How big of a problem is corner color shift REALLY? :)

[kennylovrin]

I've started to consider to sell off my dslr kit lately as I've come to realize I just miss my r-d1 as soon as I pick up my 5dmkii. However there is one thing that I miss when I use the r-d1 and that is just a slightly higher resolution.

Therefore I've seriously started to think about getting an m8, 8.2 or maybe even an m9 used if they drop a bit in price when the m10 is out. Realistically though it seems I'm leaning towards one of the m8 models.

So anyway, compared to the r-d1 I'm fairly well aware of the differences, both good and bad, and I think I can deal with it overall. There is just one thing that bothers me a bit because I can't seem to get a clear picture of the problem, and that is the color shifts in corners using non-leica/uncoded lenses.

When I read online it seems like it is a complete pain in the butt, but at the same time I'm having problems finding any clear answers/examples.

So really, how much of a problem is it? There is just no chance I will ever sit and adjust color shifts in every single image, that is just not acceptable to me. Of course that is my personal opinion, but its the way it is.

Is it really such a big issue as it sounds? 🙂

Thanks!

 

[Doug]

At www.reidreviews.com (a subscription site), it appears that each lens review has a section describing its behavior with and without coding (or M9 lens recognition), some with M8 and with/without IR Cut filter. Worth the subscription, just my personal opinion. He may even suggest which coding works well, or even that no coding is needed. The whiteboard tests will give you a good idea of the nature of the problem.

It depends on the individual lens, and the issue changes between an M8 and M9. I have coded all lenses that I use on the M8, and gave in to recognizing the need for the IR Cut filters, not just for black artificial fabrics but also skin tones, foliage and incandescent-lit interiors. Actually, I also retain the IR Cut filters on some lenses used on the M9 to reduce the warm corners often produced by the camera's internal correction... and also to more thoroughly eliminate the IR.

 

[Tom Niblick]

When the M9 first came out lenses like the Zeiss 21/2.8 and the Zeiss 21/4.5, the Leica 21 Super Angulon and the CV 21 caused a red edge. The problem didn't show up on every photo, but it was bad enough that I sold my Zeiss 21/2.8.

With the latest software, you can use the 21/2.8 Zeiss and the 21 CV without fear of the red edge. However, the 21/4.5 Zeiss and the 21 Leica Super Angulon still cause red edges.

 

[willie_901]

For a brief time I considered the M8 when it first came out. While I decided to avoid the M8, I remember reading the corner artifacts with shorter focal length lenses were easy to eliminate using specialized post- processing software.

 

[kennylovrin]

Thanks for your replies!

It does indeed sound to me that this is quite a problem? I can't help but find it a bit curious that this is even accepted for a camera of this price class..

So the problem is exactly the same on the M9, it's just that on the M9 the menu allows you to select lens so coding isn't as imporant in itself?

The upside is that I saw the Cornerfix app actually works on the RAW data (I thought it needed TIFFs), so that makes it at least somewhat more feasible for me to use it.

By the way, does this only start to become a problem around 24mm and wider? I'm not too fond of wider than 35mm generally so it may not be a problem for me as much then in reality?

 

[Roger Hicks]

Depends on the picture. My own view is that it normally matters a lot more to pixel peepers and obsessives than to those who just want to take pictures. There are several 21mm and 15mm pictures in http://www.rogerandfrances.com/subscription/m8.html (M8 review).

Cheers,

R.

 

[Richard G]

I notice it with my ZM 21 Biogon C 4.5 which I knew would be a problem when I later got the M9. But I valued it for its wonderful lack of distortion and I have been very happy with it on the M9, including in colour. The red fringing is real and noticeable in some shots. This is the only lens where there is any problem. I have the ZM and the Voigtländer 25s and they are perfect on the M9 and the Zeiss C Biogon and v4 Leica 35s are perfect also. I have downloaded Corner Fix but haven't yet fiddled with it to see how it works. I am happy to live with my 21 C Biogon and will see what Corner Fix offers.

 

[jaapv]

Using Capture One it is a total non-issue. take one shot through something white opaque, even a sheet of paper, upload in C1, set as LCC profile and all shots with that lens can be corrected with one mouseclick....

 

[Tom Niblick]

Thanks for your replies!

It does indeed sound to me that this is quite a problem? I can't help but find it a bit curious that this is even accepted for a camera of this price class..

So the problem is exactly the same on the M9, it's just that on the M9 the menu allows you to select lens so coding isn't as imporant in itself?

The upside is that I saw the Cornerfix app actually works on the RAW data (I thought it needed TIFFs), so that makes it at least somewhat more feasible for me to use it.

By the way, does this only start to become a problem around 24mm and wider? I'm not too fond of wider than 35mm generally so it may not be a problem for me as much then in reality?

If you buy a M8, code your lenses and use the filters and you will not have a problem. If you buy a M9 and stick with the focal lengths 28 - 135, you will not have a problem. If you add a 24 or wider to your kit, either master a simple software fix or pick a wide angle lens that is known to be trouble free.

When the M8 and M9 first came out, everybody squawked and raised a big stink so you will find all of those old complaints on the internet. At this point in time, red edges are not a big issue.

 

[kennylovrin]

Thanks guys, it does indeed sound like it isn't as much of a problem as I first thought.
I have no issue with running a few images through cornerfix now and then, especially now that I know it works on a RAW level. I was more worried that it sounded like every single shot would need this treatment.

Again, thanks for the input!

 

[icebear]

Trust M9 users only, not what gets bloated up on the web. I never had any trouble (so far) with a couple 1000 pics. But my shortest focal lenght is a 25/2.8 Biogon for which I manually select 24/2.8 from the M9 menue. The manual coding works but wears off after 2 to 3 changes of the lens. For me that was a waste of money. I would not consider an M8 or 8.2.

 

[x-ray]

No problems with M9 and my 25 Biogon and sold my 21 Elmarit but didn't have any shift, just darkening in the corners.

 

[hteasley]

A 21mm Super Angulon on an M9. Lots of vignetting, lots of purple at some edges, some green on other edges. This is about as bad as it gets: extreme angle from a very close rear element.

And I like it sometimes (like in this shot).

 

[kennylovrin]

hteasley, even though i can see the color shift in that image, it's looks pretty cool 🙂

icebear, you say you wouldn't consider an M8. but then again, the price difference between the M8 and the M9 is pretty huge right, so do you actually mean that you'd rather not have a digital rangefinder than buy the M8? I guess then that in my case it would mean either just use the R-D1, or get an M9 and don't bother at all with the M8.

 

[hteasley]

hteasley, even though i can see the color shift in that image, it's looks pretty cool 🙂

Thanks, I appreciate it. That shot is uncorrected, btw: no lens was specified (I tend to leave it on auto detect, and forgot to set something when I put the SA on). This is one of the first color shots were I like the color vignetting: usually, I convert to B&W in cases like this.

 

[Godfrey]

Color shift and such can be a problem. Depends on the particular lens and sensor combination, as well as both your subject matter AND intent/expectations in the imaging qualities.

The solutions are reasonably simple: choose lenses that work with the sensor the way you prefer. Use corrective software when applicable. Learn and set appropiate expectations for your work given the equipment at hand.

What is amazing is that a digital RF camera capable of working with up to a century old lenses that were never designed for it exists at all, not that it doesn't work perfectly with any lens. A scant decade go it was unclear that it was possible to build such a thing without also creating an entirely new line of lenses, and whether anyone would even be interested enough to buy it.

Remember, it is you who make the photographs. Like with all tools, with any camera you exploit what it can do and work around what it doesn't do well.

No camera is perfect.

 

[ajramirez]

No issue at all with 24mm 3.8 Elmar M on M9.

Cheers,

Antonio

 

[Ron (Netherlands)]

Can be a big issue: try a VC 15mm on a Epson R-D1

 

[Godfrey]

Example of a lens that doesn't work all that well, natively, with the M9:

Voigtländer Color Skopar 28mm f/3.5 (LTM)

The array of images shown below show a set of test calibration pairs, one set with the M9's correction set to 11804, another to 11604, and one turned off. The camera was set on a tripod and focused on the wall, two strobes were used to even illumination as much as possible (in a two minute setup... ;-). Calibration exposures were made at each correction setting, then test exposures were made at each f/stop, and the the entire batch of raw files were fed through CornerFix.

From this test, it is obvious to me:
- that the Color Skopar 28 is not well matched to the M9 sensor
- that CornerFix renders that fact irrelevant.

The processing in all cases was done at Lightroom's defaults for the M9 raw output.