D200 Banding Conditions Thread

Discussion in 'Nikon DX DSLR' started by PJohnP, Jan 6, 2006.

  1. PJohnP

    PJohnP

    Feb 5, 2005
    Peter Bendheim has looked at banding conditions for the new D200 in another thread (Bendheim Findings), and shown a "recipe" for this problem.

    Peter's also rightly pointed out that Nikon should be addressing the question.

    However.

    Judging from how Nikon responded this last spring to the focus issues on the D2X, Nikin will only move with some specific findings to present to them to get this addressed (especially in light of the current battery questions, which could occupy centre stage for them). So, I'm not going to ask members of the Cafe to run a zillion tests (I, too, have paying work I need to do, and understand not being able to act full-time in lieu of a testing laboratory for Nikon). Nikon will eventually address this, but perhaps we can derive some understanding of the banding phenomenon and be able to minimise it, just as people minimised shadow noise issues with the D2H and the D100 by using proactive shooting setting(s) on those respective cameras.

    Here's what I'll ask of members of the Cafe shooting with D200 who are seeing banding issues :


    • Summarise (briefly) shooting conditions for shots with banding;
    • Save the NEFs of any significant banding phenomenon shots for examples;
    • Continue to note any additional (not yet reported, that is) conditions or circumstances for banding; and,
    • Note any changes in shooting settings or conditions that eliminate or reduce the banding effect.

    Members of the Cafe and other fora have determined some important issues in the past with respect to equipment. I'm unsure if Nikon reads the postings on the Cafe or other fora, but even in their absence, we've found some intelligent workarounds in the past that allow us to address problems. We may well be able to do this here. :Smart:

    If we develop a significant list of repeatable conditions with examples, I'll take the lead on trying to get this dataset into Nikon's hands for action or pass it on to anyone who has better contacts to do so.

    Thanks for any constructive contributions to assist the D200 owners in the Cafe.



    John P.
     
    Last edited by a moderator: Mar 15, 2017
  2. PJohnP

    PJohnP

    Feb 5, 2005
    I'll start with what I've seen as conditions for banding.

    • ISO400 or more;
    • Fluorescent light (edit - also made a shot with natural backlight and fluorescent overhead);
    • Significant overexposure next to dark area, possibly underexposed - high contrast edges; and,
    • Processing to address under-exposure in PS or NC (edit - mixed light shot has no processing applied).

    I tested with ISO100 and 200 yesterday with no banding. I used Peter's "recipe" noted in the first post to this thread. My banding appears much less than others' examples.


    John P.
     
  3. Daylight

    Poked my head out the door...very early today. Light drizzle, grey white sky,strong glare, no sky detail.

    Some test quick snaps at ISO400...just snaps with 12-24, tried to not overexpose for the sky (which had no detail) or underexpose for the shadow, f5.

    Results and crops follow...

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    Some people at other forums are saying this is a non-issue and that it can happen with any camera. Well, if you have a 10MP camera to enable you to print large, or sell for stock...in my mind this is definitely not a non issue.

    It would be a non issue for me if it only happened at very high ISO's, infrequently and under isolated lighting conditions, but this has proved in several examples from several people not to be the case.

    But it has happened at ISO200 upwards (so far) but not as yet at ISO 100. Without artificial lighting. And yes, in high contrast and difficult situations, but not at the cutting edge or envelope of current DSLR technology.

    A cheap point and shoot like my 5400 would not do this, why should a medium DSLR doing this be a non issue. Others have said it's only raw...well I have been shooting raw and jpeg and this effect is in both files.

    I can also say that this is totally unpredicatable...I took further shots in the day of the same scene as the weather was pretty much the same, and the shots were free of the cordrouy effect.

    In my past posting history at other forums, I may have minimised issues that other posters have had with their cameras based on my personal experience of one unit. I now know what it feels like when the shoe is on the other foot, so to speak, and it's been a personal lesson for me in empathy.

    Best

    Peter
     
  4. nfoto

    nfoto Guest

    I agree with you, Peter, that the occurence of banding is (almost) unpredictable. I have spent days on end trying to provoke it and largely, but not entirely, failed.

    Sadly I have data that indicates it also may be exacerbated by the processing software as well, since the same NEF in Nikon Capture or Bibble come out differently with respect to banding.

    Currently I'm collating the data I have on this issue for my review.
     
  5. Thanks, Bjorn. Although I have been shooting at RAW+Jpeg, I've only looked at the Jpegs so as to isolate the effects of the RAW processing software. The above shots are just Jpegs, with some sharpening.

    PS - Sounds like your computers are up and running again...if so I'm really pleased for you.


    Peter
     
  6. Iliah

    Iliah

    Jan 29, 2005
    nowhere
    I must concur
     
  7. nfoto

    nfoto Guest

    My computers have been up and running all the time. I run distributed net working. It was my Linux file and mail server that crashed, and unfortunately, the D200 files went into the cybervoid with it. I have been able to reconstruct about 70% of the files from other sources, though, but this debacle has put me off the D200 focus as it were.
     
  8. Gale

    Gale

    978
    Jan 26, 2005
    Viera Fl
    I see a halo from over sharpening on my monitor.
    Would not a raw file be better . With no sharpening added as well.

    I have noticed processing with NC 4.4 on the D200 files sometimes does not seem quite right. Not sure how to explain it though...

    I have not seen banding yet. I have tried to provoke it in processing with no luck with underexposed image. Shooting manual 80-400 VR, VR on..Iso 200...

    Seen nothing on overexposed images yet either.
     

  9. Yes, you are probably right, Gale. I have standard routine that I resharpen slightly with after I've convered the image for web posting and resized it.

    I haven't quite got a handle on the optimum sharpening for me of d200 files. I find each camera to be different from the D2X to the D70, and I'm still looking at an optimum workflow with the D200..if only this issue hadn't interupted my normal routines...:mad:
     
  10. caero

    caero Guest

    I am starting to wonder about this issue.

    1. It seems a lot more evident in some raw converters (Nikon Capture) than others (RawMagick etc)

    2. I haven't seen a single shot yet with these conditions that didn't have this kind of banding. Some just seem to have it more than others.

    puts on a tinfoil hat

    3. I am starting to think that this issue has a very high correlation with the very low amount of cameras that seemed to be released initially. Could it be that Nikon either:

    A. Wanted to release the camera in so little amounts, to be able to replace all of them if things weren't as they were supposed to.

    B. Knew about the banding issue and only released a little amount of D200s to see if the issue would go by unnoticed.


    I am by no means an expert on any of these things, but the whole issue baffles me that it could even come out of a factory that has some sort of QC.

    Furthermore I am annoyed because I sold my D70 and was hoping that this camera (the D200) would be the thing I was waiting for (a digital F100). This issue is really heartbreaking :|
     
  11. caero

    caero Guest

    Well I posted this at the other forum, so I might as well post it here as people seem a bit more knowledgeable.

    The below shots are 100% crops of direct conversions of a uncompressed NEF taken by a dpreview member called rfc. I did the conversions in NC and RM Lite without anything done to them (no change in sharpening settings etc etc.)

    NC Conversion
    [​IMG]

    RawMagick Conversion
    [​IMG]


    As you probably can see the RML version shows a lot less banding (some say none at all, but I am personally still able to see it slightly) than the NC version.
     
  12. PJohnP

    PJohnP

    Feb 5, 2005
    Following on a comment by Iliah, there's the issue of what fraction of the D200 series cameras are afected by this issue in a significant manner. My camera exhibits some - but not a great deal - of the banding phenomenon. This doesn't mean (Peter, pay attention ! :wink: ) that banding doesn't occur, just that the circumstances for it with my camera are much more limited. Peter and Sune's cameras seem to exhibit more of this, Bjørn's less.

    We seem to be seeing some kind distribution of the degree of the problem, although it's premature to say what, if any, mathematical distribution applies. We simply have insufficient data.

    With that in mind, and hearkening back to the fore-/back-focus issues of times past, can we develop some kind of standard image, be it a print out or what have you, that we can use with backlight conditions to further test this ? As lovely as Peter's wife is, she's not available to model around the world, and I'm certainly not shipping a trilobite specimen weighing several kilos all over either !

    While we've better determined that the issue exists, we're currently not all shooting under the same conditions. Again, as noted in the first post to this thread, we're not going to become a testing laboratory, but we do need to have a constant reference to discuss this against.

    But at this moment, I'm unsure that the reference should be.

    I'm travelling for three weeks now on business, but I'll try to think about this further, and urge other (brighter than mine) minds to consider it as well. The more standardised the test we can develop, the better it will be to diagnose the issue for us, and easier to discuss the nature of it.


    John P.
     
  13. bpetterson

    bpetterson Guest

    If any of you got the Nikon D1 in 1999, you will recall that we had some problems at that time all sell namely banding and magenta. In some cases the oscillator was exchanged and uses learned how to cope with the magenta. Banding in many cases was solved by adjusting the settings that you selected.

    What we should consider doing now is seeing if we can come up with a work around for this so-called banding problem. I call it so-called because I do not have any banding problem with my camera. The D200.

    Another interesting fact. All previous D cameras that were shipped to the U.S. at first had serial numbers below 1000. That was not the case of the D 200.

    Now let's see if we can come up with a work around for this situation until Nikon can fully assess.

    Birger
    N26 42.435
    W80 02.477
     
  14. PJohnP

    PJohnP

    Feb 5, 2005
    Birger :

    I think the banding issue is highly variable. I'm getting almost no banding, unless I work to create it. Others are getting significant banding in normal shooting. It's a frustrating issue when it's occurring on some cameras and not others. However, I respect the photographers who are getting this enough to give full credence to their findings - they're not newbies who can't manage their settings adequately.

    The idea of the workaround is a pretty good one, and I think we're moving closer to this, but we do need additional data from those people who are getting more banding effects to be able to develop a plan. I'm going to quote from another thread on this subject where Iliah and I were discussing this.

    Lots to think about.



    John P.
     
  15. bpetterson

    bpetterson Guest

    On the D1 we switched tone and saturation to normal.
    This is what I set up my D200 to.
    For the magenta we set the magenta to -19 in PS.
    Nikon did not have or was not allowed to send out firmware opdates.
    In some cases I think that if you sent in your camera, it would be updated [secret] with the same # . Notice I said "I think".

    On the D2X again I say 'I think" Fenton and RR were the main ones
    to get the focus fixed.

    I don't think that Nikon will admit exactly why when and if a firmware update is provided.

    Birger
     
  16. Thank you John, first, for managing this thread the way you do.
    As I said in a previous post, my D200 shows banding under the now known "Peter's recipe" (ISO 400 + 2.0 EV, light metered on the highlights, curve adjustment to open the shadows) though later posts by Peter have demonstrated banding can and does occur in many other circumstances.
    I have already shared some observations :
    1- Banding occurs in the up-down axis of the sensor set horizontally, thus a vertical shot will generate horizontal banding.
    2- Banding occurs in the upper and under areas (or left and right) of highlights.
    From this, I speculated that it may be linked to a memory effect of the sensor. Bjorn has refuted this hypothesis.
    Here are some new observations I have made since.
    1- Looked at at high magnification, the bands are not clearly contrasted in terms of a bright one alterning with a dark one, as the "memory effect" hypothsesis suggests. Instead, they look more like "patterned" noise :
    [​IMG]

    2- A RAW converter called "Developper Studio" from SilkPix (http://www.isl.co.jp/SILKYPIX/english/) has a functionnality to control what its authors have named "geometric noise". This functionnality does take care of the banding in the same image (though it introduces jagging):
    [​IMG]

    3- The same scenery, shot at ISO1600, lend a coarser noise, but seems to be free of banding as such :
    [​IMG]

    Because of the dispertion of the banding around areas of highlights, I tend to continue thinking that it has to do with the sensor saturating, but also that the phenomenon may be made problematic (revealed actually) because of the way the firmware manages noise. The shots shown here where taken with a D200 with its noise parameters set to "on-normal' reduction from 400 to under 1600, and then "high" from then on. Perhaps the noise reduction applied with the "normal' setting is in part responsible for letting the banding show through, while the stronger reduction used in the ISO 1600 is better effective at that.

    Peter, can you make more experiments using higher ISO, and various noise reduction per ISO parameters ?
    I will try some time soon shooting at 400 but with the highest noise reduction avaliable.

    Hope this helps in some way, and that my post is not just more "noise" interfering here !
     
  17. I'm seeing banding in some of my shots, usually with visible fluorescent light sources in frame - shows up in the dark high contrast areas adjacent to the light source.

    The banding gets worse if I push an exposure. If I don't make EV adjustments during the conversion then it's barely visible at 100%, and it's gone during resize. I'm prepping shots for print so we'll see how bad it gets on the final images.

    During conditions where banding is visible, I have equally obnoxious noise to deal with. I take it as a sign that I'm just pushing the shadows too hard during processing. Bringing the shadows back down usually takes care of it.
     
  18. fks

    fks

    Apr 30, 2005
    sf bay area
    hi sune-

    robert posts here as RFCNIKON. his post of a night shot that suffers from banding was shocking.

    i haven't seen any banding on my photos yet (fingers crossed), but i have barely 400 shots, mostly in overcast weather.

    ricky

     
  19. nfoto

    nfoto Guest

    As far as I can glean from the D200 shots salvaged after my server crash, banding around strong highlights is virtually absent, only getting low-light banding noise.

    Several points worth noting, some of which are gradually becoming addressed:

    1. How is the camera set up? I did set "my" D200 as the D2X, viz. exposure fine-tune -1/6 EV, image contrast: low, no in-camera sharpening, long-term NR: on, high-ISO NR; either ON or OFF depending on the tests.

    2. Which RAW converter, if any, was used? I've seen similar differences betwwen NC 4.4 and Bibble 4.5 as in the depicted image with RawMagick.

    3. "Banding" noise is a generic term and what I see are three different kinds of it. Firstly, the alternating dark/light, often very short, striping of a comb-like nature surrounding intense highlights set against a dark background. Secondly, striping over all over the shadow areas going from from one edge to the other (thus different from the comb-like pattern #1), some times also including middle-toned areas. Thirdly, striping comprising the *entire* image, dark and light areas, seen for night images only.

    4. ISO setting. I've seen striping #2 at 1250 ISO, others at 100.

    5. Light sources. No striping seen by me under fluorescent lighting.
     
  20. PJohnP

    PJohnP

    Feb 5, 2005
    Bjørn :

    Your third point is an excellent one. I think that we need to differentiate for this thread between small localised "comb" banding in areas of high contrast (the phenomenon I personally experience with my D200), and the latter two band effects (shown by Peter, Robert, Christian). Based on the discussions to date, that's a reasonable division. We'll call the high contrast area effect, "Banding Type I", and the succeeding ones "Banding Types II and III".

    Banding Type I seems to have a clear recipe for repetition (Peter's Recipe), but it's addressable with care in exposure, and can be managed. That doesn't per se excuse the condition, but it allows for better diagnosis.

    Now, it may be that Types II and III are indeed separate, or they could be related, but these (for the moment) are the more pressing issues for most people. Type III may (or may not) be related to low light noise conditions, but it would seem on the face of it to be similar to Type II, just not as prevalent.

    And this division of conditions does, at least, give us the opportunity to see that some cameras may have a definite fault, differentiating them from some other D200s that do not exhibit it. OTOH, we just may not have the right recipe yet for Types II and II behaviour - I have hopes we can see where this difference in behaviour derives from.

    It's a start, at least.



    John P.

    P.S. I'm on assignment in Alabama this week, so I'll only be able to assist on this thread in later evenings, and for short periods before dawn each day. Anyone who can summarise or move the thread forward shouldn't wait on me to assist (although I appreciate the vote of confidence from Christian - thanks !).
     
Loading...
Similar Threads Forum Date
is this type of banding ok on d200? Nikon DX DSLR Jan 22, 2009
D200 anomaly -Help - Banding or what? Nikon DX DSLR May 10, 2008
D200 banding issues Nikon DX DSLR Mar 5, 2008
D200 Banding and Nikon Service Frustration Nikon DX DSLR Feb 13, 2008
Banding Fix update....D200 Nikon DX DSLR Feb 28, 2007