Mescalamba wrote:
Im fine with that correction as long as it works. Had both G1 and 14-45mm lens. Problem was, that it didnt work enough to cover all CA and fringe. Certain amount of that purplish-blue fringe is not possible to remove even today with last LR4.1. Same goes with too much of distortion correction. That lens lost a lot of its resolution thanks to that.
IMHO bit of CA and distortion correction is fine. We all do it sometimes in LR (or different SWs) to remove that bit of both that our lens produce. But its very bad idea to use it as part of lens design.
Latest 12-35mm f2.8 from Panasonic has pretty big amount of fringing too.. and thats some really expensive lens. Tho at least its amount thats correctable in LR. Still if I would pay so much for that lens..
Plus its really stupid in perspective. I have ages old Minolta lens that produce no fringing and only bit of CA in corners. Its old design, old lens and still much better than most Panasonic lens (in fact, probably all of them except few of those Leica designs).
Considering one of points and strengths of m4/3 and pretty much all mirrorless is ability to use almost all lens ever made, its very bad decision to depend on software corrections for native lens so much....Show more →
Yup, Plasticsonic has a habit so far of thinking their lenses are made of gold and pricing them as such when it's clearly NOT the case! Some are good but none of them are truly excellent (aka "exceptional"). And almost all of them release with a price inflation of around 2x and sometimes more! I wish people would boycott all new Plastisonic lens releases until they get their prices right or something - but whatever.
The in-body correction problem for me is that we never know how much color detail, edge sharpness, light spot definition, or micro-contrast the removal process is destroying without testing the lens on a camera body without that. Having done so with the 14-42 I noticed the lens actually improved a little bit without it. Distant light points (like street lights etc) were more interesting and less smeary and micro contrast was up slightly. I would be OK with the in-body correction if I could turn it off somehow! Maybe we can get Vitaliy Kiselev to include that in his patch utility?
Bif, I recall it as the internal CA correction is affecting in-camera JPG images only, no? These days my G1 see no use as I robbed myself and bought an E-M5.
In addition to the problems caused by the fishtank glass I also suspect the Panasonic and Olympus stack of glass surfaces in front of the sensor adds to the problem seen in the original image by Patrick.
I'm not sure I'm remembering it correctly or not but apparently there is a file chunk in the RAW defining the corrections which all RAW converters (silently) respect and apply. RPP might be an exception, I dunno.
IIRC I think you are remembering correctly.
Then it is the JPGs that are affected only and then it is up to the user to decide what raw converter to use, no? RPP is snobbish but RawTheerape might be an option though.
In my opinion ACR and Lightroom (and other converters as well perhaps) should let the user decide what automatic corrections that should be applied. Now we are were we are. In any case, it is possible to avoid the "problematic" lateral CA corrections for anyone who so wishes.
RAW Therapee and DCraw (or DCraw based) convertors ignore all corrections. Not a pleasant sight if you shot something contrasty at 14mm with 14-45mm from Panasonic.
You can check results for corrected and uncorrected. Not exactly amazing even after correction, actually pretty horrible lens.. maybe they are using recycled vine bottles for glass..
Jonas B wrote:
In addition to the problems caused by the fishtank glass I also suspect the Panasonic and Olympus stack of glass surfaces in front of the sensor adds to the problem seen in the original image by Patrick.
Sure, but some perspective is in order. The sensor filter stack is about 5mm thick and native lenses are designed with its effect already negated. The tank glass is probably around 10mm thick, plus there's perhaps 10cm of water, and the lens designer had no idea TOP was going to shoot through that!
Aiming straight through glass will keep the effect to a minimum, while any tilt relative to the flat surface will compound the problem. Overall, concern about systemic lens aberrations when shooting into a fishtank is just about ridiculous.
acme_d wrote:
Sure, but some perspective is in order. The sensor filter stack is about 5mm thick and native lenses are designed with its effect already negated. The tank glass is probably around 10mm thick, plus there's perhaps 10cm of water, and the lens designer had no idea TOP was going to shoot through that!
Aiming straight through glass will keep the effect to a minimum, while any tilt relative to the flat surface will compound the problem. Overall, concern about systemic lens aberrations when shooting into a fishtank is just about ridiculous.
/Acme
Well, thank you, I guess...
I wasn't thinking about "systemic lens aberrations". Maybe I should have expressed myself better. I'll try again; the filter stack in front of the sensor has shown to produce some smearing and also increased halos around extreme highlights. I think that 's one of the reasons behind the flare seen on the fish.
I wasn't thinking about "systemic lens aberrations". Maybe I should have expressed myself better. I'll try again; the filter stack in front of the sensor has shown to produce some smearing and also increased halos around extreme highlights. I think that 's one of the reasons behind the flare seen on the fish.
Ironically, since it's flat, the main side effect from the filter stack is spherical aberration. The wider, or faster, the lens, the greater the impact. It's negated in lenses that have been designed for the system, but you will see it in some degree when adapting film lenses.
Every time light crosses a boundary between materials of differing refractive index on the way to the sensor, those materials have entered into the optical system. There is no free pass just because they were ahead of the sensor, inside the lens, or outside of the lens. Because it's much thicker, and contains two such transitions (air to glass and glass to water), the effect of the fishtank, even if perfectly lined up upon, is bound to be far larger than the effect of the filter stack, and, unlike the filter stack, it's effect has not been negated by the design of the lens.
These same issues apply to underwater housings, where terrestrial lenses stare out via glass ports. These housings, however, do ensure that the alignment is as good as it can be.
acme_d wrote:
Ironically, since it's flat, the main side effect from the filter stack is spherical aberration. The wider, or faster, the lens, the greater the impact. It's negated in lenses that have been designed for the system, but you will see it in some degree when adapting film lenses.
(...)
I have used micro 4/3 system cameras since the G1 back in 2008 and before that a couple of 4/3 cameras. I own and use the same combo Patrick used for his image. I have a box of adapters and the cameras has certainly showed what they can and not can do.
When I talk about noticeable more flare in extreme high contrast areas from these systems compared to systems without the same thick stack of glass I do it from experience. But wth, if you don't think it is worth mentioning this problem as something adding to the problems (as I mentioned in my first reply) I won't repeat it.
Jonas B wrote:
I have used micro 4/3 system cameras since the G1 back in 2008 and before that a couple of 4/3 cameras. I own and use the same combo Patrick used for his image. I have a box of adapters and the cameras has certainly showed what they can and not can do.
When I talk about noticeable more flare in extreme high contrast areas from these systems compared to systems without the same thick stack of glass I do it from experience. But wth, if you don't think it is worth mentioning this problem as something adding to the problems (as I mentioned in my first reply) I won't repeat it.
The filter stack is an optical component like any other. It flares like any other, to a degree determined by the quality of its coatings, and by its geometry within the optical system. Flats tend to more readily produce visible flare than curved surfaces, so you certainly may have been able to generate such a flare.
But all of this is moot in the context of shooting into a fish tank. It's like worrying about smog while puffing on a cigarette.
OK, here is a better comparison, though not perfect. Both images were F5.6, 1/200, ISO 1600. WB was set at 12000K to match the lights, Raw conversion, auto curves, 100% crop.
Here is what I notice...
1. Fringing is comparable between the two. Stopping down to 5.6 really helped the PL25.
2. The 5D II is a cleaner image. I used no noise reduction on either image. Maybe the OMD needs some but I am not sure. I am just learning the camera. Also, the noise difference is really only noticeable at 1:1. At 1:2 it is not noticeable.
3. The colors are a bit different between the two bodies. I believe the OMD is more natural and accurate.
Jonas B wrote:
IIRC I think you are remembering correctly.
Then it is the JPGs that are affected only and then it is up to the user to decide what raw converter to use, no? RPP is snobbish but RawTheerape might be an option though.
In my opinion ACR and Lightroom (and other converters as well perhaps) should let the user decide what automatic corrections that should be applied. Now we are were we are. In any case, it is possible to avoid the "problematic" lateral CA corrections for anyone who so wishes.
Yeah, but it's not just ACR and LR. It's CaptureOne, Aperture, ACR/PS/Bridge, LR, SilkyPix, DXO, ACDSee Pro, and pretty much any of the popular editors going. RPP is an interesting tool for some things but it's not really anything anyone would want to use as their "photo editor" or really even in their workflow. (Of course it takes all kinds but... )
I agree tho that it should be an option - but as I understand it if they want official information from Plastisoniic they have to implement it Plastisonic's way - via contractual agreement. (I'm not 100% certain of that but it's been said by at least one developer I know so I tend to think that's correct.)
-- -- -- -- -- -- -- -- -- --
Pat,
That looks better. Considering only "fringe" I like the OMD image a little better than the DSLR one.
)