jaapv Posted November 26, 2009 Share #21 Posted November 26, 2009 Advertisement (gone after registration) Well, to create that accidentally in post, one would have to work the TIFF in 8 bits. I could see the effect in RAW conversion, so it is in the file. Link to post Share on other sites More sharing options...
Advertisement Posted November 26, 2009 Posted November 26, 2009 Hi jaapv, Take a look here Rainbow skies..... I'm sure you'll find what you were looking for!
plasticman Posted November 26, 2009 Share #22 Posted November 26, 2009 Andy - I don't wanna sound like a broken record but (at the risk of sounding like a salesman for the app*) have you tried Raw Developer on the files? Here's a screenshot (which I've reduced in size) of a detail of an image shot with an unfiltered 21 on the M8 - with RawDev at default settings on the left, and C1 v4 on the right. The grab was taken a while ago as a comparison for my own purposes, and naturally it's gone through the laptop's own LCD png profile and then back to sRGB and then jpeg compression, so it really is only for general information purposes - but the way that RD renders the blue channel as against C1 should still be visible (I hope). Possibly worth a try anyway. * which - just for the record - I'm not. 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 Simply register for free here – We are always happy to welcome new members! ' data-webShareUrl='https://www.l-camera-forum.com/topic/104802-rainbow-skies/?do=findComment&comment=1131781'>More sharing options...
adan Posted November 26, 2009 Author Share #23 Posted November 26, 2009 plasticman - interesting... Check your PMs. Link to post Share on other sites More sharing options...
innerimager Posted November 26, 2009 Share #24 Posted November 26, 2009 I also use RD and love it. I tried to get sky banding yesterday pushing exposure very hard and got no banding in RD. Link to post Share on other sites More sharing options...
plasticman Posted November 26, 2009 Share #25 Posted November 26, 2009 I gave the DNG a very quick run-through RawDev - default M9 settings all the way through, then a direct 16bit TIFF export to Photoshop and there I just went through the simplest conversions: resize, resample to 8bit, convert to sRGB, export to web. The result is definitely better than the original posts, but the gradation in the sky is still visible - maybe this is one of those sorts of images that really needs to be exposed spot-on: plenty of situations where the M8 has that problem (or virtue?). Highlight Recovery did nothing to help incidentally - there didn't appear to be any color there to recover. 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 Simply register for free here – We are always happy to welcome new members! ' data-webShareUrl='https://www.l-camera-forum.com/topic/104802-rainbow-skies/?do=findComment&comment=1131884'>More sharing options...
barjohn Posted November 26, 2009 Share #26 Posted November 26, 2009 While the Raw Developer image looked better the banding is still visible. Link to post Share on other sites More sharing options...
plasticman Posted November 26, 2009 Share #27 Posted November 26, 2009 Advertisement (gone after registration) While the Raw Developer image looked better the banding is still visible. Yep - I think it's definitely irretrievable. But actually each stage in the output process (and mostly the jpeg conversion) visibly increased the posterization, so maybe a print could be coaxed from it if special care was taken. One thing to note is that my version has less contrast than Jaap's attempt - that's in the neutral nature of RD's workflow - but you can also increase contrast directly in Lab colorspace, which should help to control posterization problems. I thought it was worth a try - and I still love the RD workflow. Link to post Share on other sites More sharing options...
adan Posted November 26, 2009 Author Share #28 Posted November 26, 2009 Just for the record, the cyan right above the roofs and treetops at left in that "leg" shot is "real" - sky discoloration due to Denver's "brown cloud" of winter smog. I've gotten it in winter sky shots like this for 16 years on Velvia and K-chrome - long before digital. And the same smog probably produced the underlying (or overlying ) overall gradation in the sky, but not the banding artifacts, since they go away in the bracketed shots where I exposed correctly, or at least less. The hawk people shots were in smog-free Vermont. This is why I like being an early adopter. I'm the perfect fool to test anything foolproof. Link to post Share on other sites More sharing options...
Mokkacream Posted November 26, 2009 Share #29 Posted November 26, 2009 The bad news is: that's because the effect still shows up identically in big uncompressed DNGs. 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. Could you check the files again, please: I would guess, that the uncompressed is the compressed DNG indeed and that you accidentally inverted them. Anyway, what you see is broad banding due to high overexposure in both DNG and even exaggerated in the compressed file due to significant reduction of color depth. The bow form is due to the punctual light source, the sun. Other causes, which would exaggerate such banding, are: spreading the tonality (e.g. by adjusting "exposure", midtones or "contrast" to much) Further more the higher the ISO, the more pronounced it will be. The usable dynamic range decreases from 7 to 4 stops at ISO 1250. Anyway, this is obviously not the case, because you didn't make adjustments afterwards. I must admit, that I do not really know what happens within the camera to the RAW file, when set to "pull 80". I don't trust it. You could easily avoid this by correct exposure. You already demonstrated a much better result with setting the exposure to -1EV "underexposure" in your example. Link to post Share on other sites More sharing options...
Mokkacream Posted November 26, 2009 Share #30 Posted November 26, 2009 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. You demonstrate a very pronounced banding in your picture due to the much higher dynamic range of the nature compared to the camera's sensor Link to post Share on other sites More sharing options...
SJP Posted November 26, 2009 Share #31 Posted November 26, 2009 Maybe I am being dense but isn't this standard for all digital sensors? They have a maximum exposure (counts per pixel) and if you exceed that abandon hope. Whatever you do this sort of shot has to be exposed for keeping the highlights inside the dynamic range. After that you can play around to your hearts content in PP to pull the shadow detail or blow the highlights, or enhance dynamic range. So point straight into the sun & lock or use manual and in both cases check the histograms for blown highlights. Link to post Share on other sites More sharing options...
adan Posted November 26, 2009 Author Share #32 Posted November 26, 2009 Andreas - the DNGs are correctly labeled. Those who have worked with the "leg sculpture" shot in other RAW processors can confirm this - they had the same 35Mbyte file to work with. Note the volume and location of the car traffic in the background. Absolutely the original problem is that my shots showing this are overexposed - at least in the sky. At a time of day/year (in the N. Hemisphere) when the low sun produces the greatest brightest/color gradations. I was just noting the fact that the transitions seemed more abrupt with the M9 than with similar overexposures/skies/light with the M8. The M8 didn't seem to punish me quite as severely for errors BTW my "better result" was 0 EV correction - the offending shots were at +1/3 EV, for reasons I listed earlier. A distinction without a difference, perhaps. Definitely my M9 is no longer set to +1/3 EV. I would agree that the Pull 80 ISO is essentially also an overexposure, and thus to be used with care - like a hand grenade. Link to post Share on other sites More sharing options...
ho_co Posted November 26, 2009 Share #33 Posted November 26, 2009 Maybe I am being dense but isn't this standard for all digital sensors? My thought as well when I mentioned it above (http://www.l-camera-forum.com/leica-forum/leica-m9-forum/107297-rainbow-skies.html#post1134975). Link to post Share on other sites More sharing options...
pgk Posted November 27, 2009 Share #34 Posted November 27, 2009 Maybe I am being dense but isn't this standard for all digital sensors? They have a maximum exposure (counts per pixel) and if you exceed that abandon hope. Whatever you do this sort of shot has to be exposed for keeping the highlights inside the dynamic range. After that you can play around to your hearts content in PP to pull the shadow detail or blow the highlights, or enhance dynamic range If only life was that simple. For the scenario that you make the sensor would have a precise response with an abrupt high end cut-off. But my suspicion (based on my own observations) that at the high intensity end what actually happens is that the linearity of the sensor's response fails so any (little) recorded information is far harder to interpret - hence the banding, the variation depending on raw processor etc. This also explains how 'recovery' may to some extent work. I'd guess that each individual sensor may vay slightly in its response/abilities at extreme intensity levels which may explain Andy's observations on his M9 relative to his M8 files. Just my ideas - I'm not into electronics so I have no idea of a more technical explanations based on the way the sensors record, I just experiment with files (correctly exposed and otherwise) and try to figure out what effects are usable. Link to post Share on other sites More sharing options...
thrice Posted November 28, 2009 Share #35 Posted November 28, 2009 If only RAW Developer weren't for mac only Link to post Share on other sites More sharing options...
adli Posted November 29, 2009 Share #36 Posted November 29, 2009 I am not sure if this is an M9 only phenomenon. Here is a picture taken with my M8 and the 18mm SE, showing similar artifacts (not posted for artistic reasons, not particulairly interesting picture). No postprocessing except white balance correction and jpg-conversion in Lightroom. 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 Simply register for free here – We are always happy to welcome new members! ' data-webShareUrl='https://www.l-camera-forum.com/topic/104802-rainbow-skies/?do=findComment&comment=1134615'>More sharing options...
ho_co Posted November 30, 2009 Share #37 Posted November 30, 2009 adli, just curious-- is that from an 8-bit source or 16-bit? (and would that make any difference?) I know PS offers the option of working in 16- or 8-bit space (terminology probably wrong?), and I guess Lightroom does as well? Link to post Share on other sites More sharing options...
adan Posted November 30, 2009 Author Share #38 Posted November 30, 2009 adli - Interesting. I guess I'll just accept that I was fooled into overexposing the M9 based on what looked like darker DNGs in their default state in ACR and on the LCD. Certainly I have avoided more rainbow skies with it by moving back to 0 EV compensation, avoiding ISO Pull 80 in contrasty situations (bright sky, dark subject), and revising my use of the settings in ACR to make use of "fill light" and other tools to get the brightness back. Howard, I think at the point the raw developer is working on the picture (which is where one fixes or can't fix the exposure/skies/banding), the data is not yet defined as 8-bit or 16-bit. It is whatever the camera captures (in the case of the M8, always a 12-14-bit range compressed to 8 bits for storage by a root-2 function; in the case of the M9, either 12-14 bits uncompressed and stored within a 16-bit wrapper, or compressed to an 8-bit file that can be restored (sort of) to a 16-bit file as it is being developed in C1 or ACR or LR or whatever. But in any case, not 16-bit or 8-bit in the sense that a TIFF, say, is frozen as 16-bit or 8-bit - until it has been developed by the RAW converter according to the user preference. So the "source" would be whatever the camera's default RAW parameters/format are - neither 8-bit or 16-bit, exactly. Anyway - I'll be off line for 10 days or so. Maybe there will be new firmware by the time I get back. Have fun, folks! Link to post Share on other sites More sharing options...
Alnitak Posted November 30, 2009 Share #39 Posted November 30, 2009 I've seen this on a few shoots when I overexpose the sky, but if I expose for the highlights and bring up the shadows in post, its not a problem. I used to see a problem just like this with my Nikon D100 7 years ago. Link to post Share on other sites More sharing options...
jaapv Posted December 1, 2009 Share #40 Posted December 1, 2009 Well, working in 8 bits is not advisable in postprocessing to avoid posterizing in all cases, but I used 16 bits throughout and it did not eliminate the effect entirely. Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.