4/3 and full frame

[Dave Humphries]

It's probably 2 years since I last posted on CiC (SHAME!!) and,lurking around, I chanced upon this elderly thread.
It would be very interesting to hear how many posters moved into 4/3 or M4/3, and whether this was a temporary or a permanent move.

Hi Mike,

Welcome back.

Well, since I wrote above, given that both I and my daughter shoot distant wildlife, when she decided to move up from my old bridge camera to a DSLR, we considered all the options and the benefits of 2x (over my 1.5x) crop factor and not wanting weighty camera and lens led her to buy the Panasonic GH2, a M4/3 format camera for which we bought the 14-140 and 100-300 lenses, so she can shoot at 600mm FFE against my 450mm FFE. No plans to change, so it must be permanent.

I'll stick with my 1.5 cf Nikon, but have dabbled with a Mega-zoom bridge camera (Nikon P510) which goes to 1000mm FFE at f/5.9 as a much cheaper and lighter way to get closer than spending between £1k5 and 7k2 on a long lens for the D5000.
In good light, it definitely produces better results than my Nikon DSLR with 70-300mm (450mm FFE) cropped down.
The downsides are;
I cannot push the iso so hard in poor light (my limits are 400iso max on P510 against 2000iso on D5000)
The AF can be maddeningly slow and I miss so many shots
I cannot crop down so much from full image on the P510 at 1000mm as I can with the D5000.
I am currently considering the Canon SX50 which goes to 1200mm at f/5.9, or the Panasonic FZ200, which only goes to 600mm, but does so at f/2.8!
The latter allowing 2+ stops faster shutter speed or lower iso, but with far less benefit in focal length over my DSLR.
Decisions, decisions

Cheers,

[Inkanyezi]

Hi Mike,
I didn't make the move from full frame to µ4/3, as I never had a FF digital camera, but I chose the micro for several reasons, one of them was the excellent viewfinder. Much has also happened in the couple of years since this thread was started. A few years ago, the small µ4/3 sensor indeed displayed rather much noise if ISO was cranked up, and the image quality could be compared to much older sensors for APS-C format. For example my EOS 10D had a small, almost unnoticeable edge over the Lumix G1.

Now as the G1 quit inadvertently, I got an OM-D body, and the difference is huge. The newer µ4/3 sensor is in fact comparable to many present APS-C sensors, although of course a sensor of the same class in a larger format will be better. However, there's presently no Canon APS-C sensor that is noticeable better, so it's not only a matter of sensor size.

Also the viewfinder is better than on the older camera. The EVF of the G1 was far better than any bridge model in all respects, speed, resolution etc. There was no noticeable shutter lag, and the viewfinder lag is mostly negligible. The newer EVF of the OM-D is better in all respects, and some of its features have expanded the possibilities, as the tilting touch screen, where you by the touch of a finger can set focus and snap the image in a much simpler way than ever before.

And of course there are comparable cameras also with a larger sensor, as the NEX 7 or 6 that have twice as large sensor of the same quality and a few extra goodies, as for example focus peaking, enabling convenient work with older lenses or with tilt. But there is one feature with the OM-D that makes it better for me, its sensor stabilising. The image stabiliser in the OM-D is the best I have ever tried. It is simply marvelous. And it works with all lenses, those with adapter. also tilt, and even with very long focal length.

My former digital cameras were compacts, and I still use them. But I was working with photography in the past and have used most any size of camera. I am very impressed by the little OM-D, but it is not as small as I would have liked to have it, although smaller than a 35 SLR. Its image quality is really impressive, and I think it is permanent. I wouldn't want a larger camera, and I don't really think there's a substantial difference that would benefit me. I have tried a medium format camera with 65+ back, and although markedly better on many counts, I wouldn't want to carry one around, even if I could have afforded to get one, which I cannot.

And I guess the race isn't over yet. Incremental improvements are continuing, so we may expect even better performance from products to come. So I think the "move" is permanent.

[ajohnw]

I have full frame, aps and m4/3. My aps camera a 300d is rather old and just a mere 6mp and I was thinking about upgrading it. Instead I bought an olympus e-pl1 to try m4/3. Pretty happy with the shots that this took but pretty soon bought the clip on viewfinder which adds rather a lot to the cost. Photographically speaking all I would complain about is rather low dynamic range at the dark end of things. The info is there but needs processing to bring it out. High iso shot in iAuto mode for jpg's were fine up to 1600 iso. Trying to process that myself rather than letting the camera do it was rather difficult. Other problems - noisy view in low light and often problems with af in low light as well. By low light I mean typical dim living room especially shady situations. The 14-42mm zoom lens that comes with the Pen's is an excellent lens unless you happen to get a bad one - personally I suspect that people who have had this problem haven't got to grips with the camera and most especially the AF. I was pretty hooked at this stage. Main problem was lack of controls. The e-pl1 is very compact like and the Pen's do have a lot of settings variation so I next bought an E-P3. This does give rotary controls that are easier to operate when using the same clip on viewfinder. There are other bells and whistles as well eg touch and object on the screen and the camera focuses on it and takes a shot, also varying camera curves which alter how light levels are distributed in a jpg. The AF in low light and the view though the viewfinder where vastly superior to the E-PL1 so much so I don't think they use the same sensor.

More or less hooked at this stage. There are all sorts of lens adapters for manual focus lenses available for m4/3 but there is a problem on Pen's. No image stabilisation while focusing. It only works when a shot is taken. There is a magnified view for accurate focusing an 7x is about right but it's impossible to hold longer lenses still. Where manual focusing is used with Pen lenses it can be a problem as well. More or less ok at shorter focal lengths but increasingly difficult past that. As telephoto use is likely to need more manual focusing people buy Panasonic lenses for the built in image stabilisation. The only problem with that is that Panasonic have a strong tendency to correct problems in lenses in software so they are best used on Panasonic cameras. How in important that is depends on how far shots are pushed size wise - prints etc but it can show up in shots for the web as well. The worst example I have come across is the Panasonic 45-200mm which I don't think is a very good lens anyway especially at the long end. The 100-300mm is much better.

So to cut a long story short I bought an Olympus OMD-EM5 instead of upgrading the aps 300D. That came with the 12-50mm zoom. I am getting the feeling that this lens isn't as good as the Pen 14-42mm. Not sure yet. It's also a lot longer than the 14-42mm. The EM5 can be bought with that lens. It's very compact but not water proof.

My metric for lenses is that I should be able to crop a photo out of a full frame with a longer length of 1000 pixels or more and display it on the web without any problems. Oly 14-42mm fine, Panasonic 45-200mm no way over a large portion of it's focal length range, Panasonic 100-300mm, getting there but some reduction is probably a good idea- a larger crop reduced to 1000 pixels. Olympus 40-150mm getting their but a bit weak at 150mm.

The Olympus has a rather complicated menu structure and my feeling is that it has to be gone through thoroughly. It takes a while to sort out which metering and AF modes to use. Several hundred shots and there are still some other facilities I haven't tried. The EM5 comes set up dslr like by default Odd really as some of the Penish facilities are rather useful. It has more options than a Pen too. I set mine up Pen like and then found that this might not be the best thing to do.

The Panasonics are cheaper and probably more easy to use option wise as there aren't so many. I discounted Panasonic from day one due to experiences with their previous cameras. Some aspects made the camera look good spec wise but were junk in practice. Maybe a bit like the 45-200mm but being fair that is a cheap lens and it's not junk. I just expect better. Some of it's failings are probably down to the zoom range exceeding 3:1.

The crop factor of 2:1 rather than 1.6:1 on aps is in my view is irrelevant but I believe that Sony Nex aps sensors are better as a for instance. The EM-5 may measure up to these as it does use a Sony sensor. Nex's being aimed at compact users are easier to use. Bigger lenses though. I haven't had my EM-5 long. Few shots and I'm finding it ok to use ergonomically.

Delayed electronic view.?Having used them I would say that's bunkum on Olympus cameras. Pass on Panasonic. Looks like the EM-5 has a slower update rate or maybe faster not sure. The problem is that there has to be sufficient light for the sensor to maintain a fast update rate. I believe this was a problem on some now rather old cameras. Actually I wouldn't mind a very slow update rate for focusing in very low light. No doubt people who mention this can see TV's flickering - Oly screens update faster than that which is significantly faster than our eyes can. The E-P3 and EM-5 have a live view boost mode for low light. Normally the viewfinder is illuminated to indicate what the shot is likely to look like. The live view boost makes it as bright as the camera can. Surprisingly bright in some cases.

No time to check my odd ball typo's so hope this all makes sense. I've started early this morning on this one and have been away from it several times for long periods during the day - can lead to some confused english.

John
-

[rpcrowe]

I don't know how this impacts real-life still image shooting but the Canon APS-C (1.6x) format is actually a bit wider in relationship to the height (22.2mm x 14.8mm or about 1:1.5) than is the 4/3 format (17.3mm x 13mm or about 1:1.33)... The height/width of the 1.6x format has the same relationship as the that of a full frame camera (36mm x 24mm or about 1:1.5). The 4/3 format height/width relationship is roughly the same as the old 16mm film format. Althoughgh I shot 16mm for years and years and did pretty darn well using that format; I always felt constrained by the height/width relationship of the format... I enjoy shooting with the 16 x 9 format of my Panasonic HDC-TM9000P video camera. The far wider width to height relationship (1:1.78) of video as opposed to the 4:3 relationship (which is the relationship of old time television) is easier for me to shoot with.

I am mentioning this because, I think that the 1:33 relationship of the 4/3 system might make me feel cramped when shooting stills as compared to the 1:5 aspect of both full frame and 1.6x cameras.

I am not comparing the aspect ratio of shooting video on a 4/3 system with shooting video on a 1.6x or full frame camera. I am mentioning the 16mm film vs. modern video aspects just to illustrate that I feel cramped with the 16mm aspect ratio vs. the modern video aspect ratio and I am extrapolating that I would feel as cramped comparing the 4/3 aspect ratio with the full frame or 1.6x ratio when shooting stills.

Another aspect of still photography about which I am quite sensitive is shutter lag. I originally bought an Olympus C5050Z P&S camera so I could post images of my Maltese various verious pet and rescue Internet sites. Shooting film (I was using a canon 35mm film SLR setup) required me to shoot, process and scan to a digital format. This took time and I sometimes lost opportunities to place dogs because beause I could not provide images quickly enough. The C5050Z P&S camera cost, during the 1990's, as much as an entry level DSLR and kit lens does now, so it was a pretty decent camera.

However, I could not abide the shtter lag in shooting moving subjects like puppies. I got sick of pressing the shutter button and waiting (even with prefocus) for the camera to capture the tails of my puppies as they scampered out of the frame.

I have not used a 4/3 system but, I would want to ensure that the shutterlag using the LCD and or electronic viewfinder was that which I could live with. I do know that using live view with most DSLR cameras increases the shutter lag dramatically and I also know that some DSLR cameras have a longer shtter lag than other cameras, even when viewing through the lens.

I am not considering changing to the 4/3 format but, if I were, I would want to mak sure that I could live with the shutter lag of any camera I bought.

[Inkanyezi]

I'd like to add a tidbit on the Panasonic 45-200, which I also found rather soft from start, when I had it on the G1. On the OM-D I use the sensor stabilisation instead of optical, and it's like a totally different lens, because it is much sharper when the "Mega OIS" is disabled. The Olympus stabilisation is the best I have encountered, and it is even seen in the viewfinder when the shutter button is halfway down. It is really amazing, and the Panasonic optical stabilisation doesn't even come close. Optical stabilising always degrades the performance of a lens, while sensor stabilisation does not.

[ajohnw]

I'd like to add a tidbit on the Panasonic 45-200, which I also found rather soft from start, when I had it on the G1. On the OM-D I use the sensor stabilisation instead of optical, and it's like a totally different lens, because it is much sharper when the "Mega OIS" is disabled. The Olympus stabilisation is the best I have encountered, and it is even seen in the viewfinder when the shutter button is halfway down. It is really amazing, and the Panasonic optical stabilisation doesn't even come close. Optical stabilising always degrades the performance of a lens, while sensor stabilisation does not.

Not sure about that really. I posted some full res moon shots using the 45-200mm at 200mm on here even using the E-P3 2x digital zoom. I could hear the Pansonic image stabilisation banging about while I did it. Hand held and no signs of camera shake. I did switch the camera is off - easy to forget. Where the lens disappointed me I had taken a test shot of some canadian geese that I would have expected to be able to crop out. Insufficient detail in feathers etc. I would say it's better up to maybe 150mm but still soft. This is typical of many wider range zooms really but I feel it's worse than many.

One my lens metric I should mention one other point. The Pen's are 12mp. That comes out at 24mp on full frame. That seemed fairly sensible to me given decent quality lenses. The EM-5 is 16mp or 32mp in full frame which in my view is past the point where we can really expect sharp pixel level shots from any manufacturer really. That really needs to be considered in respect to the 12-50mm zoom. The increased pixel count may very well relate to my initial miss givings about it,

Shutter lag. Can't say I have noticed anything on any of the olympus cameras but I usually shoot stationary subjects, check where the camera has focused and shoot. I have panned moving ducks etc at times and the only problem is the AF getting it wrong unless in the single square mode. That may mean assuming a later crop to get the subject where it's wanted. I'm still working on that. A slower aperture would probably be easier. I shoot in P mode a lot while finding out what a camera does. The aperture can be changed at the touch of a wheel. The E-P3 onwards are reckoned to have the fastest initial AF available but are relatively slow in continuous focus. I didn't like continuous focus at all on the E-PL1 so haven't really used it since. I feel it's aimed at video where the odd error doesn't matter so much. It may be a lot better on the other cameras.

Frame format is selectable on the camera.

John
-

[Inkanyezi]

I don't quite get it on the metrics. The area of the µ4/3 sensor is ¼ of the "full frame", linearly half, hence 12 megapixels on the small sensor would make 48 on the larger one with the same density and size of sensels, and 16 compares to 64 accordingly. But that comparison is rather meaningless unless you want to know where diffraction limiting kicks in (although the AA filter softens the image before diffraction would be noticed). A lens made for the 24x36 mm frame seldom is as sharp over the image circle of APS-C or 4/3 as lenses designed for those sizes. At these pixel densities, the difference may be clearly visible. My sharpest lens so far is the kit lens for Panasonic, 14-45 mm, and the only three lenses for 24x36 I have that rival it are Carl Zeiss Distagon 35/2.8, Sonnar 85/2.8 and Rodenstock Trinar 50/3.5. Also all my Olympus Pen F lenses, 38/1.8, 38/2.8, 60/1.5 and 70/2 are incredibly sharp, although the two tele lenses, 60/1.5 and 70/2, don't have good enough AR coating to be used stopped down.

At the pixel density of the OM-D, maybe the AA filter could be removed to get a sharper result, just as for the D800E, which would also alleviate the problem of flare caused by reflections between the filter and the rear of the lens when older lenses with less stringent AR specifications are used. I wouldn't be surprised if more cameras would be announced without the AA filter, which I believe is a limiting factor for the ultra-sharp 4/3 lenses.

[ajohnw]

I just assumed crop factor related to area without checking as that would be a sensible way to do it - max print size for instance in a round about way. The correct way must be via pixel size as you imply. The fact that a 16mp camera asks more of a lens than a 12mp one is still factual though.

People often talk about diffraction limits and pixels size but do they really think camera optics are diffraction limited? The chance of the number of elements in a typical camera lens coming out diffraction limited must be remote. My understanding is that they are designed to produce a certain sized circle of confusion. I believe that was about 0.020mm on good quality film lenses. Most lenses resolution improve when stopped down and always have done especially across the frame so diffraction in that respect clearly isn't involved. Typical F1.4 lenses gave their best at F4 - F5.6. F1.2 were a step to far. F1.8 and F2 F5.6 F8 according to quality.

I believe the AA filters are going for other reasons really. Difficulty in producing good ones as the pixel count goes up. I also note that Nikon have switched to Sony sensors. The E-PL1 can easily suffer from fringing on high contrast boundaries and I believe that is down to the AA filter. The E-P3 and I suspect the EM-5 don't have such an extreme problem. Maybe they have no AA filter either. It isn't clear which Olympus cameras use a Sony sensor other than that certain models are on clearance.

Taking bayer masks and various interpolation techniques to sort out colours and eventual jpg compression or similar I don't think that diffraction is really meaningful on camera lenses. It will be at some point but not where people expect. On a 12mp Pen I'm sure there is no effect even at F11 probably at F16 too. Black and white subjects such as test chats are likely to be rather different.

On the 14-42mm and general sharpness this may be of interest to some. Taken with an E-PL1 and the MK1 lens. A little over sharpened due to my monitor at the time but that doesn't explain the evenness of detail.. This lens also has virtually zero fall of to the corners at all apertures as well. Info in the exif but I generally shoot at the longer end for perspective. I feel this gives something for the 12-50 to live up to. I intend to try the 14-42 on the OM as that will give a direct comparison. When I post this shot the detail usually causes some rather odd artefacts due to size reduction. It links to a full sized image. To see that it may have to be opened.



Also on this shot I am rather surprised how much difference calibration and the increased dynamic range of my new monitor makes to the detail in the blacks. It also shows how much detail pens bury in blacks even in jpg's. From reviews even more is available in the OM. Much to Collins disgust I often PP jpg's. Some Nikon users do too. It's interesting to not that Pen's and Nikon dslr's have similar style camer tone curves. Canon were more traditional.

John
-

[Inkanyezi]

Oh the diffraction limit actually bears little relation to lens design or manufacture, but it is a purely physical property of the aperture within the lens, and it relates in a way to pixel pitch. When the pixel pitch would not discern diffraction, the sensor is not diffraction limited, but when the Airy disk can be resolved on the sensor, that's when limiting sets in. This happens on a sensor at a certain aperture regardless of lens construction or focal length. Then a "point" in the image is not a point, but the airy pattern. Actually, the pattern is there, to a lesser degree, also at larger apertures, but it defines the limit of the "soft spot" of a lens. Hence with a sensor that has larger sensels and lower resolution may have a wider limit for the "soft spot" toward smaller apertures.

Generally, diffraction is not a problem in cameras, and with large sensors it goes unnoticed, even for sensors with high pixel pitch. But it also sets a limit to the reasonable pixel pitch on a sensor. It doesn't make sense to have the same pixel pitch on a medium format sensor as on the small sensor of a cellphone camera. For the same reason, it makes more sense in medium format lenses to correct them to less resolving power, but better sharpness toward the edge of the image circle.

So with very small sensors, stopping down the lens much is not recommended, as diffraction is a limiting factor, while also with larger sensors and larger focal length lenses diffraction is present, but does not influence the image so much, as a larger airy disk can be accepted.

But it does not reflect the quality of lens making in any way.

[ajohnw]

There is a simple way of viewing optical lens systems in relationship to being able to achieve diffraction limited performance. It's called optical path difference and is based on the speed of light in glass, accounting for that and ensuring that all rays received by the lens travel an identical distance when they enter the Airy disc. In fact for a perfect Airy disc the point the rays must hit is of zero size. Nigh on impossible with refractive optics when colour is considered. Same for off axis images. The other aspect is Rayleigh's critereon for near perfect optics. That is that the wave front error mustn't exceed 1/4 of the wave length of light. Green light is usually used as the metric so 1/4 wave is 0.140 um or 0.00014mm or for some people as they may find it easier to visualise 5 1/2 millionths of an inch. Errors in thickness and curvature in a lens directly relate to this so in real terms there is no way say a 6 element camera lens can be made to this level of accuracy. This is a very simplistic way of explaining the effects of errors but it does give an accurate idea of the problems. It also illustrates why off axis images can't realistically be related to diffraction especially on a flat focal plane. This area also relates to why high end refractive telescopes do not use ever increasing numbers of elements only the minimum that can do the job and even these have to be produced extremely accurately. These also tend to be rated on the basis of producing a certain sized circle of confusion anyway really only it's the theoretical size of the diffraction spot.

Doh hit the wrong button. There is one side effect that could be said to benifit m4/3 and small sensors. Not that it will really. It's possible to design a certain size of system of lenses that are diffraction limited Scale the size up and they no longer are. The most extreme example of this is microscope objectives. Basically F ratio sets the size of the Airy disc and focal length sets the image scale and hence the resolution for the same F ratio. As the design is increased in size distances increase in one way or the other and the effects of errors in manufacture or the design are magnified. Actually there is some one currently working on an an unusual microscope. The idea is to have a whole image of an entire embryo that allows it's development to be studied down to a cell level. He started with high F ration camera lenses and they fell woefully short and then employed an optical designer who immediately pointed out that his problem was lens size, he couldn't view them in the way microscope objectives are viewed. He had chosen the lenses on the basis of F ratio or NA as it is usually referred to in the microscope world. Seems they are making progress with a lens that would stretch even NASA's capabilities.

The diameter of the Airy disc at F11 in green light is 0.015mm. Bigger for red and smaller for blue. Any other size can be obtained by dividing by 11 and multiplying by the F ratio. It's interesting to consider that in relationship to the shots people take and the settings they use. The effects of 4 pixels to obtain colour is bound to come into it as well. It's also probably possible to work back to actual resolution figures from tests bearing in mind that these are taken where the contrast in the image is 50% of what it actually is and on what are effectively black and white 100% contrast targets. It's pretty natural for me to wonder about this sort of thing from an interest in microscopes as here the detail level produced by the objective is known so for colour a simple approach is to use far more pixels than would be needed to record all of the details in a black and white image.

John
-

[xpatUSA]

A small point amongst the Airy tech posts . .

The OP mentions roaring about in the desert and other places.

Hmmm . . . When I remove the lens of my newly-acquired micro-4/3" I see the sensor - no mirror - no dust filter - no removable, cleanable IR filter - nada. Aren't deserts always a bit windy?

[ajohnw]

The olympus cameras have a sticky substance around the sensor to catch dust etc when it's vibrated off it.

Not sure I would fancy desert dust in a conventional dslr - no moving parts in mirrorless cameras.

John
Silly me - apart from the shutter.