DP1 Review on DCR Compares Leica M8 to DP1 on page 20

[barjohn]

Advertisement (gone after registration)

See the following link for a comparison of the M8 against the DP1 in color resolution tests. It points out a weakness in the M8 that might explain some of the color issues occasionally encountered. Sigma DP1 Review: 20. Compared to...: Digital Photography Review Very interesting.

[Advertisement]

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

[blakley]

That's really interesting. I'm trying to figure out exactly what's going on at the sensor here. If you download the file & look at the individual channels, the two center panels (red/blue and magenta/green) are very unsharp in the red channel, but all other panels are sharp in the red channel.

 

All panels are very sharp in the green channel except magenta/green, which is slightly unsharp.

 

In the blue channel, the lower left panel (yellow/magenta) is extremely unsharp with clear haloes around the magenta areas. The upper left panel (cyan/yellow) is somewhat unsharp, lower middle (magenta/green) is very unsharp, upper middle (red/blue) is slightly unsharp, and upper right (yellow/blue) is again extremely unsharp with haloes around blue areas.

 

In my photos with filter-less lenses on the M8 I've observed IR sensitivity in the red and blue channels but not so much in green. The performance illustrated in the review (out-of-focus performance in areas with lots of red reflectance but not in areas with primarily green reflectance) would make some sense if the M8 had been shot in IR-rich (e.g. tungsten) light without a UV/IR cut filter; focus problems could be attributable either to the fact that IR doesn't focus at the same distance as visible light, or to halation due to a high intensity of red and infrared light compared to exposure of other parts of the scene.

 

Unfortunately the review doesn't identify either the IR content of the ("daylight simulation") light source or whether a UV/IR cut filter was used on the M8. It's also not clear what the IR reflectance properties of the color target panels are.

[AlanG]

...Unfortunately the review doesn't identify either the IR content of the ("daylight simulation") light source or whether a UV/IR cut filter was used on the M8. It's also not clear what the IR reflectance properties of the color target panels are.

 

DPReview certainly is aware of the IR issue. It looks as if all of the test charts were adjacent to each other when photographed. So IR focus shift shouldn't be an issue. Plus the colors look good, so again, IR shouldn't be an issue. Maybe someone needs to shoot the same blue/red test chart and vary the lens focus to see if this does make a difference.

 

Keep in mind that Bayer sensors typically have twice as many green pixels as red or blue ones. So a red/blue test shot uses fewer pixels than would be used in an other color combination. But that doesn't explain why other cameras that have Bayer sensors don't show this problem.

 

I think the issue probably is a result of how the M8 processes color information. It seems to me if focus were an issue, this would show up on the b/w test chart as well. (Doesn't that use all of the pixels?)

 

What is interesting to me is that DPReview included test shots from an M8. Does that mean they are planning to finally release a thorough review of the camera?

[jaapv]

Somehow this doesn't tally with either the consensus on the resolution/sharpness nor the colour rendering of the M8 sensor. I wonder what is going on. It may well be that Blakey's post is part of the equation. Somebody should repeat this test with and without IR filter.

[vanhulsenbeek]

What is interesting to me is that DPReview included test shots from an M8. Does that mean they are planning to finally release a thorough review of the camera?

 

 

That review has been there since July 2007:

Leica M8 Review: 1. Introduction: Digital Photography Review

 

What I would like to know is if Dpreview used out-of-camera JPEG's or worked from RAW files from the respective camera's. And if so, how were these converted?

 

And would that all that have made a difference?

 

And indeed, as Jaap points out, were IR filters used?

[barjohn]

I contacted DP Review to get answers to the questions posed above. Their response was:

No IR filter used and JPG used from camera.

 

I sent a new set of questions. 1. Light source and color temperature and 2. Why no IR filter given M8s over sensitivity to IR.

[jaapv]

Advertisement (gone after registration)

Well, John, you've cleared up the mystery Maybe you should have asked about firmware as well, given the dodgy response to colour shifts by the old firmware.

[AlanG]

That review has been there since July 2007:

 

You are right. I guess I was still asleep this morning when I wrote that. (I was thinking of the preliminary review.)

 

In any case, this color resolution chart, IR, filters, jpeg, raw conversion, firmware all go to show how many factors can come into play when testing a camera.

 

I guess what is hard for me to understand is if the this test did have significant IR radiation hitting the target and that does cause the red channel to be somewhat out of focus, then why don't we see some signs of that on the black and white test chart when examining the Red channel only? Doesn't that also reflect IR? If anything the colored areas would absorb some of the IR regardless of their color whereas white would reflect the most IR.

 

It may come down to how the processor in general or in camera jpeg in particular processes detail from various colors.

[blakley]

I emailed Lars Rehm and he confirmed that there was no IR filter used; he thinks, however, that this should have made no difference. I replied that I'd love it if he could repeat the test with a filter to confirm his conclusion empirically.

 

I don't understand why Alan thinks IR focus shift shouldn't be an issue in this test.

 

If you have a panel which reflects no IR (e.g. a green panel made with non-IR reflective paint) and a panel which reflects a lot of IR (e.g. a red panel made with paint which does reflect IR), both at the same distance from the lens, then the image of the green panel will reflect only properly-focused light while the red panel will reflect both in-focus and out-of-focus light - and will appear to be less well focused.

[AlanG]

I emailed Lars Rehm and he confirmed that there was no IR filter used; he thinks, however, that this should have made no difference. I replied that I'd love it if he could repeat the test with a filter to confirm his conclusion empirically.

 

I don't understand why Alan thinks IR focus shift shouldn't be an issue in this test.

 

If you have a panel which reflects no IR (e.g. a green panel made with non-IR reflective paint) and a panel which reflects a lot of IR (e.g. a red panel made with paint which does reflect IR), both at the same distance from the lens, then the image of the green panel will reflect only properly-focused light while the red panel will reflect both in-focus and out-of-focus light - and will appear to be less well focused.

 

That may be possible but rather unlikely as none of the dyes or pigments are close to pure. (As you can see when splitting the RGB channels.)

 

I can't speak for the pigments or dyes on each color section of the panels. I am simply comparing the red/blue to the white/black panel to keep it simple. In this case I don't see why the white target would not reflect more IR than the other targets. No kind of dye or paint will add IR reflectance. They can only absorb it or not.

 

One way that invisible radiation gets turned into light is through irredescence. But I am only familiar with this happening from UV sources not from IR. (Maybe this is possible.) But even if that were the case, it would be in focus as it would be within the "light" frequencies that the lens is designed for.

 

So in the case of the red/blue target, the red part absorbs mosly green and blue, while the blue part absorbs mostly red and green. (I say mostly because these are not pure sharp cutting filters.) I have no way of knowing how much IR radiation they absorb. But I don't see any way that this target will absorb less IR than the white target. What could prevent that white target from reflecting IR? Is there some kind of special IR absorbing white paint on it?

 

For instance, if there is a secondary IR out of focus image recorded in the red channel, why isn't the white target also blurred when viewing only the red channel? It seems to me that each split color channel represents a fair amount of image processing and interpolation - not just what was originally recorded by the red, green, and blue, pixels individually.

 

I believe it has to come down to how the colors get processed. Although I'm open to other explanations.

[SJP]

Alan, I think I agree with you - a B&W test chart should cover all options. Basically your are testing all colors at once.

 

The fuzzy images in the test for me are close to being incredible, are they really sure there was no focus error? The output quality of the M8 would be inferior if this test was realistic (and we know M8 isn't inferior in image quality certainly not at the lower ISO settings). Did they use DNG or jpg?

[jaapv]

All in all, we can agree that there is something strange going on here that is not consistent with the facts. If my M8 did that, it would have hit e-bay within the month, if I would not have returned it earlier. Stuff like that must be very detrimental to image quality.If not, the test is meaningless.

[mjh]

That's really interesting. I'm trying to figure out exactly what's going on at the sensor here. If you download the file & look at the individual channels, the two center panels (red/blue and magenta/green) are very unsharp in the red channel, but all other panels are sharp in the red channel.

 

All panels are very sharp in the green channel except magenta/green, which is slightly unsharp.

 

In the blue channel, the lower left panel (yellow/magenta) is extremely unsharp with clear haloes around the magenta areas. The upper left panel (cyan/yellow) is somewhat unsharp, lower middle (magenta/green) is very unsharp, upper middle (red/blue) is slightly unsharp, and upper right (yellow/blue) is again extremely unsharp with haloes around blue areas.

I don’t think IR has got anything to do with it. As has already been observed, this would affected the b&w chart as well, but in fact, this is the sharpest of all the six charts, even in the red and blue channels.

 

Presumably, this is just an effect of the demosaicing algorithm used. Given that only 25 percent of all the pixels are sensitive to red (or blue), one should expect the red channel to be blurry. However, any reasonably sophisticated demosaicing algorithm will use the data from the higher resolution green channel to improve the sharpness in the other two channels. Whenever there is a change in the amount of red from one red-sensitive pixel to the next, and a change in the amount of green at roughly the same spot, the algorithm will assume that the gradient for the red channel is just as steep as that for the green channel, even when the red-sensitive pixels are spaced wider apart. Obviously, this works best for b&w patterns where changes in the amount of red, green, and blue are perfectly correlated. It still works when there is sufficient contrast in the green channel that corresponds with the patterns in the red and/or blue channels, but when even the green channel fails to provide the necessary detail, the M8’s demosaicing algorithm apparently resorts to a conservative approach of merely interpolating the missing values in the red and blue channels. In fact, the differences in sharpness observed in the red and blue channels are just a more extreme version of what we see in the green channel: good contrast in the green channel enables the demosaicing algorithm to make the red and blue channel resolution to look better than it actually is, but when even the green channel fails, we see what resolution for red and blue is really like with a Bayer sensor.

 

Keep in mind that Bayer sensors typically have twice as many green pixels as red or blue ones. So a red/blue test shot uses fewer pixels than would be used in an other color combination. But that doesn't explain why other cameras that have Bayer sensors don't show this problem.

Apparently, the Nikon DSLRs used for comparison in the review apply a less conservative algorithm. Looking at the red/blue chart, it seems obvious that neither the D60 nor the D40x offer better resolution, only the patterns show more contrast and appear to be sharper. This is merely what you get from sharpening. In fact, the low-contrast pattern recorded by the M8 looks cleaner that the results obtained with the Nikon models. It has to be said, though, that the sharpened look of the Nikon images is quite typical of what you get from modern DSLRs; it’s not just Nikon. Leica’s conservative approach to demosaicing is more unusual.

[barjohn]

Below is Phillip Askey's response to my question of lighting and IR reflectivity.

 

John

 

 

Not sure what you're driving at but the M8's IR sensitivity is only an issue with certain man-made materials, the color test chart doesn't reflect enough IR to be an issue. The IR filter simply would have made NO difference in this test. We used our normal daylight simulation softboxes which use high CRI bulbs at about 5000 Kelvin.

 

Does that satisfy?

[blakley]

Alan, I think I agree with you - a B&W test chart should cover all options. Basically your are testing all colors at once.

 

There's no reason to believe that's true unless we know a lot more about the materials used to make the chart. Some blacks do not reflect much IR, some reflect a LOT. This is why the "purple blacks" problem occurs on some synthetic black materials with an unfiltered M8 and not with other black materials.

 

The same is true of whites - some pigments reflect a lot of IR and others don't.

 

There just isn't enough information to know what the underlying cause of the effect is.

 

I'm skeptical that it's the demosaicing algorithm, though, because I've not seen this effect in my own photos in situations with contrasting colors (including reds and blues).

[blakley]

Below is Phillip Askey's response to my question of lighting and IR reflectivity.

 

John

 

 

Not sure what you're driving at but the M8's IR sensitivity is only an issue with certain man-made materials, the color test chart doesn't reflect enough IR to be an issue. The IR filter simply would have made NO difference in this test. We used our normal daylight simulation softboxes which use high CRI bulbs at about 5000 Kelvin.

 

Does that satisfy?

 

I guess I'd still like to see the test performed with a filter, to replace opinions with facts.

 

If the problem isn't IR, it will still be interesting to know what it is.

[barjohn]

I agree. I think I will ask Phil if he would repeat the test with a filter and make the raw files available.

 

John

[c6gowin]

It would be great if the test could be re-performed using a filter. However, I agree with JAAP, if the real world results with the M8 were as bad as the test photos indicate then it would be very obvious. I am still very happy with the results from my M8 and they don't look anything like the DPR test patches.

[Advertisement]

Hallo,

Look here: Leica M

[AlanG]

It would be great if the test could be re-performed using a filter. However, I agree with JAAP, if the real world results with the M8 were as bad as the test photos indicate then it would be very obvious. I am still very happy with the results from my M8 and they don't look anything like the DPR test patches.

 

Keep in mind that the fuzzy test result was mostly a problem only on the red/blue color combination when shooting jpegs. And even with that combination, it may not be apparent unless you are shooting a test target under circumstances that would provide maximum resolution. And then only if you really scrutinize the image. So that is a pretty small percentage of photos in the "worst case" scenario.

 

What is apparent to me from most DPReview camera tests these days is how good an image is possible from almost every camera out there. But what you enjoy using, what you can afford, or what works best for you in a specific or in general situations is another matter.

[AlanG]

I guess I'd still like to see the test performed with a filter, to replace opinions with facts.

 

If the problem isn't IR, it will still be interesting to know what it is.

 

You could do your own test and see if you can duplicate the issue - with and without a filter. Then also try shooting raw and run the image through a few converters.