Blog readers and others occasionally send me questions, and I like to reply publicly so that others who are interested in these topics can also see the answers. Here are a few from the past few weeks. (Some posts are edited slightly here.)
Dave writes:
“I recently was reading a thread here where you posted about the 24-105 vs the 24-70. First I wanted to say thanks for your perspective. I had every intention of selling the 24-105 after much of the hoopla I’ve read about it not being so great and ordering the 24-70 MKII. My first day out with the 5D3 was in a museum (No flash) and the IS on the 24-105 in combination with the ISO on the 5D3 gave me shots I never would have thought possible. In short I’m very happy with the 24-105 when I had zero expectation of keeping it.
The question I have is that my 24-105 does have a fair amount of distortion @ 24mm and I noticed you also use the 17-40. Do you find the 17-40 a better performer @ 24mm and do you like the 17-40 in general?”
There are several interesting questions here, so let me try to take them one at at time.
First, though, an article I wrote a while back that looked at the pluses of the 24-105mm lens: In Praise of the Canon EF 24-105mm f/4 IS L lens
Image stabilization (IS) versus larger apertures – The maximum aperture of the Canon 24-105mm f/4L IS is f/4. An f/2.8 maximum aperture is available on a Canon 24-70mm zoom and larger apertures come on a number of primes. This would seem to leave the f/4 zoom in the dust for low light photography – but maybe not. In situations where the limiting factor is camera stability (and the photographer’s ability to hold the camera steady at low shutter speeds), IS can provide several stops of additional low light performance if the subject is not moving too fast and/or if some subject motion blur is acceptable or even desired. In a situation like this one – photographing in a museum – the IS can work in light that would require a f/1.4 lens without IS. In addition, working at f/4 gives you slightly more depth of field. This can make focus just a bit less critical – important when working handheld with AF – let you keep more of the subject in focus.
Distortion at 24mm on the 24-105mm f/4L IS – One of the perceived negatives of this lens is that it shows more barrel distortion (a bowing outward of the image along the sides of the frame) at 24mm. It also exhibits more corner light fall off (or vignetting) at the largest apertures and especially at wide angles. In most cases you won’t notice either of these, but if you image contains lines parallel to the frame edges you may well see the barrel distortion when shooting at 24mm and you may well see the vignetting at f/4. (All lenses vignette to some extent – so it isn’t a question of whether a lens vignettes, but of how much. Most wide zooms will also show at least some barrel distortion.)
I rarely notice the barrel distortion. First, most of the time I shoot at focal lengths that don’t exhibit the issue. Second, when I do shoot at 24mm the issue is not visible with most subjects. But sometimes it will be visible at 24mm. However, most current raw conversion software auto-corrects for this. Some worry that this will diminish image resolution, but I’ve look for degradation and I cannot see it. Likewise, the software can automatically correct for vignetting and do so in ways that are essentially imperceptible. However, keep in mind that vignetting can be a beneficial effect in your photograph – in fact, it is not uncommon for photographers to add some in post! It serves to diminish the importance of peripheral parts of your image and can bring more attention to central elements.
What about resolution? Some people claim that the 24-105 is not sharp at 24mm. I have not seen this to be the case at all. I never hesitate to use the lens at 24mm on account of sharpness – though I do recognize that I might have to adjust for barrel distortion or vignetting in post.
17-40mm f/4L vs. 24-105mm f/4L IS – When it comes to comparing lenses this way, there are a few things to keep in mind to help counteract potential Lens Performance Obsession Syndrome. First, no lens is perfect. Each has what I call a “personality” – the sum of its strengths and weaknesses which may make it more or less suitable for various uses. So it isn’t necessarily that one is “better” than the other as much as it is that the two are different. Second, it is possible to become overly concerned about specific aspects of purported lens performance than necessary. In a number of cases, things that can be measured turn out to not matter in actual photographs. If we want to get real picky, at 24mm the 17-40 might have slightly better center resolution than the 24-105, though it would likely have less resolution in the corners wide open… where the 24-105 could be sharper but might exhibit more vignetting and perhaps barrel distortion. In general, I would rarely choose one over the other at 24mm for any of these reasons – frankly, I would likely use whichever one was already on the camera!
What about alternative lenses? – People frequently compare this lens to Canon’s 24-70mm f/2.8 II (or the older version) or even the 24-70 f4 lens. Here, too, the “lens personality” issue comes into play. I have the 24-70mm f/2.8 L II. It is a fine lens that produces excellent image quality and I like using it for certain subjects and circumstances. However, when I use it I miss the larger focal length range of the 24-105 and I often wish that it had IS when I shoot handheld. It is also a bulky and heavy lens. In the end, sometimes I’ll use one of these lenses and sometimes the other!
You and others comparing the lens options might be interested in this article of mine: Canon 24-70mm f/2.8 L II vs. 24-70mm f/4L IS vs. 24-105mm f/4 L IS.
Photographers shooting Canon cropped sensor bodies should take a serious look at the EFS 17-55mm f/2.8 IS lens. It nicely covers the core focal length range for cropped sensor DSLRs, has excellent image quality, goes to f/2.8, and includes image stabilization. In general, most photographers will not likely be as well served by the 24-105 or the 17-40 on cropped sensor bodies.
Bottom line: For my kinds of shooting, I find that the 24-105mm f/4L IS is a great performer on full frame Canon DSLRs. I like its relatively light weight, it relatively large focal length range, the inclusion of IS, and the excellent image quality it can produce, especially when photographs go through a good post-processing workflow.
Mike write:
“Just came across your site this morning and I am really enjoying your logical approach. I’m an intermediate shooter using TS1i body. I like to shoot night cityscapes and, in addition to the kit lens, have been using the ‘nifty fifty’ and the Canon EF 70-300mm f/4-5.6 IS USM Telephoto Zoom Lens (when travelling I rent the 24-105 L, but liked your point about the 17-55 non L for crop sensors)…anyway…I’ve read conflicting advice on whether to turn AF on or off when using a tripod…thoughts?”
The answer to the IS/tripod question used to be easy – turn IS off when the lens is on the tripod. With many older IS-equipped lenses, the image will “wander” if the camera is stable and IS turned on. (My 100-400mm lens is a case in point – the image moves enough in the display that it can wreck my composition!) However, many newer IS-equipped lenses apparently “recognize” when they are tripod-mounted (or, more likely, simply recognize that the camera is not moving) and you don’t have to turn the IS off. My guess is that the 70-300mm lens, being a newer one, will work OK on the tripod.
However, you can easily test this yourself. Put the camera and lens on the tripod. Use “live view” mode so that the image is displayed on the rear LCD. Use the live view zoom controls to display at 10x magnification. Half depress the shutter and watch the image – if it is stable you should be OK. (If you were doing this test with my 100-400 you would quickly notice the problem – and you would switch IS off!)
For my part, I always turn off IS when the lens is on the tripod – this is simply a habit at this point. I also normally use manual focus and live view when shooting from the tripod (with some wildlife photography providing an exception to this rule), to the whole setup for tripod use is different for me.
Someone (sorry I can’t find the email now) wrote to ask about the supposed “noise banding problem” with the Canon 5D2 and possibly other Canon DSLRs. This is a common topic on photography forums – where the discussion frequently drifts away from photography and towards forumtography.
Some will claim that the 5D2 images will show this noise banding in certain specific situations – images shot at low ISO (typically ISO 100) with subjects containing very dark shadow areas and a very wide dynamic range. All digital photographs contain noise but the issue here is that the noise is not completely random, but rather that it exhibits some linear cross-hatching pattern. Some will tell you that problem is significant and always visible, while others will say that it my only be visible in exceptional cases or when significant post-processing is used to light shadow areas.
I have looked at this quite a bit since I frequently shoot at ISO 100, photograph scenes with very wide dynamic range, like to lighten shadows to make details visible, and create large and high quality prints. Frankly, I haven’t found it to be an issue at all. I can sometimes find a bit of banding in noise, but only in rare cases, when inspecting at 100% magnification on the screen, and when making some fairly radical adjustments to the image in post. Among the many tens of thousands of photographs I have made with my 5D2, there has been exactly one in which I thought, maybe, that this might be slightly visible. It turned out that it wasn’t, but if it had been I could have fixed it using certain post-processing techniques.
So I do not regard this as a problem. It is possible to find some rare instances in which banded noise might be visible if you go looking for it, but it just isn’t going to be a problem in real photographs. To some extent this issue tells us as much about photography forums as it does about cameras – there can frequently be a set of forumtography characteristics that create problems out of thin air or very little evidence and then blow them up into imaginary problems. This is one of them.
The lenses in this article are available from site sponsor B&H Photo:
- Canon EF 24-70mm f/2.8L II Lens
- Canon EF 24-70mm f/4.0L IS Lens
- Canon 24-105mm f/4L IS Lens
- Canon EF-S 17-55mm f/2.8 IS Lens (for cropped sensor bodies only)
- Canon 100-400mm f/4.5-5.6L IS lens
© Copyright 2013 G Dan Mitchell – all rights reserved.
G Dan Mitchell is a California photographer and visual opportunist whose subjects include the Pacific coast, redwood forests, central California oak/grasslands, the Sierra Nevada, California deserts, urban landscapes, night photography, and more.
Blog | About | Flickr | Twitter | Facebook | Google+ | 500px.com | LinkedIn | Email
Text, photographs, and other media are © Copyright G Dan Mitchell (or others when indicated) and are not in the public domain and may not be used on websites, blogs, or in other media without advance permission from G Dan Mitchell.
Here’s a common situation when noise banding is present in a repeatable way: take a night photograph of an illuminated building in a fairly dark environment, exposing correctly for the building. Next, brighten the dark areas so that you see some detail in the sky rather than pure black (for instance with a +100 shadwow). The sky will be full of obvious noise. I am surprised you think this is an imaginary problem, since you make such images yourself. Sure, Canon users can get around it with multiple exposures, but Nikon users don’t need to do so.
QT:
If what you are saying is that other, newer sensor designs continue to improve on dynamic range, I can’t argue with you there. As we all know, the Sony and Nikon cameras using newer Sony sensor technology seem to have expanded the dynamic range… in the same way that Canon sensor expanded it from what had been possible with early digital cameras.
I’m not quite sure what to say about turning a black sky some color other than black. If the sky was black, then black is the properly registered color. If it wasn’t black, then an exposure that captured that color would seem to be a good choice.
At some point, any camera will be unable to capture the full dynamic range of a scene – at least any camera that we can acquire today. The problem is that scenes that exceed the camera’s dynamic range often exceed it by many stops. Night photography, especially urban night photography with its combination of bare light sources and other essentially black areas is a classic challenge. I know this from doing quite a bit of urban night photography myself, and the noise-banding issue has never been a limitation for me in this work.
I have seen noise banding in my own photographs on few occasions, but not in night photography. The one image on which I thought it was actually going to be a problem was a photograph that included some large areas of reflected sunlight on the ocean and some nearby cliff faces in very dark shadows. For several reasons, the exposure-blending approach proved problematic on this image, and the image that gave me a usable image of the water was very dark in these shadows. Making the problem worse, I had shot at ISO 100, which is where the noise banding will be most visible, if it can be seen. In the end it turned out that the slight banding in these dark shadows, while barely visible at 100% on the screen, was not visible in a print. If it had been, I would have done a slight noise overlay layer blended in the area where it occurred.
I have occasionally also been able to see it at 100% magnification in green areas of smooth gradients, for example in OOF background bokeh in close up shots of foliage – though none of these were visible in the print either.
So, no, it hasn’t been a problem for me.
Of course, I understand that you know enough (and then some!) about these things to have a rational understanding of what noise-banding is and isn’t. When I wrote this I was mainly thinking about those forumtographers who go on an length about how Canon cameras are defective and unusable due to this issue – and that is pretty much nonsense.
Take care,
Dan