Might want to check your camera's folks.
The US site appears down right now, or at leas the link to check your serial number generates a 404 error...
Here are details of the issue at hand. Looks like those shipped now are already fixed.
Glad they're sorting out problems fast now. They have learnt their lesson from the D600 debacle.
I am not a painter, nor an artist. Therefore I can see straight, and that may be my undoing. - Alfred Stieglitz
This whole issue is very interesting to me. My D800e has done this same thing from day one. I attempted several times to reach Nikon about it, but never go anywhere. From reading the posted link, the white spots are definitely the same issue that the D800e has, identical. I have the latest version of firmware on my D800e.
The article implies that a firmware update from Nikon will fix this issue, however my camera serial number, when entered on the Nikon site, came back with a "send camera to Nikon".
I have used the D810 2x out at night in exposures of 1' 40" and 2' 35" and it showed this issue. I called Nikon today, and they want the camera back. So mine goes back tomorrow. I hate to send it in just for a firmware update, but so it goes.
One note, Capture One Pro will get 96 go 98% of these spots out by using the single pixel noise slider. This slider does a great job and does not really detract from over image details. LR on the other hand requires a pretty harsh application of noise reduction to the entire file. Which mean two runs, one for sky, one for non-sky. This issue really only seems to show up in sky, or solids.
I wrote about this same issue on the D800e over a year ago:
Hope they can turn this around quickly and I wish they would make a fix for the D800e. Surprised to see in the featured post that they showed a D800e not having this problem. Mine was an early unit, I need to remember not to be on the bleeding edge so darn much.
Sigh. They better pay the damn postage this time. The D800 cost me a FORTUNE in shipping for service. Sony are much better: it's quick and free.
Oh, did I mention that mine is affected. I'd noticed it a bit, though I rarely make this sort of exposure. Poo.
Glad I didn't rush out and buy one yet.
These QC issues are proving to be very annoying.
I spoke to Nikon UK today, postage is free and turnaround is 3 days. Also they say you don't 'have' to send it in right away, inferring that once the rush is over, turnaround time quite possibly will be shorter. Mine is affected too but i hardly ever expose that long.
Steve Brickles2 Member(s) liked this post
Looks like it's both a firmware and hard code fix as they are changing part of the algorighims on the processing. Sending mine in today.
The postage is prepaid, ground in the U.S.
Still puzzles me that this was noticed so fast by Nikon on the D810, but no one but myself seemed to notice it on the D800e. I do wonder if the D800e has the same fix, but it was applied to later bodies and thus never released as a service fix. That's too bad. I would like to have someone with a more current D800e test this.
I got all the paperwork ready, then on a whim did a couple 30 second exposures of the back of the lenscap. I can't see anything other than black, no problems whatsoever. How long of a long exposure is needed to set the white dots?
TIP: When the D750 comes out, don't be in the first wave of buyers. D600, D800, D810. Not. Good. Enough. Nikon's financial woes are worrying enough without them testing the loyalty of their user base like this.
I would try for about 1 minute. Also try it at night with a solid night sky. If the camera has the issue they will start to show up as faint grain. Once showing they stay illuminated for all following exposures.
The fix appears to only take a few minutes as they reprogram part of the chip then add a new firmware that takes advantage of the re-programming.
Most won't see this unless they are into stacking time lapse or doing really long exposures in daytime with a ND 10 stop.
I noticed the problem with my D810 as soon as I started to shoot during a 3 hour session last week. My D800e takes longer for the dots to show up but they eventually get as bad.
With best regards, K-H.
There are innumerable light sources in the sky below the human vision threshold.
Shooting the inside of the lens cap, with the lens and cap wrapped in heavy black cloth, is valid.
It's this post that confirmed to me that sometime during the manufacturing run of the D800e, Nikon figured out that this issue existed on the 800 series cameras. I noticed on mine from day one and tried to get Nikon to fix it.
As your 800 is clean it's safe to assume this fix was implemented to later models but was not offered to earlier cameras which is unfortunate.
My D800e looks just like your D810 shot.
Glad to see Nikon is fixing this on all 810s.
We can agree to totally disagree. See my earlier post in this thread where I wrote about this problem on the 800 and 800e. I have examples of the effect.
Shooting a lens cap may help for colored stuck pixels but this problem is best seen with the camera in real use. I have literally thousands of exposures taken at night with either the 800 and e and the problem is very easy to see.
It's a very distinct issue and I have never seen it from any other Nikon camera or Canon.
The night sky cannot be characterized in any meaningful manner by any person using any technique.
Therefore it cannot be used as a valid test target for any valid tests.
Of course you can take pictures of it, but they don't mean anything.
If the problem only shows in images of the sky, not in images of a completely black featureless entity, then the problem is not in the camera.
The internet is littered with comments from people who claim that specific issues can only be seen in images of "valid" subjects, not in laboratory situations like using gray scales or resolution targets.
I challenge anyone to explain why a laboratory test target is somehow not a "valid" subject.
I never stated a Lab test target is not valid. Sad you feel the need to use the term littered, as I feel my claims are both well documented by myself and my website and others who engage in this type of time lapse photography. I am sorry that you feel a need to discredit my findings, which are from hundreds of hours work working with the 800 class Nikon bodies.
I have noticed this problem with all long exposures with all Nikon 800 class cameras since I started using them in early 2012. I am shooting mainly all long exposures, 1 minute to 4 minutes with moonlight illumination for the night sky, so I am working with a black sky, or a sky colored by light pollution. The effect shows up very easily on all the exposures I have taken in this manner, and the effect is different than any other Canon or Nikon I have used, in that the images will be covered in thousands of stuck faint white pixels.
My point to Jack was that, I constantly see this issue in the night photography I have done and mainly in exposures exceeding 1 minute. I also feel a photographer needs to be shooting more than one test frame as this is caused by thermal heat build up (Nikon themselves have admitted to this) thus the effect may not be seen on a single test.
Believe what you wish, test as you will. You also might try a D800e or D800 or now a D810 out one night and shooting a series of 1 minute to 2 minute exposures over a period of 20 minutes and see if don't get the same effect.
There is no such thing as a black sky, particularly when resolving the level of illumination achievable with a 4-minute exposure.
Look at images of the sky posted by shooters in really dark areas, like Central America.
Speaking of which, the Mayan culture that lasted for about 2000 years in that area was quite adept at astronomy.
They defined constellations as the dark areas that are found infrequently against the mass of stars.
This is what you are calling a "black" sky.
Here's an example of what I'm talking about (stolen from another thread here on GetDPI):
Last edited by Leigh; 20th August 2014 at 21:42.
Mine is one affected based on the serial number. When i try to print a shipping label i get this
The value you provided for the Ship-From State exceeds the maximum allowable length of 5 characters!
Please trim Ship-From State and try again.(0)
Okay. I took a pair of 2 minute exposures into the lenscap. I take the second one and open it in Capture NX-D, it appears totally black corner to corner. If I then crank the exposure slider up the full +5 stops, I just barely detect a slightly lighter gray -- call it 8/8/8 to 10/10/10/10 gray on the 0-255 scale. Now if I increase gamma massively to +2.0(!) I get say approximately an 80/80/80 gray. I see this as a problem in obscure minutia for my uses as I *never* process anything like that. I thus am reluctant to even bother shipping the body in because all else works so well. Thoughts?
Here is an example of the white dots. Iso 320 as I remember, 1' 40". This is from a shot taken at midnight, blue is from moon illumination. Dots clearly show.
Nikon US has agree that "these" are the dots they are referring to.
They will never be in the same locations, from one shoot to another, however during a single night, they stay in pretty much the same place, but over time become even more numerous. They are not distant stars or other natural light, as if stars they would be trailed just like the "real" stars.
This is from a brand new D810, however my D800e does the exact same thing.
They're all the same length, moving in the same direction, and of varying intensity.
Any amateur astronomer would recognize them.
I showed that shot to my wife, a very accomplished and experienced astronomer, and she identified them instantly.
You can find numerous night shots right here on GetDPI that show exactly the same features.
If they're real sensor artifacts they will appear in the test shots using the lens cap.
If you find no such artifacts in the lenscap shots, they're star trails.
Sensor pixel problems are at single locations, not dragged across the sensor as seen in your example.
Leigh, I fully understand what star trails are, read the article I wrote on the front page, I have been doing this for a while.
In the previous post I mentioned the solid lines are star trails. I even noted that in my post. The long lines are stars that created trails at 1 minute and 44". The single white dots are NOT. Please look again, zoom in, zoom in again. I can't believe you can't see single white dots. There are hundreds. Make sure you click on the attachment, if it's not loading at a 100% view, my fault. But when I look at the post, I can click on the image and it's large enough to clearly see the white dots.
I am stunned you feel I don't understand what 'star trials" are.
At a loss on this one. Not sure what you are seeing, but Nikon saw it, they saw the dots, they asked for the raw, they agreed, Yes there are white dots and these are what they are trying to address.
The dots don't trail. They stay static over two hours.
Look at the attachment again, I just did. They dots which are about the size of 1 pixel, are clearly there.
One of the main problems of the web is that the experience level of its members spans an almost infinite range.
I've found it best to assume nothing, and let others provide the missing information.
This results in far fewer mis-understandings than assuming expert level proficiency on the part of every poster, and being wrong 99% of the time.
I opened the image in a graphics program and the dots are quite apparent when I enlarge it.
Apologies for not understanding the focus of your concern.
However, my criticism of your assertion that a night sky is "black" is absolutely valid, and remains so.
No problem at all. I was getting frustrated and sorry for tone. Sometimes the web text etc can't get the meaning across.
This issue is only a problem for me as I stack, and stacking creates gaps, either my me or just the intervals and the final images look better gap less. When I use "Startracer" software to move the images, via Ptgu, the dots then become thousands of dotted lines, that will ruin the image.
The LR noise reduction at high settings gets it out but it destroys the rest of the image as too much blur is done. Capture One single pixel does a very good job, but many times I seem to prefer the LR output for night work.
As the camera works into the night the dots will start to number into the thousands and it's really impossible to manually take them out. I have tried to get Nikon to address this issue for over year on the D800e, but they never did. However based on a link in the post, later model d800e's don't have the issue, so it was fixed under the covers in later cameras. Sad that the 800 family can't have the same fix done.
Thanks. I can understand how the issue could be quite prominent under some circumstances.
All of my wire's astro work has been and still is with film rather than digital, so I have no basis for comparison.
This is unfortunately a problem with sensors of all types.
When the desired signal is so small as to be the same order of magnitude as the noise, there's a problem which is very difficult to cure.
Hope Nikon can resolve it to your satisfaction. Good shooting.
Performance limits -- exactly.
Software -- I only have Capture NX-D for the D810 files at the moment.
Here are some pics full frame downsized to 1200 px with some added comments following:
Base exposure processed out zeroed:
Plus 5 stops exposure:
Plus 5 stops and plus 2 gamma:
First comment: If I zoom in to 200% and examine process 1 closely, I see nothing untoward, not even noise at single pixel level. If I zoom to 200% on process 2, I do see evidence of single pixel noise, but they are not white dots, they process out of NX-D as dull to brighter green, but it is combined with presence of excessive noise in general, like ISO 3200 stuff. If I zoom in on process 3, then the single pixel noise does indeed appear as a mixture of gray and white dots, but then noise is so high, like for ISO 25K, I would expect nothing less.
Next comment: CaptureOne has a very excellent single-pixel noise removal option. In fact, this is such a common problem in MF capture, I have it set on as default for ALL of my processing specifically because it attacks and kills hot pixels -- I am thinking it is likely the reason why I never noticed the issue with my D800 or D800E. I am hopeful that it will work with the D810 as well, but obviously 1) I don't have a very severe problem to begin with and 2) we won't know for certain until the a C1 release comes out that will accomodate the D810 files.
So in conclusion, my "end of day" decision for me is I am NOT sending my unit in at this point in time.
"Perfection is not attainable. But if we chase perfection, we can catch excellence."1 Member(s) liked this post
Thanks very much for the follow-up, Jack.
Yes, I expect this is hitting the noise limit of the sensors.
In reality, these are "energy" sensors as opposed to "light" sensors.
They respond to temperature and other environmental conditions as well as photons in most cases.
Sent my D810 in yesterday using the prepaid label downloaded from the Nikon UK website. Got the acknowledgement of receipt this afternoon with an 'approximate repair completion date' of 4 September.
Looks like I'll be relying on my 800E for the next couple of weeks.
John L Dobson
Editor, Ffestiniog Railway Magazine
I sent mine in.. I figure might as well get it taken care of. They said it should take 7 to 10 days max before i have it back
Hopefully no longer, as the actual fix take about 10 to 15 minutes to apply. I noticed that with mine, they are giving UPS the full 7 days to deliver it even though it's in LA now. It will just sit in the UPS warehouse for 2 more days. Glad kept the 800e, as this "repair" will be taking about 2.5 weeks total, all in slow ground transport. I should have just paid to have it go 2nd day air.
Have just been advised that my D810 is due to arrive back, by courier, sometime tomorrow. That will be exactly a week since I posted it.
John L Dobson
Editor, Ffestiniog Railway Magazine
The postage is prepaid, ground in the U.S.
John L Dobson
Editor, Ffestiniog Railway Magazine
I received my D810 back today, and just finished a 20 minute test, with 2 minute 10" segments, iso 160.
Results were very impressive. There are still occasional white dots, but now they number in 10 to 20 instead of 2000 to 3000.
I can see no reduction in image quality, basing this on my first series taken before the fix.
The LCD view on the D810 is a bit of a fooler, in that often I see breakups in blues when viewing at 100%, however when you open the raw file there are no problems.
If you work with long exposures with this camera I feel the fix is worth the hassle.