A Sane Attitude to Rededge

[UliWer]

Advertisement (gone after registration)

Can you show us an example of the 35mm with red edges?

 

I am not asking this because I don't believe what you say, but I would like to compare it to what I have seen with other lenses and what I know about the problem with the M8 (yes, it already existed with the M8).

[jaapv]

with 35lux there are red corners - no question about. would you like to recommend not to use in winter time? there is lot of upsides with M9 and the superbe lenses, but this is a major downside.

Is your lens coded? If not did you select it manually? Did you use the latest firmware? All of that makes a crucial difference.

[SJP]

Yes. Problem is, I think the variability is inherent in the sensor design, at least in the sense that the sensor responds differently to different lenses.

 

Regards,

 

Sandy

Wouldn't it be possible to make a camera specific "fine tuning" option in the firmware? If it is microlens drift then the correction should be a smooth function that could be tweaked by the perfectionists and set to a generic value for the rest of the world. Leica could even provide an online service to analyse images emailed to them and suggest good starting values. No need to send anything to Solms - plus now they have some work cut out for them for the M10. We were wondering about that yesterday at the Dutch users meeting. "What should the next generation offer?" .... and we couldn't think of anything.

[sandymc]

Wouldn't it be possible to make a camera specific "fine tuning" option in the firmware? If it is microlens drift then the correction should be a smooth function that could be tweaked by the perfectionists and set to a generic value for the rest of the world. Leica could even provide an online service to analyse images emailed to them and suggest good starting values. No need to send anything to Solms - plus now they have some work cut out for them for the M10. We were wondering about that yesterday at the Dutch users meeting. "What should the next generation offer?" .... and we couldn't think of anything.

 

You could set up a fine tuning option - problem is, you'd be fine tuning a lot of different points - probably at least 8, maybe more. At that point, I think they'd be better off using something similar to CornerFix to automatically analyses the vignetting, then upload a profile to the M9. Which CornerFix could do, btw, all that's lacking at the moment is the ability to upload a DNG camera profile into the M9.

 

Sandy

[meino]

You could set up a fine tuning option - problem is, you'd be fine tuning a lot of different points - probably at least 8, maybe more. At that point, I think they'd be better off using something similar to CornerFix to automatically analyses the vignetting, then upload a profile to the M9. Which CornerFix could do, btw, all that's lacking at the moment is the ability to upload a DNG camera profile into the M9.

 

Sandy

I am just a simple IT man, but I have been thinking over this problem. Would it be possible to create a algorithm using a number of input parameters, that create the necessary corrections.

For the input parameters I was thinking of:

- Lens code.

- Distance of the exit pupil from the sensor.

- Diafragma for which the optimal correction is calculated

- 2 parameters that define the up/down and left/right deviation from the center point of the image.

 

It would be nice if we could define these parameters in a small text file which the firmware would load when present on a SD card

 

regards

Meino de Graaf

[sandymc]

I am just a simple IT man, but I have been thinking over this problem. Would it be possible to create a algorithm using a number of input parameters, that create the necessary corrections.

For the input parameters I was thinking of:

- Lens code.

- Distance of the exit pupil from the sensor.

- Diafragma for which the optimal correction is calculated

- 2 parameters that define the up/down and left/right deviation from the center point of the image.

 

It would be nice if we could define these parameters in a small text file which the firmware would load when present on a SD card

 

regards

Meino de Graaf

 

Meino,

 

That would indeed be ideal. The problem is, I'm not aware of any "model" of vignetting behavior in either an M8 or an M9 that you could base that on. Maybe Leica or Kodak have the knowledge required to produce that, but if so, they're keeping very quiet about it.

 

Having said that CornerFix's profiles are actually very small - only 2.2k, although they have quite a lot more numbers in them, and they are not a "parametric" description of the lens. So you could upload a CornerFix profile, as you suggest, quite easily, but I seriously doubt that the processor in the M9 could run a CornerFix-type correction in a reasonable time. Which is why I would suggest injecting a DNG camera profile into the image, rather than have the M9 calculate pixels - injecting a profile takes almost no time at all.

 

Regards,

 

Sandy

[Jerry_R]

Advertisement (gone after registration)

In Micro43 world - dedicated RAW developers (among others LR) make corrections of RAW files just after opening them, based on metadata written to RAW file by camera.

 

This could be enough. But Olympus and Panasonic talked and co-operated with Adobe.

Leica can't?

[mjh]

In Micro43 world - dedicated RAW developers (among others LR) make corrections of RAW files just after opening them, based on metadata written to RAW file by camera.

 

This could be enough. But Olympus and Panasonic talked and co-operated with Adobe.

Leica can't?

If one does come up with a simple solution to a complex problem that had everyone baffled so far, chances are that the simple solution doesn’t work.

[pgk]

In Micro43 world - dedicated RAW developers (among others LR) make corrections of RAW files just after opening them, based on metadata written to RAW file by camera.

 

This could be enough. But Olympus and Panasonic talked and co-operated with Adobe.

Leica can't?

A problem here is that other than identifying itself (if 6-bit coded) the lens on an M does not communicate with the camera body, so values such as aperture and focus distance (ie variables) are not available to record in the metadata. This will make such adjustments less useful as Leica M metadata probably has less recorded parameters to work from.

[Jerry_R]

chances are that the simple solution doesn’t work.

u43 users do not complain.

 

values such as aperture and focus distance (ie variables) are not available to record in the metadata. This will make such adjustments less useful

Do red corners depend on focal or aperture? As far as I know, we use primes. And most people do not create as many profiles for CornerFix as apertures.

 

Is anyone here optimistic?

 

I think subject has been explored. Let's see if Leica will bring something to M9 or it will be solved with M10.

One is sure - more competition would accelerate this...

[mjh]

u43 users do not complain.

Because the corrections performed are in no way related to the corrections necessary to prevent red edges. Just because you can successfully correct for vignetting, distortion, and chromatic aberration it doesn’t follow that you could correct for everything else just as easily.

 

It’s a tough issue.