Kamis, 29 November 2012

MILC 4th Birthday Report Part 2, Concept, Implementation , Ergonomics


MIRRORLESS INTERCHANGEABLE LENS CAMERAS      [MILC]
Fourth Birthday Report   Part Two
Ergonomics, Concept and Implementation
Author  AndrewS   November 2012 

Ergonomic Evaluation   This is the second part of my review of current MILC-with-EVF cameras.  I have owned and used Samsung NX, Panasonic G/GH and Olympus EM5 cameras and lenses. I have also made, remade and modified  numerous camera mockups. I discovered many things about ergonomics from all this and have published my findings about ergonomic principles and the operation of specific cameras in many articles on this blog. It is quite possible to make some observations about the ergonomic operation of a camera using information from product photographs, listed specifications and published user feedback. I have used these  sources of information for my comments on some of the cameras referred to in this article.  I am not beholden directly or indirectly to any maker or vendor of photographic products.
Personal Bias ?  Some readers may think the material which follows is just Andrew taking the opportunity to air personal preferences.  Fair enough,  that is inevitable with any opinion piece.   However the presentation is informed by years of hands on experience which I hope will make it useful or at least interesting for those who read on.
Photograph courtesy of camerasize.com
On the left Canon EOS 5D2 [43mm sensor] with 24-70mm f2.8 lens, In the center, Canon EOS 60D [27mm sensor] with 17-55mm f2.8 lens, On the right, Panasonic GH3 [21.5mm sensor] with 12-35mm f2.8 lens. Each of these combinations gives the same field of view range [the 17-55 has slightly less FOV than the others] and aperture.  
Photograph courtesy of camerasize.com
Same camera bodies as above. On the left, 70-200mm f2.8 lens, In the center 70-200mm f4 lens, On the right 35-100mm f2.8 lens.  The 35-100mm on the GH3 matches the 70-200mm on the 5D2 for field of view range and aperture.  There is no exact match for the 60D so I included the 70-200mm f4  for an approximate comparison, albeit with smaller FOV and aperture. These two photos show that lens size is the main contributor to total kit size particularly if longer lenses are included.
Concept and Implementation     In this essay I will consider each maker's MILC-with-EVF offerings from two perspectives:
* Conceptual Integrity.
* Ergonomically Coherent Implementation.
What do I mean by conceptual integrity ? Perhaps some examples will help explain the idea. Through the 1970's and 80's most 35mm SLR's looked very similar. Many medium format cameras were based on the original Hasselblad shape and also looked quite similar.  Then Rollei had the (not so) bright idea to make a 35mm SLR in the form of a medium format camera. This emerged as the SL2000F of 1981, followed by the 3003 in 1985. These cameras flopped in the market place, no doubt for various reasons, but I believe one of those reasons was that the basic concept was wrong. A user interface which worked just fine at medium format size did  not downscale to 35mm size at all. The engineers could downsize the components but not the user's hands.
Here is another Rollei themed  example. One of the most famous cameras of all time was the Rollieflex 2.8F rollfilm Twin Lens Reflex. This was an excellent, some might say sublime, example of conceptual integrity and well designed implementation.  Then someone made a plastic, mini, digital version scaled down to 75mm height. I don't imagine this was ever intended to be more than a conversation piece but apparently it did make pictures. This example is a bit silly but it is a gross example of a product with poor, in this case laughable,  conceptual integrity.
For a more recent example let us consider the Panasonic L-1 of 2007 and its twin-under-the-skin Olympus E-330. These were 4/3 system DSLR cameras. Instead of  the standard top mounted prism and viewfinder arrangement they had a sideways mounted focussing screen and system of mirrors leading to a viewfinder over towards the left side of the camera (as viewed by the user). Why did Panasonic and Olympus adopt this design ?  It did not work any better than a standard shape DSLR and the user interface was not an improvement. Maybe they were trying for a more compact package.  Did they want it to look like a rangefinder ?  Who knows ?   In any event buyers failed to embrace the idea and it was discontinued.   I think at least part of the problem was a lack of conceptual integrity.  I will return to the ongoing problems [well, I think they are problems] with conceptual integrity being experienced by Panasonic in Part 3 of this review.
Readers will bring their own experience and opinions to this type of discussion so I do not ask  the reader to agree or disagree with anything I write but to use this material as an additional source of ideas [if such be desired] about camera design and consumer choices.
 
Leica   
Concept: Leica has been making M-rangefinder ["messsucher"] cameras since the M3 of 1954. The current model M9 operates in essentially the same fashion but with an electronic sensor instead of film.  These cameras are mirrorless and have interchangeable lenses. I am very familiar with  M Leica and SLR film cameras featuring all manual control, having used them for many years. They require regular practice to attain the level of skill required for best results and as a result are satisfying to use. The main controls for primary Exposure [Aperture, Shutter Speed] and Focus [Manual Focus by ring and scale] operate by direct mechanical connection with tactile and kinesthetic feedback to the user. Cameras configured like this made perfect sense in a  mechanical world.
Implementation:  Mechanical aspects of the Leica user interface work just as they have done for almost 60 years which is no bad thing. However some of the electronic functions are poorly implemented.
In all my years of using film cameras I yearned for one improvement to the way cameras operated. This was the ability to magically change film speed in mid roll. Then along came digital and the magic became real. The ability to change ISO any time moved "film speed" up from  Prepare Phase to Capture Phase along with the other primary exposure parameters.  Or at least it could if implemented effectively. But some camera makers just don't get it.
To alter ISO on the M9 you have to press and hold the ISO button which is one of 5 identical buttons on the left side of the monitor screen and while so doing  rotate the control dial on the right side until the desired value is reached. To do this you must stop taking photos, lower the camera from the eye, shift grip with the left hand from the lens to the left side of the body, shift grip with the right hand to get the thumb onto the dial, find the ISO button,  push and hold the button while turning the dial then return your hands and fingers to the capture position.  Lest the reader think this ergonomic deficiency might be a Leica problem, check out the recently released Nikon D600, which has almost exactly the same ISO button arrangement.
Fujifilm    Fuji's entries in the MILC contest are the  X-Pro1 and X-E1.   I have some hands on experience with Fuji cameras. I owned an X10 compact for several months and have had the opportunity to use an X100 owned by a family member.  Unfortunately some of the design problems which I encountered on those two cameras have been transferred across to Fuji's MILC's.
Concept:   Fuji's MILC's are fully electronic but  the X-Pro1 in particular is styled and configured to reprise an M Leica. It's a look-a-Leica.   However unlike the Leica which still uses mechanical lenses, the Fujis are all electronic devices. For a  mechanical camera the lens focus ring, lens aperture ring and shutter speed dial are located where found because of their mechanical connections.  The arrangement has conceptual integrity. But an electronic camera can assign any type and location of UIM (user interface module) to operate any function.  If the layout of a 1950's era mechanical camera has attained such ergonomic perfection that improvement is not possible then so be it, all cameras should have that design. But improvement is readily possible in the electronic era. So why look backwards ? Of course some camera buyers will declare they consider the retro look and layout really cool, or maybe hot, whatever, but I think it just lacks basic conceptual integrity. The X-E1 with it's all electronic viewfinder is a more versatile proposition [it can accept macro, zoom, very wide angle and telephoto lenses which rangefinders cannot accommodate with an optical viewfinder] but the lens aperture and shutter speed controls are still in the same place.
Implementation:   I mention just a few of the many implementation problems with these cameras. The lens axis is inset from the left side of the camera (as viewed by the user) about 75mm. This makes sense on the X-Pro1 [but not the X-E1] to keep the lens out of the OVF field of view. But it also leaves insufficient room on the right side for a proper handle. When the maker gets around to offering telephoto lenses for these cameras the lack of a handle will become a real issue. The rear control dial, which doesn't actually have a function in normal Capture Phase, is located where the right thumb will press on it in normal hold position. Fuji put the rear dial in the same place on the X100 and X10 leading to a brisk trade in aftermarket thumb rests. Simply relocating the dial slightly to the left and ergonomically shaping the thumb rest built into the camera would easily fix all this. The 4 way controller is of the "5 buttons" type. The buttons on this type of UIM are difficult to locate and operate by feel. the "Rocking Saucer" type is much more user friendly. The AF button is the bottom one of three [X-Pro1] or four [X-E1] identical buttons on the left side of the monitor. You have to press this to start the process by which the active AF area can be moved around the frame. But to do so you must go through a task sequence similar to that described above for changing ISO on the M9. It means, in effect, that one of the operational advantages of the mirrorless design has been neutralised by poor and completely unnecessary UIM design.  
Sony     Sony Corporation has a history of innovation. Their cameras have often been leaders in the technnology race, not always backed up by good implementation.
There are currently  two NEX-with-EVF cameras, the NEX 7 and NEX 6. 
Concept:  I regard the  Sony NEX cameras, especially those with EVF as the most unambiguous current expression of the MILC genre.  They are not a reprise or pastiche of some other camera type. They are not emulating a DSLR or a classic rangefinder or somebody's favourite 1980's film SLR.  Their design directly expresses the unique characteristics of the new camera genre, the MILC.  Bravo Sony. 
Implementation:  Here we find many problems.  NEX cams use the 28mm diagonal sensor size. So, while the bodies are miracles of compact engineering the lenses must fit the sensor and thus are about the same size as any other maker's lenses designed for that size sensor and appreciably larger than Micro Four Thirds lenses. If the achievement of smallness was an integral part of the NEX project [and the bodies surely leave one in no doubt about that] then the lenses are a constant impediment to realisation of that goal. Actually I do not see smallness past a certain point to be an inherently virtuous characteristic. Cameras which radically downsize become difficult to handle and create their own design problems. For  instance the NEX 6/7 rubber eyecup protrudes beyond the top and left side of the body where it is likely to get snagged every time the camera is placed in or removed from it's carry bag.
There are many issues with the user interface, both hard [buttons, dials etc] and soft [menus].  The first round of  NEX-without-EVF models had a user interface stripped of  most hard UIM's. This may have been satisfactory for the snapshooter never wishing to leave one of the "All Auto" modes.  But Sony and most other MILC makers apear not to have understood initially that many buyers who wanted a camera with interchangeable lenses also wanted the level of user control over camera operation which came as standard with a DSLR.  The early models had a user interface which was extremely frustrating for the controller. Later models including the 6 and 7 do acknowlege the requirements of controllers but the user interface of the 6 and 7 has been laid over the top of  [in the case of scroll wheels and Mode dial, literally on top of]  the original physical and electronic architecture from the 3 and 5 making for a combination lacking coherence.  Consider, for example, the function of  the AEL AF/MF button. As described in the Digital Photography Review NEX7 report, the behaviour of this button is defined by no fewer than four menu options on three menus, with the resultant function "....not entirely obvious from the manual...." I regard this as completely unacceptable especially given that it is easy to create a user interface in which button functions can be set in direct and simple fashion from a list.
In a recently published review on his SansMirror website, Thom Hogan wrote  "To put it simply the the NEX menu system is a mess."
There are also specific issues with the physical user interface.   My studies of functional anatomy applied to camera design, which you can read about elsewhere on this blog, indicate that the best finger to operate a scroll wheel is the right index, as it has high tactile and kinesthetic sensitivity and is the only one not required for holding the device. There is enough space on the NEX cams for a scroll wheel in the optimum location just behind and at the same height as the shutter release button. But with the NEX7, Sony's designers chose to allocate no scroll wheel to the index finger but three scroll wheels to the thumb. This makes the user interface ergonomically unbalaced. The thumb must stop gripping the camera while trying to manage three scroll wheels. Meanwhile the index finger sits there doing nothing. With the NEX6, they finally got the message that a Mode Dial is a really good idea for an electronic camera but then stuck it on top of and on the same axis as  the thumb operated scroll wheel where it might be bumped off  position inadvertently. Then there is the location of the video button about which there have been many complaints on user forums. 
Nikon   With it's long and proud history of making well regarded cameras one might have expected Nikon's first foray into the mirrorless sector would be with a product which achieved excellence in all ways including ergonomics.
Concept:  One might have expected the new MILC to clearly define  an upgrade path on the "Nikon Way" with a coherent and familiar user interface from compact to MILC to DSLR.  But no, they came up with the 1 series which appeared to have been designed with no connection to other Nikon products at all, apart from the adapter for existing lenses.
Implementation:     Then they gave the 1 series truly awful ergonomics making one wonder whether anyone at Nikon actually understands the basic principles of functional anatomy and user interface design. My question is, how could they make cameras, many well regarded, for 75 years yet be unable to transfer (mostly) good ergonomics from the established product line to the new ?  I don't pretend to know the answer to this question but I do have a theory about it. The shape and layout of early SLR designs were dictated by the mechanical and optical relationships between the various parts. Thus the lens mount, flipping mirror, focus screen, pentaprism, film pathway, wind on and rewind cranks and shutter release button were all located where they had to be for their mechanical or optical connections to function. The height and depth of the camera body were determined by the need to make room for the film and the flipping mirror. Each new model made small iterative changes to the details without changing the basic SLR architecture. But when confronted with the task of designing a completely new camera from a blank sheet with no set mechanical requirements, the designers appear to have found themselves completely adrift, without a book containing  basic principles of ergonomics and guidelines as to how they should proceed.
Canon   Canon was the last of the major camera makers to release a MILC, four years after the Panasonic G1.  Canon's product development people had abundant opportunity to examine all the competition, identify the strengths and weaknesses of each then deliver a category killer product.  What we got was the most underwhelming, derivative, uninspiring new MILC of them all.
Concept:  I struggle to identify any clearly expressed concept  behind the EOS-M.  It looks like they took an EOS 650D, chopped off  the useful parts for holding and viewing then  priced it the same as the 650D.  Is that a concept ???   
Implementation:  Actually I shouldn't even be commenting about it here as I elected only to deal with the cameras having a built in EVF, which the EOS-M  lacks. It lacks almost everything else too, such as a handle, decent thumbrest, any chance of attaching a viewfinder, built in flash, swing out monitor, Shooting Mode Dial, suite of hard UIM's  or fast autofocus.  Maybe it is just a "toe in the water" exercise for Canon which appears not to have noticed that there are a lot of other toes in that particular water and many of them look, and operate, almost exactly the same as the EOS-M.  Presumably Canon has a range of follow up product options on the launching ramp and ready to go, so we shall see what comes next.
Olympus M43     Olympus was a founding adopter of the Micro Four Thirds format but in the first three years produced 7 camera models with not a built in EVF to be seen.  Maybe they were trying to protect their existing 4/3 format DSLR line.  The OMD-EM-5 may be the product which turns Olympus' camera fortunes around, with a lot of help from Sony, which supplies the sensor and has become a significant investor in Olympus Corporation. The EM-5 is clearly designed to be a highly competent all rounder, able to take on almost any photographic challenge. And it almost succeeds, despite problems with concept and implementation.
Concept:  This is a very modern, all electronic digital MILC for the 21st Century yet someone at Olympus thought it would be a good idea to make it look (somewhat) like one of their 1980's OM series SLR film cameras. In my assessment the result is conceptual dissonance between the image and the reality. It is not a 1980's film SLR so why make it look like one ?  I bet the majority of potential buyers for the EM5 neither know nor care that it looks somewhat like an OM4ti.
Implementation:  The conceptual dissonance inherent in the EM5 has ergonomic consequences. There is no room for an inbuilt flash. The control modules are small and cramped, making several of them awkward to operate. There is no built in handle so you have to buy one as an accessory. Having done so the top deck of the camera is cluttered up with three scroll wheels and two shutter buttons, taking up valuable camera real estate which could have been used more productively.  I bought an EM5 earlier this year and was impressed with it's image quality and some of it's advanced features.  But I sold it  because, despite spending a lot of time and effort trying to learn it's unique characteristics,  I could not engage with this camera's user interface.  Every time I went to use it,  I felt like a student sitting for an exam and repeatedly failing.
 Pentax    Pentax has two MILC's, the Q/Q10 and K-01, neither with EVF unfortunately. The Q is like a small sensor compact but with interchangeable lenses some of which are designated "Toy" lenses, which might be a clue as to the Q's place in the camera world.
The K-01 appears to be an attempt to produce a MILC but keep faith with the scattered tribes of K mount lens owners. Then they got a guy who does furniture and jewellery to design it. Hmmmmm....
Ricoh    Ricoh's entry into the MILC sector is the  GXR with it's unique system consisting of base module with handle, controls and monitor, [but without built in EVF]  plus interchangeable lensor [lens plus sensor] modules. If you want to buy a new lens you must have the sensor which comes with it in the same module.  There has not been a rush of buyers to embrace this concept.
I see camera design as having parallels with  politics. Designers and policy developers have to be seen as innovative yet safe, as dramatised in the 1980's British TV programme, "Yes Minister". I suspect that Sir Humphrey Appleby might have described the Pentax K-01 and the Ricoh GXR as "courageous", the kiss of death for any policy thus described.
Samsung   I really don't know why Samsung bothers with cameras, the production of which represents a very small fraction of one percent of total  corporate product output. I bought a Samsung NX10  soon after it's release in 2010. It had better image quality and better ergonomics than the Panasonic G1. I used this and it's slightly updated successor the NX11 for two years with a variety of lenses, some of which were of good quality. But after boasting it was going to conquer the world of MILC's, Samsung failed to deliver the goods and was soon overtaken by the best of the MILC crowd.  These days it's not clear to me as a consumer what Samsung Imaging's intentions might be regarding camera development. Will they persevere with the NX line ?   I can't determine this from Samsung's press or product releases and have moved to M43 where there does appear to be a committment to making cameras.
Panasonic  is last on my list and because I have quite a bit to say about Panasonic's adventures in MILC land it gets all of Part 3 of this fourth birthday review.

Tidak ada komentar:

Posting Komentar