Jump to content

Rainbow skies....


adan

Recommended Posts

Advertisement (gone after registration)

I've run across several examples of banding like this in blown wide-angle skies with the M9 (35mm uncoded and 21mm coded). The camera seems a bit more prone to this than the M8, which would go from blue corners to white blowout in a smoother transition without the funky cyan and bald desaturated gray bands.

 

I suspect (hope?) it is just a defect in the A/D conversion process that can be fixed in firmware. At ISO Pull 80, I'm not too surprised - I've seen the same thing in Canon shots at the "pulled" ISOs - but at ISO 320? I can't imagine Kodak's 6.8micron pixels have changed their saturation capacity this much.

 

Although I was using +1/3 EV comp for a while, because the M9 images looked so dark on the LCD. Which is a "pull" of sorts - but only to ISO 250.

 

But I've gone back through 3 years of similar M8 shots with dark subject against bright skies, and never seen this kind of effect. Using the exact same raw converters (ACR 3.4 and 5.6).

 

Edit: This is from a compressed DNG - but the M8 always produced compressed DNGs, too. I haven't tried an uncompressed DNG in similar light yet, but will right now.

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!

Link to post
Share on other sites

  • Replies 40
  • Created
  • Last Reply
Edit: This is from a compressed DNG - but the M8 always produced compressed DNGs, too. I haven't tried an uncompressed DNG in similar light yet, but will right now.

 

Andy,

 

That would be a worthwhile test - in practice the M9's compression scheme compresses to fewer levels than the M8.

 

Regards,

 

Sandy

Link to post
Share on other sites

The good news is: I don't have to shoot big uncompressed DNG files to avoid this effect...

 

The bad news is: that's because the effect still shows up identically in big uncompressed DNGs.

 

Backlit tree - uncompressed ISO 160 dng

Leg sculpture ISO Pull 80 - as labeled (compressed and uncompressed)

 

I also included just the blue channel, which actually gets darker again in the "gray" area where it should be still getting lighter, like a partial solarization. The other interesting point in the affected leg sculpture images is that the RGB values never even get near 255 - they top out at about 220 in all channels (but in different places) and than drop back to under 200.

 

Also included is the leg sculpture shot with 1 stop less exposure - which looks just fine, and actually produces a brighter sky.

 

No "recovery" applied in the workflows.

 

I'm still thinking something awry in the A/D conversion programming/math - unless CCDs are subject to "solarizing" effects...

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!

Link to post
Share on other sites

Andy--

I was going to ask after the first shot if it had to do with the falcon, maybe he's from a different union? I think Wotan preferred ravens. (Nice, affecting shot, BTW, except for the rainbow.)

 

Interesting effect, but only if it's intentional via postprocessing. ;)

 

 

I'm glad you've tried all the scenarios. Have you contacted Leica about it? That's obviously next on the agenda, I would say.

 

Since others aren't chiming in that they're getting this as well, it may be just your camera?

 

 

Do keep us posted.

Link to post
Share on other sites

Since others aren't chiming in that they're getting this as well, it may be just your camera?

 

Well, that's partly why I decided to put this up here before rattling Leica's cage - to see if anyone else had seen this.

 

Or to see if there were any logical "user error" explanations - which is always possible, whether in exposure or in handling the M9 files in post.

Link to post
Share on other sites

Advertisement (gone after registration)

Well, that's partly why I decided to put this up here before rattling Leica's cage - to see if anyone else had seen this.

 

Or to see if there were any logical "user error" explanations - which is always possible, whether in exposure or in handling the M9 files in post.

 

This may sound like a silly question, and please forgive me if it is, but how do camera produced JPGs look? Also, is this problem apparent on the camera LCD? My thinking being that if they appear OK, it's more likely to be a post camera problem. If the LCD and JPG images are affected too, then it's more likely to be an in-camera problem.

Just a couple of thoughts. :)

Link to post
Share on other sites

I can get a similar effect on the M8 by significantly overexposing a shot and then pulling back the exposure in the raw conversion, but its never been a problem I've seen with 'correctly' or 'viably' exposed images. It is a substantial problem shooting 'sunbursts' underwater with a variety of cameras where the cyan shift is even more distinct and very unwelcome indeed (many diving magazines have examples as its exacerbated yet more in print). I've always put this down to non-linearisation of the data at high intensities (especially the blue channel) - something like 'reciprocity law failure' in film, but with very different results. Whether different raw converters handle this in different ways is not something I've looked at.

Link to post
Share on other sites

HI Andy

Can you do this with a coded 'normal' lens as well?

I only ask because it's more likely to grab attention.

 

I'm going to give it a go later on - I have to say, I have NEVER seen it, but then I'm usually more interested in the sky, and I'm careful not to blow it out!:)

Link to post
Share on other sites

Nicole: good question - the sculpted-leg image below is the fine-jpeg shot at the same time as the weird DNGs above. Blown sky, slight cyan band, but not the gray band.

 

pgk: example below of folks exercising at dawn (red shirts) is the closest I can get in an M8 shot to equivalent exposure - the M8 just seems to behave a bit more benignly, especially as regards hue (blue/cyan), when abused in overexposure (but see my other note below).

 

jaapv: I have no personal ftp address. If you know a locale where I can upload 35 Mbyte files, PM me.

 

jono: by "coded normal" I guess you mean 50mm-ish? No, I just have an uncodable 35 and a coded 75. Certainly wider lenses reveal the effect more obviously, since only they take in enough sky for there to be the wide brightness range in the blue that sets this off. I have several coded 75mm shots taken of the people with hawks, and generally the sky is just blown white - or not - in the background (near the horizon). Only one - posted below - shows some cyan bands (but not the gray). It is, BTW, totally uncorrected in ACR - just the way it opens with default settings.

 

Now that I've looked at this in more detail (thanks to your questions and Sandy's comment on dng compression), several things seem apparent:

 

1) The "straight" dngs from the M9 "look" darker than those from the M8 - on the LCD/histogram or when first opened without adjustments at the Camera Raw defaults, so I've been tending to give 2/3rds stop more exposure with the M9 (the M8 was permanently set to -1/3 EV, and I've had the M9 set at +1/3 EV for these first couple of months) - on the assumption that the dark-looking DNG originals were due to underexposure. See the 75mm shot of the woman with hawk below and you'll see what I mean about dark-looking DNGs.

 

I was afraid that boosting exposure too much in raw development would increase noise.

 

I've also been "exposing for the shadows" manually in places where there is strong contrast and the main subject is in the shade (e.g. the original hawk people shot).

 

I also started out using ACR 3.5, which does not have the "recovery" and "fill light" controls of ACR 5.6 or LightRoom, even though it reads M9 dngs.

 

Comparing the DNGs with the Jpegs, it seems that correctly exposed DNGs from the M9 do just appear darker than those from the M8 at default settings (e.g, my default brightness for M8 shots was "0" in ACR, whereas the M9 brightness routinely goes up to "+80" or more), but that there is lots of room to raise the brightness in developing using the fill-light and recovery controls, so that I can go back to no EV correction, or even the M8's -1/3 EV correction - and likely avoid blown skies in the first place.

 

2) that being said, it does seem to me that the M9's response to overexposure is more abrupt and squirrelly than the M8's. But I no longer have an M8 to experiment with...

 

3) The final sample below is the same file I originally posted, with significant adjustments in both ACR (which left the sky mostly white but mostly unbanded) and followup corrections in Photoshop, which leaves just a trace of a cyan stripe.

 

4) I do have a query in to Leica through channels, but between the US holiday and being on the road the first two weeks in Dec., it may be a while before I hear or can share any response.

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!

Link to post
Share on other sites

Andy- To the issue of how you are handling the darker DNGs from the M9. In my brief experience, it seems that mid tones and highlights are exposed correctly, but that the shadows are blocked. I have been exposing for highlights as I did on the M8, and find it easy to recover detail in the shadows whether in the raw converter or in PS after conversion using a touch of shadow opening in shadow/highlights. best...Peter

Link to post
Share on other sites

Yeah, the short version of my post above is that this may be "design-induced user error" - the darker, contrastier LCD setting of the M9 plus the clipped black point in DNGs that I believe some have mentioned led me up the path of overexposures.

 

Which just goes to show that every camera has a learning curve - even oh-so-conservative, never-changing Leicas.

Link to post
Share on other sites

I thought I would give it a go in C1V5pro, just to see. I think it reduces the posterizing considerably, but it is still there. I can provoke it by applying a large amount of highlight recovery. Maybe the reason is that the red channel blows out the highlights before the other channels, giving a green band in a gradient-like situation. Anyway, this is what C1 makes of it. Btw, there are a couple of hairs on your sensor. Those gnomes in Solms should keep their beards better trimmed....;)

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!

Link to post
Share on other sites

I picked up a similar effect with some test shots I did with the M9 & 12mm C/V lens.

This is a straight conversion from DNG to JPG no pp.

There's also some cyan drift in the bottom left of the picture. Lens not coded / coding off in camera.

 

The sun was coming up over the roof of the building on the right side of the picture.

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!

Link to post
Share on other sites

Andy

 

Herer's the nearest that I've been able to achieve (very accentuated) on the M8 but this is via ACR with -3 stops dialed in the exposure slider on a very overexposed shot (my fault entirely when 'experimenting'). The cyan is obvious but there still isn't the real banding that you showed.

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!

Link to post
Share on other sites

Paul - you think? I see a big band of gray below the cyan in your shot - broken up by cloud formations so it is not as solid a band (and the clouds pick it up all the way down to the horizon in places).

 

Runs from between the sign and cyan band on the right all the way to the purple cloud deck on the left.

 

I think if you'd had no clouds in the composition to break up the smooth sky tonal transitions, it would have been just about where mine are. Or if I'd had clouds, mine would have looked like yours.

 

Nice scene, BTW - I hope you got some exposures without the blown sky!

 

Anyway, for the moment I'm just going to knock off the + exposure comp and wait to see if I get any feedback from Leica. I think there is something slightly odd going on compared to the M8 - but since I have to screw up to see it, I'll just avoid screwing up...

Link to post
Share on other sites

Effects as the one in your first pic has occurred to me only when you "play" too much with PS... even with "odd" exposures and rather strong processing in LR I never got something similar from M8 DNGs... Could be it's your camera processor... or even the RAW developer... but I suppose that with so many M9 adopters in the forum, if someone else would have seen this with his M9, we would have seen some posts about...

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...