Sunday, 27 October 2013

Lumix Leica DG Summilux 25mm f/1.4 Review Part 2

In the first instalment of my review, I talked about the basic stuff, like the focus speed, sharpness and bokeh. Here, I take a closer look at other aspects of the lens.

Focus breathing


The Leica 25mm f/1.4 features internal focusing, like most other autofocus Micro Four Thirds lenses. This makes the focus fast and virtually noiseless. However, there is a downside: The field of view changes as the focus is shifted. This gives rise to the focus breathing problem: As the focus moves, objects in the frame appear to change size, to be breathing.



Here is an example illustrating this. I'm using the Leica 25mm f/1.4 lens on the Panasonic GH2. I chose the GH2 deliberately, since it does not focus as efficiently as the GH3. It usually jogs the focus back and forth a bit more to confirm the focus correctness. This creates more focus breathing issues:



In the first example, I am focusing on a model car in the foreground (about one feet away), and infinity. In the second example, I am using the 10mm macro extension ring from Meike/Neewer to be able to focus very closely.

Most Micro Four Thirds lenses suffer from some focus breathing. But the Leica 25mm f/1.4 is among the worst in this matter. This is not a problem for still image work, but if you are going to be using focus pulling methods during video recording, then the lens may displease you.

For serious video use, you may want a manual focus lens, like the Voigtländer Nokton 25mm f/0.95. The well dampened focus ring makes it easy to focus smoothly.

There are two autofocus Micro Four Thirds lenses that don't have internal focusing: The Lumix G 20mm f/1.7 and the Olympus 17mm f/2.8. They have no focus breathing issues. Then again, they focus slower and with more noise.

Compared with other normal lenses


Due to the typical approximate 45mm register distance of classic SLR systems, fast lenses with a focal length of 50mm can be made relatively inexpensive. For that reason, they are quite common, often being call normal lenses.

Here is a collection of lenses with an approximate 50mm field of view:



First row: Leica 25mm f/1.4 lens, Pentax FA 50mm f/1.4

Second row: Canon 50mm f/1.4 lens, Sigma 30mm f/1.4 EX (for APS-C sensor size)

As you see, the Leica 25mm f/1.4 lens is the smallest of the bunch, but not by far. For a Micro Four Thirds lens, it is actually quite large.

To compare them optically as well, I selected the Sigma 30mm f/1.4, the newest design of the other lenses. I used it on a Pentax K10D. Not the newest camera, for sure, but I am only comparing the image sharpness here, not the focus speed or other camera aspects that have typically been improved over the years.

Leica 25mm f/1.4 @ f/1.4Sigma 30mm f/1.4 @ f/1.4

You hardly need to see the 100% crops to see that the images from the Sigma lens are very unsharp. But here they are anyway:



What's going on here is not that the Sigma lens is unbelievably unsharp, but rather that the camera or lens is poorly calibrated. This is a common problem with PDAF DSLR systems: Just try to make a search for "back-focus" or "front-focus", and you'll see that this is an issue people spend a lot of time to resolve.

"Back-focus" and "front-focus" are not issues with Micro Four Thirds. With the contrast detection autofocus (CDAF), you don't need to worry about camera or lens calibration. In all fairness, more recent DSLR cameras can also do CDAF. This way, the focusing is a lot slower, but you are guaranteed a sharper pictures. For example, the Pentax K-3 and Pentax K-30 feature CDAF (often called live view on DSLR cameras).

In the 100% crops, we also see that the Leica 25mm lens, while being very sharp, has some purple fringing artefacts around high contrast areas. This goes away when closing down to f/2.8. You only get these artefacts when you have a high contrast, like the difference between the bright sky and the dark branches in the example above. With a low contrast image, you wouldn't get these purple fringing artefacts.

Aperture and focus noise


Unlike classic manual focus lenses, Micro Four Thirds lenses are of the electro-optical type, meaning that the aperture and focus is set electronically. Older lenses were typically mechanical: Both the focus and the aperture were set by turning rings on the lens.

Changing the aperture or the focus makes some noise. This can be different between lenses. Let's look at the Leica 25mm f/1.4 lens compared with the Lumix G 20mm f/1.7, the main alternative lens:



Leica 25mm f/1.4Lumix G 20mm f/1.7
Aperture change noise73dB66dB
Autofocus noise46dB59dB

With the internal focus, the Leica 25mm f/1.4 lens has virtually no focus noise. However, the aperture change noise is rather high, among the worst of Micro Four Thirds lenses.

There is a reason for the high aperture change noise: This is a lens with a large aperture in the first place, hence, the aperture diaphragm mechanism also needs to be larger than most other lenses. And larger aperture blade make more noise than small aperture blades.

The lens is notorious for the high aperture change noise. This is often referred to online as the rattlesnake sound. What causes this is the combination of the large aperture, and the relatively high noise of the aperture change.

Unlike DSLR systems, the viewfinder image is not optical. Rather, you see the viewfinder image through the sensor. And the sensor has a limited dynamic range capability. Too much light, and it cannot convey the image to the viewfinder at a high quality. Hence, with large f/1.4 viewfinder, the camera will often have to stop down the lens for the viewfinder image to be good enough outdoor on a sunny day. Closing down the aperture generates noise.

But it doesn't stop there. When focusing, the camera has to open up the aperture again, for the most accurate focus. And after the focus is done, the aperture closes down again, if there is a lot of light. Repeating this process again and again generates the rattlesnake sound, as the aperture is jogged up and down.

To avoid this rattlesnake sound, you can set the camera up to not actively doing autofocus when it thinks you are composing the frame. You do this by setting the "Quick AF" to "Off" on the Panasonic GH3:



The "Quick AF" feature attempts to keep the image in focus at all time. This will engage the autofocus very frequently, and cause a lot of rattlesnake sound. Disabling the feature will reduce the noise when using the lens outdoors in sunlight.

Geometric distortion


Most Micro Four Thirds lenses feature some in camera distortion correction. And this lens is no exception. To examine the geometric distortion characteristics, I have photographed a square tiled wall, and then overlaid the out of camera JPEG (in black) with the uncorrected image (in red). I used the third party RAW converter software UFraw to assess the uncorrected image.

Leica 25mm f/1.4 (-8%)Lumix G 20mm f/1.7 (-11%)

The percentage in brackets is the relative distortion correction applied in The Gimp image processing software to get a rectilinear image. Hence, the distortion characteristics of the two lenses is very similar. The 25mm lens has slightly less geometric distortion though.

There has been a rumour that Leica co-branded lenses don't require any software image processing. From this example, we see that this is clearly not true. Like most short Micro Four Thirds lenses, it has some barrel distortion, which is corrected by the in-camera software processing.

Some traditionalist may be put off by this software post processing. However, I think it makes good sense. By allowing the geometric distortion to be corrected by software, the lens designers can focus their effort on other types of distortions which cannot be corrected by software. This allows them to make the lens smaller, less expensive, and better. And who can argue against that?

Night photos


Given the very large maximum aperture, it is tempting to use the lens at evenings and during the night, when it is very dark. You can take photos of scenes lit by street lights, by pushing the ISO up moderately, to around 1600, and using an aperture of around f/1.4. Here are some examples:

ISO 800, 1/80s, f/1.8:




ISO 1250, 1/60s, f/1.4:




ISO 2000, 1/60s, f/1.6:



ISO 3200, 1/25s, f/1.4:



ISO 1600, 1/25s, f/1.4:



These pictures are remarkably sharp, despite being taken in almost pitch dark conditions. We can see some halo effect around strong light sources. This is due to some flare inside the lens. You may get better pictures during night time by using a lens especially designed for such applications, e.g., the Nikon 58mm f/1.4 and the Zeiss Otus 55mm f/1.4. However, these are extremely expensive lenses, and the Leica 25mm f/1.4 does a good job anyway.

Example video


These videos were recorded at f/1.4. I kept the autofocus on, and used the standard mode on the GH3.



At f/1.4, and during daytime, the camera was using very fast shutter speeds, around 1/1000s to 1/4000s. This makes the video a bit stuttering. To get more fluid motion, it is better to use a 180 degree shutter, corresponding to around 1/50s. This would be impossible to combine with a large aperture, since the ISO cannot be set lower than 200.

To get around this, people often use an ND filter. This allows you to use a 180 degree shutter, and at the same time, a large aperture.

Camera bag


I've found that the camera bag Lowepro Munich 100 fits exactly around the Panasonic GH3 with the Leica 25mm lens mounted. So if you want the smallest possible bag to use for the camera and lens, this is a good choice. Since the GH3 is the largest Micro Four Thirds camera at this time, any other camera should also fit inside the bag with the lens mounted.



Conclusion


The Lumix Leica 25mm f/1.4 lens is a premium lens, and it performs very well. It has some less pleasing aspects: The aperture makes more noise than you might expect. When focusing during video recording, you will experience focus breathing. And the bokeh could have been better.

But all in all, it is a very good lens, and it is very fun to use.



10 comments:

  1. This halo you refer to, is this what is known as "coma"?

    ReplyDelete
  2. Thank you for the review.
    I am still surprised by the huge DOF at 1.4. So basically this lens is great for maximising light intake, but not very useful for isolating your subject if you are not right in front of it. Is this correct?
    Because given this fact and the huge form factor (for µ43 system) and high price, it makes it an all around unattractive lens, in my opinion.

    ReplyDelete
    Replies
    1. The DoF is fairly generous for an f/1.4 lens. But it depends on the distance. The shorter the distance, the thinner the DoF.

      So if you need very thin DoF, then I think the Micro Four Thirds system is not the correct choice. Then you should rather go for fullframe. That is probably one of the few advantages of fullframe today: The possibility to get very thin DoF.

      From my point of view, I don't view this as a problem. It is very seldom that I have a problem with too deep DoF. More often, I'm bothered by too thin DoF. I guess it is a matter of personal preferences when it comes to the look of the photos.

      Delete
    2. I totally agree with you. I recently switched from Nikon-DX to MFT. Love the compactness. I would love to add a "standard" lens to my kit, but there are just no reasonable options in the moment for me:

      Panasonic 25 1.4 to big and expensive
      Panasonic 20 1.7 super slow autofocus is making the 1.7 useless in low light situations
      Sigma 30 2.8 compelling but to slow

      So I guess I will wait for an olympus pancake 25mm 1.8.

      Best regards,
      Mario

      Delete
    3. I think it is worth noting that with more recent cameras, the Lumix G 20mm f/1.7 focuses much faster. There is a significant difference between the GH2 and GH3 in that aspect: With the GH3, the Lumix 20mm lens is more usable in terms of autofocus.

      I haven't tried the GX7 and GM1, but I would guess it focuses faster with those cameras too.

      Delete
    4. I have tested the lens on my Olympus E-PL5 and the focus was very slow comared to my Panasonic 14mm 2.5 or Olympus 45 1.8. That's why I've decided against this lens.

      Delete
    5. Yes, I can understand that. The E-PL5 is not the best in terms of focus technology.

      Delete
  3. Have you experimented with the Olympus M.Zuiko 25/1.8 and GH3? How does the video AF work in comparison to this Panasonic 25mm (or Panasonic 20mm)?

    ReplyDelete
    Replies
    1. No, I haven't tried the Olympus 25mm f/1.8 lens. I would guess it can focus a bit faster, as it has a slightly less thin DoF.

      Delete
    2. I tested m.zuiko 25/1.8 and the 2 panasonics 20mm/1.7 and 25m/1.4 with my GH3 for a few hours, trying to find out how they autofocus with video. In 2-3h I couldn't find out enough, but it seemed to me that whereas the panasonics found focus more often with GH3 (in AFC/AFS mode, other settings manual), the m.zuiko's focus didn't have the pumping effect when focusing. The pana 20mm I tried seemed to find focus, then lose it for a second before locking back into focus. The pana 25mm seemed to go too far with focus and then "reverse" slightly to hit the full focus (pump) and stay. Both looked irritating when viewing the video from my computer. Often m.zuiko 25mm didn't focus to the person in the middle of the screen, no matter how close the person was to the camera. But when it found the focus, it didn't hesitate but the transition was smooth.

      This was my few hours of testing in the shop. Has anyone done more investigation on AF video with these lenses (on GH3 or other panasonic camera bodies), the m.zuiko 25mm in particular? Sorry if my terminology is off, my English ain't that splendid.

      Delete