Are in-camera profiles really important?

[pico]

Advertisement (gone after registration)

I have stopped using in-camera lens profiles (Leica M9) because post-processing adjustments, some easily called from Adobe software, are better, more flexible.

 

Thinking now that the in-camera profile should just be an EXIF (or alternative) data to use in an optional manner. Or is it already optional?

.

Edited December 8, 2016 by pico

[Joakim]

But isn't at least one camera profile needed to adhere to the DNG standard so that the files can be opened by any raw converter or other software that supports the DNG standard?

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[adan]

What "in-camera" profiles are we talking about? (I also use an M9).

 

There are at least 3 things that can be called "profiles" in operating an M9.

 

- Profiled corrections for wide-angle rangefinder lens problems on a digital sensor (corner vignetting, and color stains around the edges) - the primary reason for the identification of the lens type via 6-bit coding, or the menu list.

- Profile of the color reproduction (a camera profile, not a lens profile)

- Profile-based general lens corrections not specific to rangefinders - distortion, normal vignetting, chromatic aberrations.

 

Only the first two are "in-camera" - corrections that the camera itself makes or uses as it saves the file. And only the first counts if shooting raw/DNG. The second is only used in-camera if shooting jpegs.

 

The first is absolutely necessary to use in-camera if shooting anything wider than 50mm in color. The only external post-processing correction that even comes close to doing what the camera can do itself in that regard is CornerFix. Adobe, etc. have no way at all of getting rid of the multi-colored (not just darkened) vignetting that a 35 or wider lens will produce on an M9. Without using the M9's in-camera profile for that, my 21mm would be useless.

 

The second is irrelevant to shooting raw pictures - .dngs are not assigned a color profile in the camera anyway, only in the "development" of the raw file in the computer afterwards. If shooting jpeg, the in-camera color profile can't be turned off or avoided - it is a hardwired step in producing a jpeg in-camera (along with assigning the color into either sRGB or Adobe 1998 color spaces, and "baking in" sharpening and contrast and saturation, etc.).

 

The third is never an in-camera profile either. The M9 will not correct for distortion or chromatic aberrations, even if it knows which lens you used. The M9 only corrects for the severe and M9/digital-specific problems under profile 1. But some software can use the 6-bit ID in EXIF to automatically correct for CA or distortion for you.

 

(Unlike the Leica Q - which basically has a 28mm fisheye lens until corrected in-camera to straighten the curving lines - but that's another story).

 

I religiously use coded lenses wider than 50mm on the M9, recording those lens profiles, because there is just no way to fix the problems those lenses produce on a digital sensor during post-processing outside the camera, at least not unless one has a boatload of free time to fill. I tried using an uncoded 35 Summilux pre-ASPH when I first got the M9 - even that barely-wide lens had unacceptably cyan corners unless I ID'ed it from the menu list (or eventually applied 6-bit dots for a 35 Summicron v.4 with a Sharpie). I also like having the coded-lens profiles for longer lenses if possible, even though they do not affect image quality. Just for my own reference in EXIF. Can't always tell 75 from 90 or 90 from 135 shots otherwise. But can live without them in the case of, E.G., my 135 Tele-Elmar.

 

I shoot only DNG, so the in-camera color profiling is not applicable anyway. I've built my own color profile(s) in Adobe Camera Raw, which is where they get applied.

 

My Adobe software reads the lens type in EXIF - but for the Leica M lenses, doesn't do anything with that info automatically. I still have to use a pull-down list to apply Adobe's correction for, say, distortion in a Leica 35 Summicron V.4, if I want (I usually don't want).

Edited December 21, 2016 by adan

[otto.f]

I deactivated the lens correction and the AUTO lens detection in my MM1, because I found that it obliterated the beloved character of my lenses, for instance of the Summilux 75. I find it a bit weird and un-Leica-ish that the natural vignetting of a lens is corrected, we don't do that with in an MP or an M-A either, do we?

Edited December 21, 2016 by otto.f

[Joakim]

When you open your raw image in a raw processer like Adobe Camera Raw or Capture One you have the choice to use a camera profile supplied by the software maker or an embedded profile. You can't choose to have use no profile and without either of those two the software can't handle the image properly as I understand it so the in-camera is hardly irrelevant for raw images?

[pico]

What "in-camera" profiles are we talking about? (I also use an M9).

 

There are at least 3 things that can be called "profiles" in operating an M9.

 

- Profiled corrections for wide-angle rangefinder lens problems on a digital sensor (corner vignetting, and color stains around the edges) - the primary reason for the identification of the lens type via 6-bit coding, or the menu list.

[...]

 

The first is absolutely necessary to use in-camera if shooting anything wider than 50mm in color. The only external post-processing correction that even comes close to doing what the camera can do itself in that regard is CornerFix. Adobe, etc. have no way at all of getting rid of the multi-colored (not just darkened) vignetting that a 35 or wider lens will produce on an M9. Without using the M9's in-camera profile for that, my 21mm would be useless.

[...]

 

I religiously use coded lenses wider than 50mm on the M9, recording those lens profiles, because there is just no way to fix the problems those lenses produce on a digital sensor during post-processing outside the camera, at least not unless one has a boatload of free time to fill.

[...]

 

Andy, perhaps you can find where I err. For example, using the CV 10mm with Photoshop 5.1 2 I achieve excellent correction by setting no lens profile which is processed first through ACR, sometimes then from Adobe's profile for the 15mm lens. In fact, the 15mm profile overcompensates for vignetting (for my taste) so I skip that. In all, it is quick and sufficient to batch corrections.

 

I've posted examples.

.

Edited December 23, 2016 by pico

[pico]

Advertisement (gone after registration)

Here are two images, one without Adobe's lens correction in Photoshop, the second with their correction.

10mm CV on M9 with no lens profile set. The subject was between 6 and 10 feet away.

 

No Adobe Photoshop lens correction.

http://www.digoliardi.net/10mm/later-for-comparison/10mm-w-no-adobe-correction.jpg

 

Adobe's Photoshop 15mm lens correction.

http://www.digoliardi.net/10mm/later-for-comparison/10mm-w-adobe-15mm-correction.jpg

 

I find the 10mm lens so amazing is it that my kit is now the CV 10mm, Leica 35mm Summilux V2 and 75mm Summilux V1 ... and a newly purchased additional M9.

.

Edited December 23, 2016 by pico

[adan]

Ach so!

 

If we are talking non-Leica lenses, then the whole lens profile thing becomes irrelevant. Leica's in-camera profiles aren't written for C/V (or any other manufacturer's) lenses. They may or may not work - but it's just luck of the draw, not intentional. And certainly not for a 10mm, which Leica has never had in their lineup.

 

And - that new 10mm C/V is optically designed like the v.3 15mm and 12mm. More of a mini-SLR lens than a classic RF lens (I mean, it is only f/5.6, but nearly as big as the Leica 21 f/1.4!). With digital sensors in mind, so the lens itself eliminates the weird color shifts that come from classic M Leica lenses on color digital sensors. (Not counting "normal" optical color aberrations - purple-green fringing on the tree leaves top right - not the same thing.)

 

For that lens - no, there is little to be gained in IQ from using any recorded M9 lens profile or lens ID.

 

For comparison, see the 2010 picture below (snow scene) with Leica's own 21mm Elmarit-M (pre-ASPH), with early M9 firmware where the profile was not yet correct. Bad magenta staining on the left and bottom, with a touch of cyan/green on the right side, thus the "Italian-Flag" nickname for the phenomenon.

 

Also a shot fom the same year with the OLD tiny C/V 15mm (Broncos fans) - similar edge color stains, with no profile.

 

As I said - a mess that regular post-processing tools can't reasonably fix. Unusable, unless the (now perfected) in-camera profile is used. Sadly - such does not exist for the small 15mm, but the much larger redesigned v.3 15mm works as well as your 10mm. Just different optics.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[jaapv]

 

I religiously use coded lenses wider than 50mm on the M9, recording those lens profiles, because there is just no way to fix the problems those lenses produce on a digital sensor during post-processing outside the camera, at least not unless one has a boatload of free time to fill.

Well, "no way" or "a boatload of free time.".. An LCC profile or Cornerfix are a matter of seconds to apply, just one or two clicks. It is no big deal to fix perfectly in postprocessing.

[pico]

Thank you to Adan for affirming that CV has made an extreme wide-angle lens for a Leica RF superior to what Leica has for lenses almost twice as long.

.

Edited December 24, 2016 by pico

[adan]

Jaap, since a 21 is my main lens, and I may have hundreds of pictures to process on some assignments, jumping over to a 3rd-party program even for a few clicks adds up. Life is too short. Thankfully, Leica got the profile for the M9/21 right, after 18 months.

 

Pico, well yes, Cosina eventually produced some extreme WAs that are digital-friendly (but 2 stops slower than my 21). I say eventually - because their original 15mm was no better in that regard than the Leicas - see previous post. I'll admit that Dr. Mandler was terribly remiss in not predicting the special needs of lenses for digital use in 2010 - back in 1978 when he calculated my 21 Elmarit-M. Obviously he lacked foresight.

 

I tried the new 15mm III - definitely cures the problems with the original from 1998. I may get one. Of course, Leica's 16-18-21 also solved that problem - 10 years ago. I did use the original small C/V 15mm on the M8, which cropped off the problem areas and made a nice (if slow) "20mm f/4.5".

[otto.f]

Here are two images, one without Adobe's lens correction in Photoshop, the second with their correction.

10mm CV on M9 with no lens profile set. The subject was between 6 and 10 feet away.

 

No Adobe Photoshop lens correction.

http://www.digoliardi.net/10mm/later-for-comparison/10mm-w-no-adobe-correction.jpg

 

Adobe's Photoshop 15mm lens correction.

http://www.digoliardi.net/10mm/later-for-comparison/10mm-w-adobe-15mm-correction.jpg

I find the 10mm lens so amazing is it that my kit is now the CV 10mm, Leica 35mm Summilux V2 and 75mm Summilux V1 ... and a newly purchased additional M9.

.

I prefer the uncorrected version. I can hardly see an Italian flag here, but that may be a coincidence given the shadows at the left Edited December 24, 2016 by otto.f

[pico]

adan wrote: "And - that new 10mm C/V is [...]  More of a mini-SLR lens than a classic RF lens"

 

With respect, Andy, I am trying to understand what you wrote. Mini-SLR? Are you suggesting that the C/V 10mm is actually designed for large DSLRs? Then why is the C/V 10mm  so very successful in Leica M mounts for full-frame  if it was not designed for exactly that? Why is it not available (yet) for DSLR cameras?

 

You think the lens is large, no? All of Leica's later designs are hideously  large.

 

Best,

Pico

Edited December 24, 2016 by pico

[ramarren]

 

 

adan wrote: "And - that new 10mm C/V is [...]  More of a mini-SLR lens than a classic RF lens"

 

With respect, Andy, I am trying to understand what you wrote. Mini-SLR? Are you suggesting that the C/V 10mm is actually designed for large DSLRs? Then why is the C/V 10mm  so very successful in Leica M mounts for full-frame  if it was not designed for exactly that? Why is it not available (yet) for DSLR cameras?

... 

 

I suspect Andy is alluding to the fact that the lens design is a retrofocus type, more commonly associated with SLR wide-angle lenses than with classic RF wide-angle lenses (symmetrical designs like Biogons, etc). Of course, many of Leica's more recent M lenses are also retrofocus designs as well. 

 

The Voigtländer 10mm f/5.6 is available in two mounts: M (for Leica) and E (for Sony). It's hard to say from looking at photos, but the design probably requires short flange mount distances and is intended for cameras without a flipping mirror so it can't be offered in an SLR mount. 

[adan]

With respect, Pico, "mini" (short for "miniature") and "large" are antonyms - they mean the opposite of each other. So If I said "mini," I obviously was not suggesting "large."

 

You appear to think I am attacking the C/V 10mm in some way. I'm not.

 

Ramarren gets it - the C/V lens is built with digital in mind, and thus is quite retrofocus and "telecentric" in optical design. Like an SLR 20mm or 14mm, but reduced in size. So that, as much as possible, the light rays strike a sensor at close to 90 degrees. Which avoids the color smears that older, rangefinder lenses designed in an era of film only (like my 21), produce - unless there is camera firmware correction.

 

It's one of the sad facts of digital sensors that they don't play well with classic, historical wide-angle rangefinder lenses designed for film. Everyone is having to design lenses larger than in the past (just look at the Leica SL's 50mm f/1.4, or the Zeiss OTUS 55mm, or Cosina's original vs. current 15mm lenses).

 

Here's the comparison of the 15s - V.3, V.2. and original from the film era (1998 or therabouts): http://www.sonyalpharumors.com/wp-content/uploads/2015/03/swh04.jpg

 

"Mama, it growed!"

 

My 21 Pre-ASPH f/2.8 is no shrinking violet - but it is still only 80% of the weight and 2/3rds the volume of your 10mm - while capturing 4x the light (f/2.8 vs. f/5.6).

 

BUT - because it was designed 20 years before the digital needs were apparent, it requires that I use in-camera profiles (to get back to your topic title) if I want to use that smaller size/larger aperture.

 

So it's a trade-off I have to make.

[pico]

Thank you Andy and Godfrey for your explanations. So is it true that to make superior lenses they must become larger? Looking at the trends it seems to be the case. One exception might be the 50mm Sonnetar f/1.1 (likely 1.6) which is very small albeit with serious aberrations in which I have found a place in my aesthetic. (Perfection is a distraction.)

 

I had never anticipated a rectilinear (if that is what it is) 10mm lens for 35mm. I am happy to have lived to see it become real.

.

Edited December 24, 2016 by pico

[adan]

Well, you have to define "superior." Large aperture? Large angle of view? Better on a digital sensor - or on film? High resolution? High contrast? Rectilinearity or distortion?

 

Largeness is usually not a goal in itself, but a by-product of achieving other goals. Sometimes mutually exclusive, which is why "superior" depends on what you value most.

 

Personally, when it comes to Leica M, I think the "superior" lens is the smallest and lightest for the performance. I don't know that a Leica or C/V 35mm f/1.4 is "superior" to a Canon or Nikon SLR 35 f/1.4 in terms of imaging. I do know that they are very often effectively as good - in a much smaller, lighter package.