What Sean Reid Knew When?

[sean_reid]

Since I was travelling for a shoot this weekend, I seem to have missed the debates over M8 revierws. In a thread that is now closed, someone asked if I found problems with the production camera that I held back from the review. The answer is no, I did not. Every problem that I detected in my test M8 (production firmware), or in my personal M8, has been discussed in my review series. Problems that I was not aware of when those reviews were first published were discussed in new sections that were added afterwards.

For those of may have missed, this post that I made previously may be relevant:

"In August of this year, I had two opportunities to work with a preproduction M8 so that I could make some initial assessments of it for a review (aspects such as ergonomics, controls, finder, speed, etc.). I wrote about those things in my first review of the M8. I found it to be an excellent camera save for a couple of weaknesses that I discussed in the review (weather sealing and digital controls design, ie: ISO, EV and WB). Those cameras did not have final firmware and none of my later specific picture quality discussions were based on those pre-production cameras. In addition to journalists, some well-known people such as Tom Abrahamson and others were also working with preproduction cameras to get a feel for them, including many people who are active in the LHSA.

On October 20, I received a camera for evaluation that had production level firmware installed. As is the case for all my reviews, I tested it primarily by using it for the sort of photography I normally do. I was getting an unbelievable number of e-mails from people who were, understandably, eager to see me publish on file quality as soon as possible. I worked essentially without sleeping and had Part Two of my review ready two days after the camera had arrived. Part Three was published later on.

Many people test color reproduction by shooting a color chart under controlled lighting. My own preferred color testing is less orthodox. I go out and buy the most visually interesting array of fruits and vegetables that I can find and I photograph them. People know what a Granny Smith apple looks like, they know what peppers look like, etc. Looking at a picture of produce is one way to get a sense of how a given camera's color rendering relates to the actual color of objects in the world. I find this test, actually, to be a very useful reference point for real world color rendering. Naturally, I photographed a WhiBal card as well in order to get a neutral reference point for white balance. The camera came up aces on color rendering, to my eyes and I got no e-mails from readers that suggested that they felt otherwise. But....produce clearly does not show the IR-based magenta color cast. And at that time, I had no reason to even be looking for an IR color cast.

My other pictures, then, were made of people and objects out in the world. If one looks again at the pictures in those reviews, it's possible to see several cases where there may be a magenta cast, including some jackets, a fireman's hat, etc. It certainly wasn't obvious initially though, not only to me but, apparently, also to the thousands of people who read the review and looked at the samples. When one is reviewing pictures made in the streets, etc. he or she would be hard pressed to remember exactly what color hats, coats, etc. each person was wearing. Certainly nothing jumped out at me as being unnatural. Michael Reichman and I were also talking back and forth about the camera and our impressions by phone and e-mail and neither one of us had noticed anything especially strange with the color rendering. My attention was actually concentrated largely on ISO performance and comparisons with the R-D1 and 5D.

I did detect what I thought of as "mild banding" in some ISO 2500 pictures and I mentioned that in the review. But it was rare and didn't seem like much of an issue at all. One can see evidence of it in the picture of the boy flipping burgers but it certainly doesn't jump out at a viewer.

This past weekend, several threads on this forum raised concerns about the light streaking and magenta casts in certain pictures. Neither was prevalent in my own pictures, so far as I could see, but on Monday I contacted several people at Leica and asked for official responses to these concerns. To their credit, they responded and the comments were sent in German to Christian Erhardt who translated them and sent them to me. I then added new sections to the reviews that discussed these issues and also included the official statements. Later, I was asked by a moderator of this forum if I had any official statements from Leica on these issues and I agreed to publish them here. These statements have been widely criticized but I appreciate that Leica provided them rather than ducking my questions. Many other companies have not been so responsive. I also want to reiterate here that anyone who wants to attack Christian for those statements (and I think that's really misguided) is barking up the wrong tree. Christian just made sure that Solms responded and then translated those responses. I want to publicly thank him again for doing that. I think he's a real asset to Leica.

So that's the deal. All that I can do is to report what I know when I know it. All the conspiracy theories (common for the Internet) about this are nonsense. As time goes on and we work more with the M8, we're all likely going to discover both strengths and weaknesses that we haven't experienced yet. Even a very conscientious reviewer is only one person and we, collectively, will discover more about this camera than any one person ever could.

It's clear that the M8 still needs some work. I'm hopeful that the light streaking can be solved in firmware and I suspect that it can be. The IR response issue is probably a much tougher thing to fix but I know that Leica is aggressively looking at how that can be solved. The IR block filters are the best solution they've come up with so far but that doesn't mean that other solutions are not possible in the future. I still argue that the M8, despite its problems, is a very capable camera in many ways. I had to lay out my money to buy one just like anyone else and I have absolutely no plans to return it. To do so, for me, would be throwing out the baby with the bath water. I've order B+W 486 filters to block the IR and will test them tomorrow. If they work, I'll use my camera and the test camera to shoot a wedding on Saturday. People who have concerns about the camera certainly need not buy it or keep it. But there are a core of people here who are keeping the M8 and who are experimenting with finding ways to get the most out of it. I applaud that. No one needs to deny the problems but we all can choose to proceed as we like. For those people who are just now receiving the camera, I'd encourage you to go out and use it to make your usual sorts of pictures. Test for the problems if you want to but consider also just making regular pictures with it, of whatever subjects you normally like to photograph. That way, at least, you can make a balanced decision about it."

I don't wish to reopen Pandora's box but I wanted to make sure my response to that question was available for all forum members. I don't need replies on this thread, necessarily, I just want people to have the information. BTW, Michael Reichman did tell me that he got some green blobs in a parking garage picture or something like that but it isn't something I experienced with my M8 or my test camera. In fact, while I've seen the discussions of them, I still haven't seen any in my own work with the M8.

Cheers,

Sean

[pascal_meheut]

One thing to support Sean: I looked at the pictures I made before the IR problem was known but after I found a magenta cast under incandescent light.
Some of them exhibit the "magenta clothes" problem but I did not noticed it at the time. Now, it is obvious.

It is very easy to reproduce a problem once it is detected and analyzed. And even easier to comment on it "ad nauseam" on the Internet.
Finding it is another story.

[k_g_wolf]

Thx for these remarks, they hopefully help to calm things down.

But I suggest, we call he M8-magenta/ IR issues >> PME <<
from now on : the >> Pascal- Méheut- Effect <<.

A credit to (one of) the person(s) who detected this new species ...


Best

[Jamie Roberts]

Sean--

In a raw workflow, the worst of the magenta problem can be handled by a new C1 profile.

This doesn't let anyone off the hook, but it's true. I took many tungsten ISO 2500 shots yesterday in a dark bar... the bartender wearing black. C1 m8 profile=magenta. Tweaked C1 profle= black.

I'll be happy to send you the most recent version of the tweaked one I'm using. A couple of people have been using it and so far like it quite a lot.

I'm travelling this week, and so can't tweak any further (no color management on the laptop). But once I get a DCC and have the time, I really do believe this will solve most of the problem.

It evidently works in LightRoom too.

As for streaking, that's something I've seen at all ISO, but mostly at higher ones.

The camera itself is fabulous. Files are gorgeous and can be neutral (that's what I want from RAW anyway--zip is for finished files), and sharp as razor blades, wonderful to use, discreet as heck.

[chris_tribble]

Sean - on your side!

As Pascal says, 20/20 research hindsight is so easy when you're reviewing earlier data to find evidence of a problem that's been surfaced since the initial study was done.

I'm still waiting for my M8 and plan to hold on. What ever the problems it's still going to be easier to work with than the days when I was trying to do documentary work on Provia 400 in mixed light and then hand correcting every scan in Photoshop.

Filters? Not a big deal for me - all my lenses have had UV or protect filters on since I got them - I can't afford to replace front elements and I have to take my kit to fairly difficult places...

Best

Chris

[smsmd]

Hi Sean,

Did you use the M8 during your work over the weekend? If so, did you use the IR filters?

Thanks, steve

[chris_tribble]

Originally Posted by Jamie Roberts

Sean--

In a raw workflow, the worst of the magenta problem can be handled by a new C1 profile.

This doesn't let anyone off the hook, but it's true. I took many tungsten ISO 2500 shots yesterday in a dark bar... the bartender wearing black. C1 m8 profile=magenta. Tweaked C1 profle= black.

I'll be happy to send you the most recent version of the tweaked one I'm using. A couple of people have been using it and so far like it quite a lot.

I'm travelling this week, and so can't tweak any further (no color management on the laptop). But once I get a DCC and have the time, I really do believe this will solve most of the problem.

It evidently works in LightRoom too.

As for streaking, that's something I've seen at all ISO, but mostly at higher ones.

The camera itself is fabulous. Files are gorgeous and can be neutral (that's what I want from RAW anyway--zip is for finished files), and sharp as razor blades, wonderful to use, discreet as heck.

And hooray for Jamie's comment. My Photoshop / C1 guru was amused by some of the comments on IR that he's seen on the discussion -- his main reaction was "So you've got more data coming in than you expected to get. Why complain? It's digital - it's only numbers so there has to be a way of using software profiling to fix it".

Interested to learn more on the outcomes of your work with C1 once you feel it's useful to share.

Thanks

Chris

[sean_reid]

Originally Posted by Jamie Roberts

Sean--

In a raw workflow, the worst of the magenta problem can be handled by a new C1 profile.

This doesn't let anyone off the hook, but it's true. I took many tungsten ISO 2500 shots yesterday in a dark bar... the bartender wearing black. C1 m8 profile=magenta. Tweaked C1 profle= black.

I'll be happy to send you the most recent version of the tweaked one I'm using. A couple of people have been using it and so far like it quite a lot.

I'm travelling this week, and so can't tweak any further (no color management on the laptop). But once I get a DCC and have the time, I really do believe this will solve most of the problem.

It evidently works in LightRoom too.

As for streaking, that's something I've seen at all ISO, but mostly at higher ones.

The camera itself is fabulous. Files are gorgeous and can be neutral (that's what I want from RAW anyway--zip is for finished files), and sharp as razor blades, wonderful to use, discreet as heck.

Hi Jamie,

I agree the camera is excellent in many respects. I've gotten several e-mails about profile suggestions and will play with them as soon as time allows. Thanks for the additional ideas. Later today, I should be able to start editing the wedding pics I did this weekend with the M8 and 486 filters.

Cheers,

Sean

[innerimager]

Thanks Sean. I did ask you to comment, and I am totally accepting of your statement and believe that it puts to rest the issue of whether you, at Leica's request, held back information. I trust that all interested readers will agree and hope that my question just gave you the opportunity publically to explain what you put forward in your excellent review so there will be no doubts.
best....Peter

[sean_reid]

Originally Posted by smsmd

Hi Sean,

Did you use the M8 during your work over the weekend? If so, did you use the IR filters?

Thanks, steve

Yes, the two M8s were my primary cameras and both had lenses with 486 filters. I'm sorting out some computer problems and then should be able to start coverting files this afternoon.

Cheers,

Sean

[rvaubel]

Originally Posted by chris_tribble

And hooray for Jamie's comment. My Photoshop / C1 guru was amused by some of the comments on IR that he's seen on the discussion -- his main reaction was "So you've got more data coming in than you expected to get. Why complain? It's digital - it's only numbers so there has to be a way of using software profiling to fix it".

Interested to learn more on the outcomes of your work with C1 once you feel it's useful to share.

Thanks

Chris

Its theoritically true that you can build a virtual filter in software that is as good as a physical filter. However, it isn't easy or even possible if your profiling software doesn't control enough variables. ARC is pretty good about that, so I have high hopes that a software filter can be developed to do the job. Some people are already having success with this approach.

Rex

[Leica DeOcean]

Sean:

I am completely new to the Leica world, especially as an M8 owner. I have seen the magenta color problem in the pictures I have taken. I'm wondering if you have looked at the same effect in the M8 when it is set to B&W ? I noticed that when I took a picture of my black equipment case, it was "magenta". Then my curiosity led me to switch the M8 to B&W and take the same picture under the same circumstances and I realized that the Black equipment case was showing as gray.

I am hopeful that this will be resolved when the "color" issue is resolved as well.

Any thoughts?

[guy_mancuso]

Originally Posted by rvaubel

Its theoritically true that you can build a virtual filter in software that is as good as a physical filter. However, it isn't easy or even possible if your profiling software doesn't control enough variables. ARC is pretty good about that, so I have high hopes that a software filter can be developed to do the job. Some people are already having success with this approach.

Rex

I am still a little skeptical about a software fix. The basic issue is too much IR coming in and that needs to be blocked from any light source. And that will vary. Just not sure how you can remove the IR in software , well you can't and i see this as a fix to a certain degree but you are affecting the red channels. Please not trying to discourage anyone from trying this but just having a hard time wrapping this around my head at the moment. It certainly will help in some area's but not as effective i would not think than just block out the IR to begin with. Okay back to work but Rex and Jamie please keep going on this, it s very interesting

[rvaubel]

Originally Posted by guy_mancuso

I am still a little skeptical about a software fix. The basic issue is too much IR coming in and that needs to be blocked from any light source. And that will vary. Just not sure how you can remove the IR in software , well you can't and i see this as a fix to a certain degree but you are affecting the red channels. Please not trying to discourage anyone from trying this but just having a hard time wrapping this around my head at the moment. It certainly will help in some area's but not as effective i would not think than just block out the IR to begin with. Okay back to work but Rex and Jamie please keep going on this, it s very interesting

Your right because the profiling software that is available can only block in the red channel i.e. the color slider doesn't control into the infrared. So, given the software we have at our disposal, we can only effect an approximation, but an approximation that is more than adequate. Maybe

But if the software solution were impossible, you wouldn't see success stories where a previously purple "black" became a true black without injecting an overall color cast of it's own.

Unfortunately I have to work for a living, so later

Rex

[guy_mancuso]

Me too Rex but i like the idea a lot and maybe part of that will be done from Phase

[chetccox]

I agree, Guy. I think that a software profile may correct a magenta tint, but it cannot possibly remove the effects of IR without affecting the rest of the colors. If one had an instrument to measure the infrared emmissions from every source of light in the original picture, a software program could be written to remove the IR from the picture. This could be done by having two sensors recording the exact same picture, one with normal light and one only measuring IR. The easy solution, and the only one available for the M8, is a filter. If one could write a software program to remove the IR, and only the IR, canon, nikon, and everyone else would be able to remove their ir filters and one would make millions. It simply cannot be done. The filter has the added advantage of removing the blur caused by the different focus for the IR light. Try removing that in software.

[sean_reid]

Originally Posted by Leica DeOcean

Sean:

I am completely new to the Leica world, especially as an M8 owner. I have seen the magenta color problem in the pictures I have taken. I'm wondering if you have looked at the same effect in the M8 when it is set to B&W ? I noticed that when I took a picture of my black equipment case, it was "magenta". Then my curiosity led me to switch the M8 to B&W and take the same picture under the same circumstances and I realized that the Black equipment case was showing as gray.

I am hopeful that this will be resolved when the "color" issue is resolved as well.

Any thoughts?

This will be one aspect I look at in an article I'm working on this week. It does change the tonality in B&W but not necessarily for the worse.

Cheers,

Sean

[johnll]

Software doesn't receive any IR-specific data. All it gets is the intensities at each photosite, and knows what sort of filter (RGB) was at each site. It may conceivably be provided with the IR characteristics of each of these filters. Presumably there might be quite a bit of IR coming through the red filters, probably less or even none coming through the green and blue filters. In the demozaicing, it *might* be possible to isolate the IR components to some extent that way, but it would probably only be a partial fix. I think the IR issue is going to need hardware modifications - probably to the sensor cover glass.

[rosuna]

I have a M8 and I will keep it.

Leica has suggested a possible solution: a IR filter for the lens and a firmware update. That is OK for me.

Leica could develop a special (softer) filter, because the IR filter on the sensor filters some IR light. This specially made filter will have less vignetting or resolution consequences than other external filters if Leica designs it thinking on the M8. It is not a general purpose IR filter, but a IR filter for lenses to be used with the M8.

I think Michael Reichmann and Sean Reid made a great job. I saw the pictures and did not notice the cast.

Anyway, the cast appears randomly in my own pictures, but I can reduce it by adjusting the colors in Capture One. I do many B&W pictures. The external filter will improve the final result for color pictures.

[steve_l]

sean-
It doesn't just change the tonality, it adds an IR haze onto any area of the picture that has heavy IR, due to the different focus points of IR and visible light...I saw this clearly in Guy's pictures....foliage in sunlight is much less sharp without a filter due to this, as is anything that shows the color shift....so a software fix won't do it....