix X 12-35mm lens renders the out of focus background the nicest. The Sigma 30mm EX DN lens renders the out of focus highlights with somewhat hard edges and ringing, making the background look more busy. The Lumix X 12-35mm lens gives much more smooth background.

Geometric distortion correction

Like most Micro Four Thirds lenses, this lens also uses in-camera software corrections to give rectilinear images. In the wide end, the lens natively has some heavy barrel distortion, which is corrected. In the long end, there is no geometric distortion correction at all. Read more about it here.

Even after the software distortion correction, there is some residual barrel distortion in the wide end, and some pincushion distortion in the long end, especially at short focus distance. This is a bit disappointing.

Here is an example image illustrating the barrel distortion at 12mm:

I put an orange ruler into the image, to make it easier to see barrel distortion. This was taken at infinity focus. At closer focus, there is even more barrel distortion.

Here's an example at 35mm as well, showing the pincushion distortion:

Example images

This image was taken at 26mm, f/4, ISO 200, 1/60s with the GH3:

This image was taken at 12mm, f/3.5, ISO 200, 1/200s with the GH3:

And some 100% crops from the image:

Example video

This video was recorded with the Panasonic GH3 camera, at f/2.8, in a dimly lit location, handheld:


This is a very good performing lens, with few flaws. It is probably the best lens I have ever used. However, it comes at a rather steep price.

If you want to pay the price, you get a very good lens.