Separate the cyan shift and vignetting effects

[wparsonsgisnet]

Advertisement (gone after registration)

I am confused by the (apparent?) interdependence of the cyan shift and vignetting in the M8, when wide lenses and IR-cut filters are used.

 

In the detailed testing found on ReidReviews, I see that vignetting and cyan appear at the outer edges of these kind of shots. When the M8's (version 1.09) software correction is applied, both the cyan and vignetting are cleared up, the cyan to a lesser degree.

 

However, the darkening that is due to the vignetting is responsible for the appearance of increased cyan color.

 

Also, both vignetting and cyan are dependent upon lens aperture, again the cyan to a lesser degree.

 

I have been urging the correction of these effects in post-processing software -- and some of these attempts are in current threads. If I use the tools being suggested, I will have to make my own photos from which to create masks. Would any of the monsters care to suggest how I can separate these two effects for my postprocessing bliss?

[jaapv]

Work in the appropriate colour channel?

[Jamie Roberts]

{snipped}I have been urging the correction of these effects in post-processing software -- and some of these attempts are in current threads. If I use the tools being suggested, I will have to make my own photos from which to create masks. Would any of the monsters care to suggest how I can separate these two effects for my postprocessing bliss?

 

Bill, the easiest way to separate treating the colour shift from the luminance shift is to convert your shot to the LAB colourspace and build a little curve adjustment layer (you can pretty much follow any of the advice to correct the vignette in LAB).

 

But the big difference here in the LAB colourspace is that the L channel ONLY deals with Luminance detail and the AB channels ONLY deal with colour. So if you make a colour correction, you're NOT affecting the detail of the shot in the same way you would in RGB.

 

The A channel deals with magenta-green balance, and the B channel deals with blue-yellow balance. You can have some very weird looking curves here, since you're only dealing with colour.

 

In theory, you can do the same things by creating two curves in RGB; one for colour correction and one for luminance. The way you do this is to add adjustment layers in PS and then change the colour adjusting layer mode to hue or color and the vignette proper to luminance mode. But in practice, I've found it not quite as separated as it could be Still, you might not see a big difference to your prints.

 

In LAB, if you fool with it a bit, it seems possible to completely adjust for both colour and luminance shift in one correction curve.

 

Then you convert back to your RGB space for printing. I'd only do this late in the post-processing; just before print / display sharpening (unless you're comfortable working entirely in LAB).

 

You can use masks in the way described, or you can just use them as guides for each shot and lens combo. I imagine the ones you want to be most accurate on are the very wide lenses.

 

Hope this helps, and is answering what you're after...

[Guest magyarman]

This very simple processus make by some people extreme complication.

 

First one step, take away vignetting cause by lens and sensor. This can to be with CS2. I like first go to colour adjust and make it note valuable of saturation, than move saturation to zero. Than I use it lens correction take away vignetting. Than I put back saturation to origin valuable. (This just I can better to see vignetting when is monochroma, if you can do ok with colour stay, than you dont need change saturation).

 

Secont one step, take away cyan with PanoTools.

 

Is finish. All this other with mask it, all this stuff is too many complication what it need for simple job.

[Jamie Roberts]

{snipped}

 

Than I use it lens correction take away vignetting. Than I put back saturation to origin valuable. (This just I can better to see vignetting when is monochroma, if you can do ok with colour stay, than you dont need change saturation).

 

Secont one step, take away cyan with PanoTools.

 

Is finish. All this other with mask it, all this stuff is too many complication what it need for simple job.

 

You mean lens correction in ACR? That's great but I don't use ACR

 

Panotools rocks--I use it all the time. But it's not available, IIRC, on a Mac.

 

So you might need some other post-process way to do this. A LAB curve is dead easy.

[Guest guy_mancuso]

I'm telling all my clients that i switched to B&W:D :D

 

As i am cleaning my office out I am just finding stuff everywhere red , yellow filters . I'm waiting for my enlarger to show up after 20 years. Don't buy any bags until you talk to me LOL. i feel like i got lost in a camera store and the only way out was to stuff the car with bags

[LJL]

Advertisement (gone after registration)

You mean lens correction in ACR? That's great but I don't use ACR

 

Panotools rocks--I use it all the time. But it's not available, IIRC, on a Mac.

 

So you might need some other post-process way to do this. A LAB curve is dead easy.

 

Jamie,

See my "mini-rant" in that other thread where you posted the Lab curve corrections for the wide angle shot. I agree with you that this sort of "fix" needs to be more universal. BTW, there are PTtools plug-ins available for the Mac. (Need to find that thread again.) I sitll think the point SHOULD be that these sorts of corrections can be made early and in almost any RAW converter that is going to support the M8 DNG files, preferrably via pull-down menus for folks to pick a lens and make minor adjustment tweaks from there.

 

LJ

[wparsonsgisnet]

Ferenc, I can see that you have a simple two-step way to do this.

 

The trick to me would be to automate it.

 

I think I'm sorry I started this thread because this topic is being handled pretty well in the lengthy thread on

Alternative Method for correcting 486 cyan drift

 

I agree with Jamie that I don't really want to be in PS (until print time). I would prefer to stay in C1 -- entirely in RAW mode -- to make all these changes.

 

Can we automate in C1?

 

I always use Neat Image noise elimination software as my last step. At the moment I do NOT use any sharpening. I have been unable to sharpen without introducing artifacts that bother me, and, with no AA filter, the M8 doesn't seem to produce soft images that need help.

[Guest magyarman]

Ferenc, I can see that you have a simple two-step way to do this.

 

The trick to me would be to automate it.

 

Is no trick write it action for mine schema. How you gonna make it automate any schema except you put together in batch all fotos what make it by same lens? This possibility only 2 ways: lens tell in EXIF (this you need 6bit, so than do not need correct by software!), ora you got very good memorability

[sean_reid]

Ferenc, I can see that you have a simple two-step way to do this.

 

The trick to me would be to automate it.

 

I think I'm sorry I started this thread because this topic is being handled pretty well in the lengthy thread on

Alternative Method for correcting 486 cyan drift

 

I agree with Jamie that I don't really want to be in PS (until print time). I would prefer to stay in C1 -- entirely in RAW mode -- to make all these changes.

 

Can we automate in C1?

 

I always use Neat Image noise elimination software as my last step. At the moment I do NOT use any sharpening. I have been unable to sharpen without introducing artifacts that bother me, and, with no AA filter, the M8 doesn't seem to produce soft images that need help.

 

It won't help you yet but C1 has Version 4.0 coming out this year and who knows what wonderful tools might be in that? <G>

 

Cheers,

 

Sean

[d_brown]

I agree with separating the vignetting issue from the cyan corners for correcting an image. I have been using a method similar to Ferenc's (without the desaturation part) and find it easy, quick, and it works every time. Lenses with 6-bit coding need little or no vignetting correction so all there is to do is run PTCorrect ONLY on the red channel. With only a little practice, one can look at an image and choose the correct value. There is no need for all the fancy procedures. Even files made with the CV15 and the 486 filter look great with little bother.

 

Dale