D700 - how many dead pixels are 'acceptable'

<p>I am shouting myself a camera upgrade. After many great years of image making I have sold my D700 and look forward to shooting with the D810.</p>

<p>However, my D700 buyer claims to have discovered 10 dead pixels on the sensor and is very disappointed with this. Frankly, I was unaware of a single dead pixel on the sensor in my 6.5 years of pixel peeking with the D700 and believe me, I looked long and hard at many exposures as I was proof peeping files ready for printing in large sizes, however, I never really went hunting them down either.</p>

<p>Wrong or right, existance of dead pixels or not, this unfortunate revelation led me to wonder, how many dead pixels on say a 12.1 MP sensor like the one on the D700 would begin to effect image quality and begin to irritate the dicerning photographer?</p>

<p>I don't remember seeing a dead pixel in my DSLRs but then I haven't gone out of my way specifically to look or them, either. Some of the high resolution models such as my D810 has quite a lot of hot pixels at high ISO which surprised me a bit but those pixels are so tiny that I am not concerned about them. If the hot pixel happens to fall on a face I just use the healing brush to remove it. On the D800 there were fewer of them.</p>

<p>You could consider having Nikon service map the dead pixels so that the camera is aware of them and interpolates the image for those pixels. </p>

<p>I'm guessing Dead Pixels show as black, ie 0,0,0 ? ...Where-as Hot Pixels are 255,255,255?</p>

<p>Could the buyer be seeing <strong>black</strong> dust spots?</p>

<p>Depending on the sale transaction mechanism etc and whether they are expecting a discount because of this 'flaw', I'd consider various options including Ilkka's above. </p>

<p>There is still a high demand for clean and low-count D700's, although the price is v.slowly dropping. Finding a new buyer shouldn't be hard.</p>

<p>My D800 came with 2 or 3 hot pixels. Not a big deal for normal use, but a bit annoying if you're trying to take starfield pictures (which I don't). Dead pixels, I'm not too sure about, since I haven't actually spotted any, but I believe they get mapped out automatically anyway. Hot pixels are usually much more noticeable than dead ones, unless you take pictures of blank white surfaces all the time.</p>

<p>The D700 I have still seems very clean, but might well have a few anomalous pixels. To be honest I don't look too hard for them. Dust spots are a much greater menace, but of course more easily got rid of.</p>

<p>IMHO, 10 dead pixels is no big deal. Especially on a used camera that must be a few years old, <em>and</em> they can be mapped out. If the 10 pixels were all clustered together, then I might start to worry. I also doubt that there was a single frame of film ever shot outside a clean-room that didn't have dust particles on it amounting to far more than the area of 10 dead 8.5 micron pixels. That's less than 0.0001% of the entire full-frame area! Or under 1 ppm.</p>

<p>One more thing: Are you sure the buyer is actually seeing dead pixels and not dust spots? I'd ask them to send you a sample image.<br /> Edit: Just seen that Mike suggested the same thing.<br /> "I'm guessing Dead Pixels show as black, ie 0,0,0 ? ...Where-as Hot Pixels are 255,255,255?" - Not necessarily Mike. Since a cluster of 4 photosites are amalgamated into one pixel, a dead pixel will look anywhere from black to a dark pastel colour. Likewise a hot pixel can look strongly coloured to pure white depending on the image content and ISO speed chosen. In a lot of circumstances both hot and dead pixels are almost entirely invisible.</p>

<p>RJ. Good point; I wasn't considering the 4 photosites to 1 pixel bit! Doh!</p>

<p>Yup, get them to send you a frame from it. To be honest, i'm not sure what one would look like but looking at it @ 100% I suppose it should only be 1 'odd' pixel, no fringe, nothing, just an isolated pixel surrounded by 8 'normal' pixels. No AA filter effects here!</p>

<p>Equally, I've never seen single pixel dust either...:-)</p>

<p>I had one green (hot) pixel in my D700. But, I think there was a Nikon specification out there stating a certain number was acceptable, and service would not be provided.</p>

<p>I couldn't find any dead pixels to show, but here's one of my D800 hot pixels magnified 900%. On the left it's seen against some dark foliage and on the right it's against a more neutral grey shadow area. Note how its colour and intensity changes depending on the surrounding colour. Also the fact that it doesn't show as a single clean square.</p>

<p>It was an absolute devil to find it BTW, requiring tedious scanning of the frame at around 300% magnification. Even though I knew roughly where it was in the frame.</p><div></div>

<p>It's common for semiconductors of this sort (non 100% yield factor) to be graded and sold to target application markets accordingly, so it's entirely possible that Nikon buys a certain sensor grade with some specified tolerance on acceptable numbers of pixel defects. We just don't know what that number is. </p>

<p>I'm pretty sure camera sensors develop "stuck" pixels over time. I honestly don't know whether it's cosmic rays or something else.<br /><br />I've got an old D1 that developed a handful. I have a Nikon D700 that has one (that I'm aware of)<br /><br />It's possible you never saw these stuck pixels if you always shoot raw and process with Adobe Camera Raw. ACR somehow detects stuck pixels and maps them out automatically. But those same stuck pixels will show up on in-camera jpeg files as the typical cross-pattern because Nikon firmware does not map out stuck pixels on the fly.<br /><br />APS in Morton Grove told me that for $135 they'll do a clean and check that includes putting my D700 on a computer that will map out the stuck pixels as long as there aren't too many stuck pixels. But they have no way of knowing what "too many" pixels is.<br /><br /><br /></p>

<p>Stuck pixels do not show up as 0,0,0 or 255,255,255 because a single pixel on the sensor is only blue or red or green. Because the R,G,B value of a single pixel in the jpeg file is affected by the surrounding pixels, the color of the stuck pixel can vary with position in the Bayer Pattern as well as the background of the photograph.</p>