M8 - Moire and Chromatic Aberrations?

[Gabriel M.A.]

OK, I downloaded Digital Outback's DNG file, used Adobe Raw (3.4), and like any of us who actually know something about digital workflow and the problems with CA in difficult situations, set the CA-correcting parameters. If it hadn't been for a few interruptions, this would have taken me about a minute rather than about five.

After a few steps, I uploaded. Moral of the story: don't just trust anybody's files to show what an instrument can do.

Sheez, I mean, really. Let the guys ship out the thing.

 

[ywenz]

Okay, compare apples to apples, to get true idea on what the output from the sensor is like. Take un-CA corrected M8 file and compare to un-CA corrected files from other cameras. Sheez, I mean really, that's just common sense...

 

[dcsang]

Hey... I've given enough hints in the initial post that you should know who the person is who's got the M8.

As for my "very reliable source".. I also stated that I personally have not seen the files myself but hope to do so this week.

Don't shoot me... I'm repeating what I was told and am disappointed to hear it (if it is true) as I would prefer to see the M8 succeed.

Purple fringing and moire I can sort of live with (and avoid hopefully) but I remember what the fiiles were like out of the Kodak 14n - They were craptacular for the amount of money that camera cost and were completely unusable beyond ISO100 - THAT is the thing that worries me the most wrt the M8 - Not even being able to have a usable ISO400 or higher would make the camera nothing more than a really good "daylight" cam.

Dave

 

[jaapv]

Originally Posted by jaapv
I always have problems with " an unnamed friend who is supposed to be an expert says..." type of reviews. I tend to believe guys like Sean Reid, Phil Askey and others...

yeah me too ... the word of Rumsfeld, GWBush and the other "established" are water on my tongue for me ...

I mean, really how could they be wrong, millions of people listen to them... and ahhh, I also believe everything I read in the newspapers ....

Well, to be really truthful, the only one I really believe is myself- I'll let you know as soon as I have played around with my own M8 -if you promise you'll believe me 😉 But you're right in a way- Deep Throat was more reliable than Nixon. Fortunately camera-reviewers are not politicians 😀

 

[x-ray]

I would certainly expect moire from a camera with no AA filter. This what the AA filter reduces. Without it your image is sharper but at the expense of moire. The chromatic aberations are just a part of lenses, even leica lenses. If you're a digital shooter get used to it. It's easy to remove the effects of chromatic aberations in photoshop in both the raw converter and under the filters / lens function. It's no big deal. Without the AA filter you will have purple artifacts around branches of trees / fine details and particularly when back lit. Welcome to the digital world. Nothing against the M8 but it's just a fact of digital life.

 

[Gabriel M.A.]

Okay, compare apples to apples, to get true idea on what the output from the sensor is like. Take un-CA corrected M8 file and compare to un-CA corrected files from other cameras. Sheez, I mean really, that's just common sense...

Sheez, I mean really, I'm like, totally game:

Both 100% crop images from non-CA-corrected 20D RAW images.

 

[ywenz]

badass gab, badass..

 

[jaapv]

I think you've proved your point, Gabriel. And - may I add - these shots also prove the limited -to put it mildly- value of this kind of pixel-peeping.

 

[Gabriel M.A.]

I think you've proved your point, Gabriel. And - may I add - these shots also prove the limited -to put it mildly- value of this kind of pixel-peeping.

Pixel-peeping is inevitable. But, yes, it's so open to interpretations, who uses the camera, how, then the files. The 20D is fine, but it has its limitations. So is the 5D, but the CA still annoys the hell out of me; without any filter on the M8 whatsoever, I'm sure it's going to give a lot of ammunition to people who have no clue what that means. I believe this is why Leica went with the lens-coding route, so that there would be some "automatic" CA correction. A "software" filter. But on the Canons, it's there and that's that; final. Finito.

I am more worried at this point at the wild differences of the high ISO noise performance so-called reviews so far; not what I saw myself. So I don't know what's going on with these people's cameras and/or settings.

 

[John Camp]

Purple fringing and moire I can sort of live with (and avoid hopefully) but I remember what the fiiles were like out of the Kodak 14n - They were craptacular for the amount of money that camera cost and were completely unusable beyond ISO100 - THAT is the thing that worries me the most wrt the M8 - Not even being able to have a usable ISO400 or higher would make the camera nothing more than a really good "daylight" cam.

Dave

I had a couple of Kodaks, and in most ways, short of high ISO shots (say, up to about ISO 400), they were as good as a Canon 1Ds2 -- and the color was better. They are still used a lot in studio work, where, with controlled lighting, they *are* as good as the top-end Canon, IMHO. Landscape guys liked them, too, for the detail. But you did have to spend some time learning how to use them. And, with Kodak's multiple firmware upgrades, they kept getting better.

Their reputation, again, IMHO, was generally pulled down by Canon trolls; you had to be there to understand how ferocious that assault was. People who actually used them, though, tended to think they were prretty good for their time -- I mean, how many people make their livings shooting black branches against a bright sky? Reminds me of the 'banding' assault on the D200, with the lightbulbs shot against black backgrounds.

As for the purple blooming or fringing in Uwe's shot, it looks to me like he was trying to hold some detail in the white plaster surround on the building, while also trying to hold detail inside the restaurant window (notice the almost-black leaf reflections in the window, and the detail in the salt shaker.) The white plaster was essentially selected as his highlight. So you get the purple blooms on sun reflected off chrome, which is what, 4 stops higher than the white plaster? Six stops? Tell me about a DSLR of any make that will hold that, without doing something weird.

And (psst) is you have *any* competent post-processing software, it cleans right up.

JC

 

[Bob Ross]

This is the thing that stuns me each time I play around with these M8 files - the speed and ease with which they post-process.
Starting with scanned film -I lay no claim to great expertise, but I think I'm no beginner either- let's just say life is too short... Before they are halfway decent I need about every trick in the book.
Then my Canod 10D, which taught me the basics:Convert from Raw after slamming all the automatic adjustments to zero, do levels and curves, contrast,finetune colour balance, complicated sharpening, 10 mins per file, then Digilux2, which taught me the ins and outs of noise reduction, but was better:5 minutes. And now the M8: Open in lightroom, a few clickes, export to PS, crop, a bit of Focal Blade, done in 60 seconds, perfect file ready for print. I love it!

Hi JAAPV,
I have to agree with your observations. It may be the Kodak sensor family trait, since my E-1 is the same and even pro post processors agree. I am glad to see it with the M8. My three years with the E-1 have been a challenge to my compulsion to post process learned from film scanning and fooling with my first digicam, a Coolpix 5000. Hopefully the M8 will add another layer of compulsion therapy😉
Bob

 

[John Camp]

I am more worried at this point at the wild differences of the high ISO noise performance so-called reviews so far; not what I saw myself. So I don't know what's going on with these people's cameras and/or settings.

As far as I know, there have only be two reviews (or one and a half -- Sean Reid's review, and Uwe's partial reviewe on Outback.) And I think they pretty much agree that high ISO capture is good. Sean Reid says that the ISO is conservatively rated with the Leica, and that the 1250 setting is actually closer to a real 1600, and that the 2500 is actually closer to 3200. And the 1250 [1600] Reid says is pretty good, and I get the impression that Uwe also thought it was all right -- he says it's very usable and the color noise cleans up well, leaving good detail. Things start to break down at 3200 in Reid's review, but he specifically said that that was under the worst kind of mixed-light shooting conditions -- and even that cleaned up pretty well.

I think it's going to be fine.

JC

 

[Peter Klein]

Klein's First Law of Pixel Peeping states that "Sufficiently magnified, all photographs look terrible." I've learned that pixel-peeping is counterproductive unless you qualify it by how big you're blowing something up. Otherwise, it's like looking at a Monet painting from four inches away and declaring that the painter had sloppy technique.

So I made some prints of Uwe's file, a sun-lit shot at ISO 160. I opened the DNG file in Picture Window Pro's RAW converter, which is a fairly generic tool. I then used some fairly standard adjustments--just curves and a bit of USM. I easily got a file with levels and colors very similar to Uwe's JPG examples.

I then made letter-size prints of the file, both in color and converted to black and white. I also made a couple of 5x7 prints of sections of the file, blown up to the magnification of a 16x20 print.

At letter size, I just barely see the purple fringing on the car's chrome trim. It only bothers me if I "smell" the print. On the B&W print, it just looks like normal specular-highlight glare. At 16x20 magnification, the fringing is more bothersome. I also see some color artifacts at boundaries between the white plaster and darker shadows. There are also faint reddish rings around the number "160" on the restaurant door.

If I pixel-peep at 200%-400% magnification, I see significant color speckling in the gray molding that runs across the upper part of the picture just above the door. This appears to be not moire, but Bayer-pattern artifacting on tiny flecks of wearing paint on the molding. It's barely detectable at 16x20 magnification. I can also see an occasional false-color speckle on the plaster dome above the door, but this is only visible on the screen at 200% and larger--not in the final print.

To which, I say a rousing SO WHAT? The only extra work I would have needed on this file to print at letter size in color would be on the sun-reflection highlights. Most cameras, even the best, have this issue. Heck, I've gotten such fringing on such highlights on film scans. It cleaned up with a minute's work with a quick mask and the Moire reduction transformation in Picture Window Pro. For a 16x20 print, I would have needed to do the same thing on the dark grey molding above the door. For B&W, I wouldn't have needed to do anything extra at all.

It's possible that the Leica-tailored edition of Capture One supplied with the camera deals with these issues in the converter.

Apart from the above, the general impression of quality is pretty astounding. Sean Reid is right on--the prints indeed look like they were made from medium format negatives. The 16x20 magnification prints have unbelievably subtle details that I have not seen on samples from other digital cameras save the 1DSMkII. And unlike most cameras, these details hold with magnifications up to the point where you can see the individual pixels. With most other DSLRs, details "smear out" before you can see the pixels, probably because of the AA filter. If a few color artifacts represent the price one must pay for that level of detail, it is a worthwhile trade.

Despite what some less-informed people on digital camera forums think, there are laws of physics and some real-world trade-offs to be made in camera design. Based on what I've seen so far, I think Leica made some *very* good decisions on this camera.

--Peter

 

[x-ray]

Over the seven years that I've been shooting pro digital equipment in my studio the files have progressively been getting easier to post process. Since the beginning when I dhot the original D1 Nikon both software for post processing and the internal software in the camera has improved. Remember when you shoot raw the camera is not applying any adjustments to the file but the camera has embedded that data into the file so the post processing software knows what Leica, Canon or your preferences were when shot. When you post in some software that data is seen ans applied to the image unless you over ride it. This is the beauty of raw vs jpg or tif. You make the decisions on perameters. If the software does not read the embedded information then the designers of the software make that decision but again you can over ride it. My first D1 was a job to get right, the D1x, 1D and 1Ds canons and now the 1DsII all have become much easier to process. I don't think it's so much a function of the chip as I do the internal software and processing software.

Chromatic abberations and all the other little things that go along with digital files and pixel peeping are just a fact of life. Leica makes great glass but Canon and Nikon do too. All lense makers have chromatic aberations in short lensed, even Leica and Zeiss. When you look at a digital file like in the 1DsII at 100% you are effectively looking at a nearly 70 inch print at a foot or so. If you shot film and enlarged the image to that size you would see the same problem. How many times have you made prints thirty or forty inches in size and looked at the edges at four inches? It's not a function of digital but digital has given us the ability to examine the image closer than we ever would have in the past. Figure the magnification of a 1DsII file at 70 inches.

 

[newyorkone]

Pixel-peeping is fine but somtimes I think people can't see the forest for the trees...

Here's a 100% crop from the top-left corner from the original DNG saved to JPG. No sharpening at all. I think it looks pretty good and keep in mind that it's a corner no less. Of course the raw file looks even better and the shadow pixelation is from saving to JPEG.

EXIF information indicates that a coded 50 summicron was used and was shot wide open at f/2...

 

[Stan98103]

EXIF Data

EXIF Data

I believe that the EXIF data can only tell what lens is on the camera (and the maximum apeture for that lens), not the actual apeture for a particular shot.

Cheers,

Stan

 

[butter71]

I believe that the EXIF data can only tell what lens is on the camera (and the maximum apeture for that lens), not the actual apeture for a particular shot.

the EXIF data is "Max Aperture Value" with a value of "f/2.0"

regardless, i thought that one function of the blue dot was to allow for an aperture approximation for the EXIF data?

 

[Stan98103]

EXIF data and Blue Dot

EXIF data and Blue Dot

The blue dot is an ambient light sensor, but it's only used to control the brightness of the LEDs in the viewfinder. It has no effect whatsoever on how the image is captured or processed or and it does read or estimate what apeture is being used.

If you open an M8 dng file in Bridge or PSCS2 and then go to File --> File Info and look at Camera Data 1, you can see which common parameters are recorded in the EXIF data. Additional EXIF data is found in Advanced under EXIF Properties and includes Exposure Bias Value (compensation) and white balance settings, as well as some other parameters. This includes an entry for exif:ImageUniqueID which is a 32-digit number that looks like a shutter release counter. You may have to have this reset before you take your 1 x 10^33 image on the M8. 🙂

Cheers,

Stan

 

[butter71]

The blue dot is an ambient light sensor, but it's only used to control the brightness of the LEDs in the viewfinder. It has no effect whatsoever on how the image is captured or processed or and it does read or estimate what apeture is being used.

that disagrees with the latest LFI (p28) - "This sensor measures the light independently from the one inside the camera. The aperture value is then simply calculated by subtracting the light travelling through the lens from the exterior lighting" [...] "the values for aperture and focal length are subsequently stored as Exif info in the image file"

 

[Nachkebia]

Oh my, IT talks are here! get ready for transformation forum! from photography to computer hardware! 😀