Puts's latest M8 color comments - seriously flawed

[sean_reid]

Advertisement (gone after registration)

What Sean said makes good sense and if I get a chance I will conduct an experiment using a Gretagmacbeth and shooting it on a sunny day with and without filter and profiling it in ACR to see what, if any difference it produces. In theory, at least, there should be no difference since the chart should not reflect much IR but given it is sunlight it may be that the heavy IR in sunlight does get reflected back rather than absorbed. I did get a difference between the B&W and the Leica filter when I ran my test. However, when I look at real photos I'm not sure I like what the profiler produced. When looking at the GretagMacbeth shots they looked much better and closer to the real thing with the profiles.

 

There are two variables to keep in mind with that:

 

1. The degree to which an object, including a GMB chart or the like, reflects IR in a given kind of lighting.

 

2. The degree to which the filters change the colors in subjects that are not reflecting much IR. My own observations suggest that the filters might cause a slight change in the color rendering overall but make a huge difference in the color rendering when subjects that reflect a lot of IR are involved. And, frankly, many of us have known about the latter for months now based our own experience photographing at weddings, making pictures of our families, people on the street, etc.

 

I do suggest that anyone who wants to do their own testing make some tests with, say, ten or more different textiles in various kinds of lighting.

 

Cheers,

 

Sean

[Advertisement]

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

[finkaudio]

Hi,

when I saw Mr. Puts review, I decided for me that he does not really understand how to set up a digital workflow.

See, I'm in the consulting business (Audio industry) and I'm used to read a lot of pseudo scientific stuff of no use. Hey, I can make pretty impressive graphs of complete nonsense and put a lot of formulas around it, so everybody is dead impressed. But still nonsense.

So as somebody, who used M8 with and without filters, I can only laugh.

 

Best regards

 

Karl-Heinz

[scott kirkpatrick]

I went back and reread Erwin Puts' article on "the IR filter question" just now. I know the tools he is using because I have used them myself from time to time. He is shooting a GretagMacBeth 24-color chart and using the Imatest (Imatest - image quality evaluation software) color checker routines to compare where the colors came out in RGB values with the values that the chart colors are supposed to generate. There are three areas in the process that have to be controlled if you are to achieve Erwin's objective of reproducible, standard, laboratory results:

 

* the color balance of his lighting. he should have balanced each shot to a standard neutral card, and I didn't see that he did that. I've noticed that color balanced shots in the same lighting with and without an IR-cut filter can differ by around 100-150K and 1 or 2 in the tint (using C1's scale for the tint correction -- I think this would be more like 5-10 on Adobe's tint scale).

 

* the input profile used. C1 has two different "generic" profiles and since the ICC format is a standard there are some home-brewed profiles around as well. Puts used ACR to convert the files. ACR doesn't use an ICC profile, but takes data from the EXIF plus its own "windage" factor for an initial cut. I don't know if the calibration data that Leica ships in the EXIF is different in "On+UV/IR" mode than when in "ON" mode or "OFF" mode, but now I am curious... The original tables are in the firmware, or one can try to read them as they appear in the EXIF. This information is only used by ACR and Adobe products, and doesn't influence ICC-based raw file developers. But without getting this factor nailed down, we don't know what he is doing (and perhaps he doesn't, either).

 

* the output profile used. The comparison to known RGB values also requires being in the correct 8-bit color space, sRGB, aRGB or whatever, for which the GMB patch values are known to Imatest..

 

BTW, if anyone wants to try these and many other tests, Imatest can be downloaded in a trial version that will allow something like 20 free test runs. But read the instructions carefully before trying your 20 tests. Norm Koren knows a great deal and writes very clearly. It's time well spent. He separates the math (in green) from the information in layman's terms.

 

I think Erwin Puts' observations about the noise curves may be interesting, but only when he has gotten the comparison nailed down completely.

 

scott

[wparsonsgisnet]

Puts is up against an awful lot of professional image makers here. We've got guys who have been making and using profiles for this camera for 7 months, now.

 

That's a lot of experience with color balancing. And, the gang here still doesn't think Leica has it nailed down. What IS nailed down here is the methodology.

 

Poor Puts.

[sean_reid]

I went back and reread Erwin Puts' article on "the IR filter question" just now. I know the tools he is using because I have used them myself from time to time. snip

 

I think Erwin Puts' observations about the noise curves may be interesting, but only when he has gotten the comparison nailed down completely.

 

scott

 

Hi Scott,

 

...but you're forgetting the most important thing. The GMB cards don't reflect much IR (and what's there will be largely uniform) and so the whole point of using filters is lost with that sort of test. I'm sure you've considered that already but didn't mention it in your post. Its a bit like researching the effects of exercise on a chicken's milk production. Very detailed data that is almost completely irrelevant when one considers the purpose of an IR filter on an M8 lens. For milk, we normally look to cows, goats, etc.

 

Cheers,

 

Sean

[robertwright]

Nice of you to bring the subject back around to barnyard animals....

 

His test is so flawed in so many ways...

[ken_tanaka]

Advertisement (gone after registration)

I'm not interested in playing amateur test lab technician for cameras. So I am not qualified to speculate whether or not the methodology of a true amateur lab techie, Erwin Putts', is good or bad.

 

I can say that the M8's color balancing has been a bit of a nightmare to me. I have spent many hours trying to master the M8's color balance with IR filters, voodoo dolls, etc. As recently as today I spent a couple of hours experimenting with a 28mm lens after having a particularly nasty session with it this weekend.

 

But I've concluded that it's simply impossible. Overcooked greens/yellows, muddy reds, some dead blue hues, and of course the almost ever-present cyan edge casts with wide lenses are, at least now, take-it-or-leave-it propositions with the M8. Sometimes the M8's color characteristics work in my favor, but usually they do not.

 

I know that I'm an idiot for posting this note, and am sure that one of the Leicaphiles will immediately declare me insane, incompetent, and heretical (as usual). That's fine, although I am none of the above. But my own experience with the M8's color is more similar to what Putts describes than to the ideal.

[Guest guy_mancuso]

No i would not say your insane or anything like that but maybe something is really wrong because in most cases a simple color temp change is all that is needed for many situations be it in C1 or LR. Honestly I'm really having no issues here and even my AWB is pretty good most of the time. Is there anything we can do to help is really the question. i shoot with the filters on exclusively and really not having any bad reds or overcooked greens and yellows. Ken is there a way you can test it against another M8 body, local store or something.

[ken_tanaka]

Thank you for your offer, Guy.

 

I don't have access to a second M8 body at this time. (I nearly bought a second M8 in April but decided to hold off, perhaps in the anticipation that big changes might be in the offing.)

 

I don't really think that my M8 is anomalous, although I might end up sending it to the Fatherland just to have it checked out. I spent some time last week with a very well known photographer, currently semi-retired, who had recently purchased an M8. (He has been shooting with Ms nearly as long as Ms existed.) He was getting the same color results as I have been getting.

 

I also really don't mean to be provocative with my remarks, just presenting my own experiences. The M8 can certainly record wonderful images in a range of conditions. I believe happiness and success with the camera comes from identifying those conditions that best suit the camera's characteristics.

 

For example, the current banner image on my PBase galleries is an M8 image, taken with the 50 Lux. The camera's characteristics played perfectly to the impression I wanted to convey with the image, which underwent very little adjustment in Lightroom. This image and this image were also banner images I captured with the M8.

 

Patience, I believe, will be the key. I believe that the M8's firmware will probably undergo several refinements before the camera is physically superseded. As the camera is not essential to my photography I can afford to be patient.

 

Thanks again, Guy.

[Guest guy_mancuso]

Ken if you want me to take a look at some of the raws that are not cutting it . Just use http://www.yousendit.com and send me a link and be happy to take a look at them. I have a free day tomorrow before i get slammed with a book to do for client that lasts almost 10 days and is a absolute bear to work on, so if you want just send a file and i can see what issues you maybe having.

[sean_reid]

Overcooked greens/yellows, muddy reds, some dead blue hues, and of course the almost ever-present cyan edge casts with wide lenses are, at least now, take-it-or-leave-it propositions with the M8. Sometimes the M8's color characteristics work in my favor, but usually they do not.

 

I know that I'm an idiot for posting this note, and am sure that one of the Leicaphiles will immediately declare me insane, incompetent, and heretical (as usual). That's fine, although I am none of the above. But my own experience with the M8's color is more similar to what Putts describes than to the ideal.

 

I'm not quite following you here. Erwin's article, linked at the start of this thread, uses a color chart (it seems) to compare color response with no filter (lens detection off and on) and with filter (lens detection set to on+UV/IR). It says nothing at all about: "Overcooked greens/yellows, muddy reds, some dead blue hues...etc." Instead his conclusion (flawed though it may be) was:

 

"At least for all lenses above 24mm (and possibly including that focal length too) the firmware update and the IR filter in combination with the 6-bit coding do not improve on the general image quality that can be delivered by the basic combination of the lens without coding and the attachment of the IR filter."

 

How is that related to the color concerns you describe in your post?

[jonoslack]

 

But I've concluded that it's simply impossible. Overcooked greens/yellows, muddy reds, some dead blue hues, and of course the almost ever-present cyan edge casts with wide lenses are, at least now, take-it-or-leave-it propositions with the M8. Sometimes the M8's color characteristics work in my favor, but usually they do not.

 

I know that I'm an idiot for posting this note, and am sure that one of the Leicaphiles will immediately declare me insane, incompetent, and heretical (as usual). That's fine, although I am none of the above. But my own experience with the M8's color is more similar to what Putts describes than to the ideal.

 

HI Ken

I'm the resident idiot around here as far as IR and colours go

I think there is a point that both Guy and Sean (wonderful though they are) spend most of their time shooting people and not vegetation.

 

I've done loads of heartsearching and loads of testing, and the conclusions I've come to are:

1. there IS no answer (there are many, depending on different lighting)

2. for natural work NOT using a filter is a can of worms, but it's a smaller can of worms than USING a filter.

 

I've now taken something like 15,000 nature pictures with the M8, and I've found it quite easy to account for the effect of the infra red on RGB colours in nature, but hellishly difficult to correctly balance colour using the filters.

 

Throw in the effect of differential cyan vignetting, and anyone who tells me "You don't get correct colour without filters" just elicits a bitter laugh. I spent a lot of time with Kodak, working on the firmware of the SLR/n, where they had complex routines and lens tables to get rid of magenta/cyan shifts, and it was hellishly complex, and it depended a lot on the camera knowing which aperture was used (which the Leica never can).

 

I use filters when shooting people and events, and, like Guy, I don't get a lot of difficulty, skin tones and artificial fabrics come out pretty well. For the rest of my photography I take them off and expect to do a little tweaking in post processing (but only a little). These days I'm getting the colour I want, and the greens are better and more natural than I ever managed from my Nikon cameras.

[ken_tanaka]

In general, the lesson I think I've learned after spending much time experimenting is to just leave the IR filters off unless I absolutely need them. My own tests indicate that while the Leica E filters impose less of a cyan/green overcast than the B+W 486 filters, they're both culprits in shifting colors in many "natural" scenes. I've gotten the best color results with the M8 when I just leave the lens naked.

[MarcRochkind]

Sean--

 

Thanks for your detailed reply regarding evaluating color (a dozen or so post above). Now I know I'm not crazy. Or, rather, that I'm no crazier than you are... ;-)

 

--marc

[jonoslack]

In general, the lesson I think I've learned after spending much time experimenting is to just leave the IR filters off unless I absolutely need them. My own tests indicate that while the Leica E filters impose less of a cyan/green overcast than the B+W 486 filters, they're both culprits in shifting colors in many "natural" scenes. I've gotten the best color results with the M8 when I just leave the lens naked.

Phew - I thought I was a lone voice in the wilderness! Nice to have company, there's nothing quite like a 'naked' lens;)

[sean_reid]

Sean--

 

Thanks for your detailed reply regarding evaluating color (a dozen or so post above). Now I know I'm not crazy. Or, rather, that I'm no crazier than you are... ;-)

 

--marc

 

Hi Marc,

 

My pleasure...although being no crazier than me may not qualify one as sane...<G>

 

Cheers,

 

Sean

[Jamie Roberts]

Ken, Jono--

 

So much of this is workflow related.

 

With all due respect, I can't get reasonably good colour for my work from Lightroom right now and Leica DNGs (DMR & M8) no matter what I do. That is, unless I output them with defaults and then work them in LAB in PS.

 

I am not kidding. I don't really think this is an issue, since Lightroom is v1 something and the M8 was probably an afterthought in the grand Adobe scheme of things.

 

Aperture support for the M8 is still a hack. So YMMV. I'm glad it's a workable hack, Jono--honestly. I truly do love your work, and whatever works for you is good for you, and for me when I get to see what you've shot

 

But so much of these complaints or tests depend on the interpretation of the RAW file, the profile applied, and the other workflow elements. To me, that's where the true pseudo science of Puts's approach comes in.

 

The real questions and answers here?

 

1) Do the filters fix the IR problem with black fabrics? Yes.

2) Do they restore magenta saturation? Yes--too much for some people, in fact.

3) Do they cause a cyan cast on coded lenses? No.

4) Do they cause a cyan cast on uncoded lenses? Yes.

5) It that cast hard to correct in RGB? Yes.

6) Is that cast hard to correct in LAB? No, because it's very, very consistent from shadows to highlights.

 

To me that's it. If you want an instant workflow, then the m8 isn't for you. But then, neither are Canon pro cameras either. Or the DMR for that matter. Or a Hassy or Phase back.

 

Ken--saying the M8 is good when conditions favour it is, well, really. I don't think you're insane or anything. But really, if you'd put the same amount of time understanding the M8 as you have your other cameras, I think you'd come to a much different conclusion.

 

Or not. YMMV, and who knows but you?

 

But I can easily say exactly the same thing about my 1ds2. It's good by itself when a (quite narrow) range of conditions favour it. For some people, they swear Canon pro stuff will not deliver good skin tones or textures. I don't agree; but that's me.

 

And Ken, I have no idea what you're talking about in regards to muddy colours with the M8--I'd love to see those RAW files, too!

 

A lot of photography is precisely all about narrowing the conditions for a good latent image, and then making processing compromises. The other part of photography is correcting and printing. Compromises again.

 

But hasn't it always been this way!?

 

To say the M8 entails more serious compromises than any other digital camera right now is a bit ridiculous to me. Yes, it has its flaws; so do other cameras. It's not magic, it has it's advantages, so do other cameras.

 

But in any case, you need to process them all differently, even if the process approach is fundamentally the same.

 

You probably both realise that I've shot many thousands of unfiltered shots, too. And colour-corrected all the ones I wanted to keep--exactly the same way I would with any other digital camera, BTW.

 

But I do like the Leica filters, too. On wider lenses they have a noticeable lighter touch than the BW 486s (as Sean and others have pointed out).

 

Controversially enough, I think, I get fine skin tones from unfiltered glass. I have *never* seen any shot whatsoever of no matter how pale a Caucasian where an IR filter would make any noticeable printable difference whatsoever.

 

I do see a better green / magenta balance and saturation, however, given the right profile in C1 (and the newest C1 provided profile is quite good in this regard; but still can be improved upon).

 

So I mostly keep the filters on. But they don't *have* to be on. When you're shooting a lot of black synthetic fabric though, well, they're a good bet, is all.

 

So back to Puts--I see his "experiment" as kind of stupid, precisely because it ignores all the real issues when using the camera.

 

Worse, it's dangerous to people who are just moving from film to digital and expect it to be some kind of cakewalk, or that Leica is gouging then for filters. It's not a cakewalk, and Leica is right about the filters--if you're serious about quality and repeatability under the conditions where the filters help.

 

Ok, off the soapbox now. I have another 1000 or so files to process today:) Back to C1.

[ken_tanaka]

Jamie: It's best that we just agree that we disagree on some points. I really don't have the time, energy, or inclination to rebut your remarks. It wouldn't be productive for anyone here, anyway.

 

I will however refute one of your suggestions, namely that I've not devoted enough time and effort into the M8. I've spent literally hundreds of hours with the M8 and its images since February 2, 2007. In fact, I've nearly used the camera exclusively specifically to gain deep and wide insight into it. I really hate to toot my own horn but I'd put my knowledge of this camera, of digital image processing, and of photography in general against nearly anyone else's. So it is with some strong conviction that I posit that the problems I've observed are not "workflow" (i.e. knowledge / skill deficit) related.

 

Jamie, if you're achieving photo nirvana with your M8 good for you! I'll be the last person to pee in your rice bowl. But, at least for the sake of your own health, don't take every contra opinion on the camera so damn personally. Two or more people can simultaneously be "right" when it comes to such matters.

[Advertisement]

Hallo,

Look here: Leica M

[Jamie Roberts]

Ken--

 

LOL!!! I'm not taking anything here personally. Life is too short!

 

But you said you weren't getting good colour from the camera, and so I responded to that (and your claim that it's really only satisfactory in a narrow range of conditions--whatever that really means!).

 

So you're of course entitled to your opinion! And entitled to use Lightroom too, though IMO it still sucks right now for the camera for all kinds of reasons.

 

And I'm not sure you really read my post, especially the part where I say I've had a lot of success without the filters Yes, two people (or more) can be "right" at the same time about any of this!

 

And really, Ken, I'm not a rabid Leicaphile. I don't think I've ever been gushing over reaching photographic "nirvana"--or anything like it--with the M8. I have tried to minimize its flaws in my own work, and help others along, because it's advantages frankly outweigh a lot of the flaws.

 

But I've never, ever, ever said it didn't have any. Just that any other system I've used had them too. Hardly personal

 

IOW, IMO, It's a great tool if you know what you're doing. The filters work the way they're supposed to; that was the point of my post.

 

And frankly, for the sake of others who may be reading the forum, I felt compelled to address both your fairly broad statements about the "failings" of the M8, and Puts's failure to ask appropriate questions of his test. You may have been trying not sound provocative, but frankly your list of M8 failings (and the mention of some semi-famous, er, no "famous" but semi-retired, photographer as ethical backup) was pretty provocative.

 

We all know there are people who don't like the M8, famous or not. There are those who love it too.

 

But there's absolutely nothing personal about this.

 

Personally, I like your work (what I've seen posted)--I think it's great and machine be damned; I really don't care what camera takes a great picture. That's personal.

 

The rest of this is just offered for the sake of balance. You're not the only one with strong opinions on the M8, after all, and if you say um, stuff, expect stuff back.

 

But nothing personal, rabid, or otherwise philic, other than photophilic, I guess

[jonoslack]

Ken, Jono--

 

Aperture support for the M8 is still a hack. So YMMV. I'm glad it's a workable hack, Jono--honestly. I truly do love your work, and whatever works for you is good for you, and for me when I get to see what you've shot

Thank you - I think that the point of this is that after many hundreds of hours of work I've found that I'm getting the best colour for nature out of Aperture (probably because it starts at such a low point in terms of saturation and contrast). I mostly agree with you about Lightroom, but. tell me true (do) how much time have you actually spent in Aperture or Lightroom?

 

I'll confess that since around February I've completely given up on C1 - I never got what I thought of as accurate greens out of it, whether I used filters or not, and now I have a workflow that works, I'm not going to go back and spend another couple of hundred hours trying to get decent results out of a program which seems to me to be old fashioned and clunky.

 

I do agree with you about Eriwn's stuff though - it really doesn't seem to be useful at any level.

 

 

The real questions and answers here?

 

1) Do the filters fix the IR problem with black fabrics? Yes.

2) Do they restore magenta saturation? Yes--too much for some people, in fact.

3) Do they cause a cyan cast on coded lenses? No.

4) Do they cause a cyan cast on uncoded lenses? Yes.

5) It that cast hard to correct in RGB? Yes.

6) Is that cast hard to correct in LAB? No, because it's very, very consistent from shadows to highlights.

 

This is where we really disagree

1) agreed

2) as you say, too much

3) this is the real issue - I radically and fundamentally disagree - at some apertures and if you're lucky - but not at all apertures - for event and street - fine - for landscape and nature - not a prayer (of course, I could be wrong, but you are being categorical here). Just look at Sean's estimable results on wide angle lenses and say that again!

4) Unquestionably

5) absolutely

6) haven't tried (lose the will to live). But I don't consider that this sort of colour correction makes 'accurate colour'. I've spent a couple of years trying to deal with it with the Kodak SLR/n, it's a slippery number, and is different for every shot.

 

Tell me I should go back to C1 and go through the whole procedure again, but I rejected the colours at a point when everyone said they were fine, and I still firmly and absolutely consider that using the filters for landscapes with foliage produces nasty greens with unnatural blue tones and impossible cyan drift with wide angles . . . . . . . . .

 

Actually, my real gripe is not how anyone uses the M8, but that there is a kind of accepted wisdom that you use the filters, and that this is also what Leica recommend. There are some circumstances where they are vital and useful (and I have them for all my lenses). But for the most part they stay firmly in the bag.