Saturday 22 October 2011

Portrait lenses: 5D vs GH1

One of my favourite words is fungible. I'd never heard of it till I worked in the Finance industry (and who'd have thought but its a finance namespace word). Its defined over at Wikitionary as:

Adjective

fungible (comparative more fungible, superlative most fungible)
  1. (finance and commerce) Able to be substituted for something of equal value or utility; interchangeable, exchangeable.

With that definition in mind, the purpose of todays post is to explore the fungiblity of 50mm on 4/3 vs 100mm on full frame.

background


When digital SLR cameras came out I was unimpressed by the APS size sensor. I felt that it twisted lenses into something I was not used to. My fine 24mm wide angle became a middle or nothing semi-normal and my favourite normal (EF50) became some sort of half assed tele (something like 80mm which is too short IMHO). I ended up needing to buy alternative zoom lenses to cover the angles I liked and the DoF was never what I wanted.

Then I tried 4/3 and found that the x2 factor on the lenses was much more acceptable to me than the x1.6 of APS. I did find that the optical SLR viewfinder was pathetically small and so it wasn't till micro4/3 came out with the electronic viewfinders that I was fully comfortable with focus (and yes, I had a number of focus problems with APS cameras which included backfocus and front focus issues as well as inability to clearly manual focus on the small focusing screens).

Since I started using micro4/3 cameras one of the questions I have sought answers to is the ability to substitute half the focal length at 2 stops brighter and get the same thing. Back in March 2010 I made a comparison of this however I used 35mm negative, which didn't fully satisfy my interests because I still wonder about the utility of a 5D in what I photograph.

As I recently borrowed a 5D (for testing the Olympus 21mm f3.5 wide) I thought I'd do another test to satisfy my curiosity (and of course provide information for others pondering the same thing).

Too many reviews compare things with no context other than the optical one. I think that's invalid because I don't know anyone who just buys based on what the magazines say. Most people need reach into their pocket and pull out some money. So in someways there needs to be an examination of cost benefit. So lets examine some costs.

Both the GH1 and the 5D are only available now as used cameras (which to me is a good thing as they are cheaper this way, let someone else take a big depreciation hit if you can). A good used 5D body still fetches about $1000 while a good G1 about $250 and a GH1 (if you were after video as there is no other significant difference between the G1 and GH1) about $400.

The lenses uses in this test were the Canon FD 50 f1.4 and the Olympus OM100mm f2.8 - one can expect to pay about $100 for the FD50 1.4 and about $200 for the OM100mm f1.8 So one can take advantage of cheaper lenses. Because shorter lenses cost less than longer ones of equal quality and also 50mm was once a really popular focal length.

Then there is size the 5D is about twice the dimensions of the G1

So not only do you pay twice as much for the 5D as you do the G1, you have to carry around twice as much of a monster.

People often discount this fact when making considerations and instead ruminate about "what makes the ultimate image".

This "ultimate image" nonsence forgets an important fact in photogaphy: if you don't have your camera with you then you can't take pictures with it.

Hmmm?

This is of course why I bought a G1 and sold most of my EOS gear back in 2009 (after much rumination and after buying the G1 and having if for some months just to be sure).

Clearly then this test is not without me having some pre-conceptions from the outset. I think its important to say that some of my preconceptions were upheld here, while others were challenged and I think its fair to say I learned something in here which I didn't expect to learn (but wondered about in another context)

examination


In this test wanted to examine how well the micro4/3 camera would hold up against a full frame camera in respect to portrait focal lengths. I chose 100mm for the focal length because I've always like it more than 135mm. Canon and Nikon and Olympus alike have all made good sharp f2.8 100mm lenses for some decades. I happen to own a Olympus OM100mm f2.8 which I think is one of the sharpest lenses ever made (the TS-E90mm would get my vote as the best) and the 135mm range is a bit patchy if you ask me. Likewise there are plenty of good 50mm lenses out there (I own a few of them) which of course makes the basis for this comparison.

So to compare a 50mm lens on micro4/3 to a 100mm on a full frame one needs to consider that to get the same DoF look and feel, you need to keep the aperture 2 stops brighter on the 4/3 camera. So in theory 50mm @ f1.4 <=> 100mm @ f2.8. But of course its not without problems because (among other things) the aspect ratio is different the 4/3 format is far more square looking than the 3/2 format of full frame.

Notice that when covering the same width that the 4/3 gives that bit more along the top and bottom.

Next I'll state that all images were taken
  • in RAW
  • on a tripod that didn't move
  • converted using ACR 5.6 with exactly the same parameters
  • exposure set by my Gossen light meter and manual setting applied to the cameras
This was done so as to get the images as close as possible in terms of processing (camera JPG's would be more difficult for the nature of this comparison)

Something which came out right away was that: if you are taking shots in full sunlight that using f1.4 is a challenge.

The shutter speed required is often exceeding what your cameras shutter will allow. For instance on a sunny 16 kind of day (like this was) I was using 4000th of a second to get f1.4 and even then was needing to have shadows as part of the picture or have a washout. Check the nuclear glow of the shirt back for instance. At 4000th of a second fill flash is out of the question too.

So it seems that fast lenses and daylight are not good friends.

So anyway now lets take a look at the images.

First the image from the Olympus 100mm at 2.8


and then the 4/3 camera with the 50mm at 1.4
The first thing I notice is that the colour and contrast is different (there is also bokeh but that's for another post). The FD 50 when opened up to 1.4 is really soft and dull looking which influences the colour. This is clear in the 100% view below, especially where the stark contrast border between the white velcro of the cap strap borders the darkness of the hair in shadow. I would call this "blooming".


Don't get confused and think this 'flare' only exists where there is bright light, it exists all over the image area and is what serves to make the lens "low contrast".

If we now look at the OM100mm at f2.8 we see definitely better contrast.



Its interesting to note however that the outright resolution of detail between the two lenses is quite similar.

This is consistent with what pulled me away from my 10D and 20D cameras in favor of the Panasonic G1 ... it just has heaps of detail.

On that subject its a good time to observe that the 5D turns its 36x24mm capture area into 4368 x 2912 pixels while the G1 divides its (much smaller area) 18x13mm into 4000 x 3000. This means that while the outright capture of detail is quite similar the micro4/3 is actually a little more demanding of lens quality. Since the FD50 f1.4 (or in fact almost any super bright lens) has always been regarded as a little soft fully open, it translates to a bit more than that here.

So when you stop the FD50 down to f2.8 it clears up immensely

to be more or less equal to the contrast of the Olympus 100mm lens.

but of course loosing that nice shallow DoF. I encourage you to open up each of the above images in a separate tab to allow you to click between tabs and see the differences jump out.

Perhaps the new Olympus 45 f1.8 lens would solve these problems and yield a lens which would
  • have the contrast I find in the Olympus 100 f2.8 on full frame,
  • have at least similar (if not better) sharpness
  • be more compact and light weight than the FD f1.4
The lens will set you back about $400 as it seems to be panning out, which is quite attractive. The initial discussion of the new Oly 45f1.8 indicated prices would be much higher, but if it does come in at $400 that will put it right on par with the typical 100mm f2 lens.




So you would then get the compact benefits of micro4/3 without needing to pay through the nose for smaller.

Back when micro4/3 came out there was of course no such lenses available, but the potential was there. So to actually see and explore this potential I have experimented with adapted 35mm lenses so that I can get actual experience with images, not just a theoretical idea of what I can expect.

Both lenses used in this test were designed for different cameras than the EOS series, and so need adaptors. The OM lens was made for a camera with a shorter flange distance than the EOS, but not by much. So it only has a slim adaptor. The FD being used on the micro4/3 however has a much longer adaptor because (again) the flange distance effects the design of the lens (has to clear that moving mirror).

Paradoxically the 50mm lens actually becomes bigger than the 100mm lens when both are on their respective adaptors.

Of course a 50mm lens does not need to have such a massive stand off on a camera without a mirror. In this photo I show the Pentax 50mm lens (far right) beside the 50mm OM lens (on adaptor).

pentax-110-kit


So the new micro4/3 Olympus 45mm f1.8 is likely to be about the same size as that little pentax lens. If you happen to be interested in more compact 50mm portrait lens than the FD 50mm f1.4 I suggest you consider the little Pentax. I have reviews on them here , here and (perhaps the most interesting) here.

Of course adapted lenses lose "niceties" such as Autofocus, Face Recogntion Autofocus, auto aperture and a few other things which can be important (for people who don't know how to focus a camera or set an aperture). Since that time lenses have begun to emerge which don't cost an arm and a leg, may give better image quality and provide all the creature comforts (crutches?) for the photographer who needs them. For what its worth adapted manual focus 35mm lenses cost about ¼ of what one would pay for a modern equivalent.

However there is more interesting information available in this comparison yet. While the depth of field may be reasonably equal in a gross comparison there are some significant differences.
  • The 100mm at f2.8 has a more shallow depth of field at the subject


With both images I tried to focus on the back edge of the cap. Using the magnify zoom focus assist on the GH1 it was easy to get that spot. Looking carefully its clear that the 50mm lens on the 4/3 the focus zone is much deeper even at f1.4. By the distance from the ear to the eyelashes, they are starting to go out of focus on the 100mm at f2.8 but not on the 50mm at f1.4. This means also that you won't get that on the Olympus 45 f1.8 either.

This (shallow DoF) is exactly what one wants in a classic portrait lens.

The ability to have focus sharp and defined on the eyes
(my wife wanted to keep a little more privacy) but be soft by the ears or the back of the head, especially when the head is turned slightly.

Before people mention anything about pixel peeping, this sort of thing will become obvious on a 8x13 print and more so on a larger photo. Even if you don't "see it" obviously here it will be what makes it "feel better" when you examine prints.

This is not to say that the micro4/3 is bad by any stretch of the imagination. Indeed its pretty bloody good.

While playing around with these images I noticed something else, and that was the 5D also had a better dynamic range. You'll notice that the blacks in the shadows were better resolved? Thats because they had less 'noise' (as well as not being covered up by the flare)



The above are both taken at f2.8 (*where the FD lens became quite equal to the OM), the details in the shadow (arrows) are present on the 5D (right hand) while being muddy on the GH1. At the same time the highlight transition to washout (elipse) was similar. See the differences in the hair clarity in the shadows? If you were trying to dig a little into the image to pull out more shadow detail then the micro4/3 would give you a little less clear shadow detail than you'd get out of the 5D.

Lastly I thought I'd present the difference that one gets using a camera like the 5D in terms of outright printable size. The 5D image is slightly more pixels wide (4368 vs 4000), so if you were wanting to keep the aspect ratio at 3/2 then you'll get a slightly bigger print from the 5D, so printing at 300dpi from native:
  • 14 inches wide at native pixels on the 5D
  • 13 inches wide at native pixels on the GH1
Which is not exactly startling is it. On the plus side the 5D will tolerate more enlargement of pixels (upsizing) from native because it has a slightly better image quality (as identified above) and has a bit more to start with. In my comparisons above I've scaled the 5D back to the 4000 pixels of the GH1 so as to make it easier to compare. Here is the actual pixel comparison from these images.


So there is a minor increase in size of features in the picture, but personally I feel that if you wanted to print BIG or you want to capture every last skeric of detail then the 5D would do it that bit better.

Ultimately my call here is that the micro4/3 gives you 95% of the image feel and 99% of the image quality that a 5D will give. If the pursuit of that few percent advantage is worth doubling your money in buying a camera (and lenses)

PS: and now, after a long delay Part II is completed (sorry about the delays but some times life gets in the way)

4 comments:

Mulan said...

I'm just reading this, and this is an excellent comparison.. Just the one I looked for!!

Thanks!!

Anonymous said...

bright sunlight + wide aperture + fastest shutter speed you can get = :-(

bright sunlight + wide aperture + fastest shutter speed you can get + ND filter = :-)

obakesan said...

Anonymous,as soon as you figure out how to do that and not have it effect the fill flash please let me (and everyone else) know.

Anonymous said...

"On that subject its a good time to observe that the 5D turns its 36x24mm capture area into 4368 x 2912 pixels while the G1 divides its (much smaller area) 18x13mm into 4000 x 3000. This means that while the outright capture of detail is quite similar the micro4/3 is actually a little more demanding of lens quality."
An important point, I think, when APS-size sensors are being loaded with more pixels, but still being packaged with consumer-style lenses.

I recall when the Nikon D800 was released (full frame, 36 megapixel) the reviews were very good but there was a warning that if your lenses weren't up to the mark, the D800 would reveal that whereas it mightn't have been apparent previously with lower megapixel full-frame (or APS) sensors.

There is much to know, much to learn, and still try and be creative.