[Marc Sabatella]

They do, it's just masked in camera or in RAW development. The *ist D can save as TIFF (IE less error due to jpg compression) and doesn't mask hot pixels.

I see; so you *are* talking about dead/hot pixels and not noise. Once more; yes, dead/hot pixels *do* gradually develop over time. But that's *not* the same thing as noise. Noise performance does *not* gradually degrade, and that's what the OP claimed and what you originally claimed you saw. Two totally different things. Defective pixels *do* happen over time. Reduction in noise performance does not.

[distudio]

I see; so you *are* talking about dead/hot pixels and not noise. Once more; yes, dead/hot pixels *do* gradually develop over time. But that's *not* the same thing as noise. Noise performance does *not* gradually degrade, and that's what the OP claimed and what you originally claimed you saw. Two totally different things. Defective pixels *do* happen over time. Reduction in noise performance does not.

Strictly speaking hot pixels are noise, they are not wanted signal but if the initial query was specifically referring to random noise then yes you are likely correct. However I would not be surprised if there was limited but measurable degradation in random noise performance also (as hot pixels become more susceptible to self saturation I would expect that the same mechanism affects all pixel just that the degree of effect varies).

Another interesting article:

http://www.harvestimaging.com/pubdocs/130_2009_IISW_cosmicrays.pdf

[FullertonImages]

I'm got 3 years of hard use and over 20,000 clicks on my K10d. It's got a few dead pixels, but nothing major, and the noise seems to be about the same as before. It does however get nasty hot purple spot bleeding in from the frame edges when I do long exposures, like 10 minutes or longer. I don't know if that's age related, or if it always did it, since I've only tried them recently. It's a bummer too, since that's my favorite camera for long exposure, since it's the only one that you can turn off DFS over 30 seconds. I hate waiting for another 30 minutes with my K20d or K-7.

[clark]

honestly you're probably just (subliminally) looking at high ISO shots from say the kx and thinking to yourself 'my camera is so noisy'

I often think my sensor is screwed because it's so noisy, but I look back at some of the first pictures I took on my ist DL and they're just as noisy

it's like how before HD, standard definition looked fine, but now I have HD I could never watch SD again without noticing the quality drop

[Mike.P®]

I thought my K10D was taking worse shots than when I got it (3 years old and around 50,000 shots) but after using the firmware hack to change the focus it is much better now.

I managed to get a K-7 and kit lens for a very silly price a few days ago (cheaper than I could sell my K10D for) and its very nice and all that but imo the IQ is not any better (or very little anyway) than the K10D .. infact I am thinking of selling the K-7 and waiting for the next Pentax body.

[axl]

My first K10D gave up recently, after issues with continuosly deteriorating AF (camera kept back focusing more and more) last month the sensor decided to go. I posted an example here.
It's shame, because I really really like K10D (never quite liked K20 and K-7 is small for my hands)
Camera was bought new in XI. 2008 and made ~30.000 clicks in those 18 odd months. I hope my current K10D GP will fare better... fingers crossed!

[ducdao]

My first K10D gave up recently, after issues with continuosly deteriorating AF (camera kept back focusing more and more) last month the sensor decided to go. I posted an example here.
It's shame, because I really really like K10D (never quite liked K20 and K-7 is small for my hands)
Camera was bought new in XI. 2008 and made ~30.000 clicks in those 18 odd months. I hope my current K10D GP will fare better... fingers crossed!

Hmmm 30,000 clicks for a K10D is such a short lifespan compared to the 100,000 clicks claimed by Pentax. Have you tried to contact Pentax and see what they have to say about that?

[Marc Sabatella]

100,000 is the stated life expectancy for the shutter, not the sensor. Normally, sensors would outlive the shutter, but one might occasionally die if you're unlucky.

[distudio]

I see; so you *are* talking about dead/hot pixels and not noise. Once more; yes, dead/hot pixels *do* gradually develop over time. But that's *not* the same thing as noise. Noise performance does *not* gradually degrade, and that's what the OP claimed and what you originally claimed you saw. Two totally different things. Defective pixels *do* happen over time. Reduction in noise performance does not.

I just dug up a noise profile pic from 2004 and shot another on the same camera which apart from firmware revision and slight deviations in battery condition and temperature were, for this intent, identical. It's plain to see the degradation of the sensor over this period in the form of additional noise.

Free File Hosting Made Simple - MediaFire

[ducdao]

100,000 is the stated life expectancy for the shutter, not the sensor. Normally, sensors would outlive the shutter, but one might occasionally die if you're unlucky.

Ahhh good catch, Marc!

[ducdao]

I see; so you *are* talking about dead/hot pixels and not noise. Once more; yes, dead/hot pixels *do* gradually develop over time. But that's *not* the same thing as noise. Noise performance does *not* gradually degrade, and that's what the OP claimed and what you originally claimed you saw. Two totally different things. Defective pixels *do* happen over time. Reduction in noise performance does not.

When you guys mentioned dead/hot pixels, what do you really mean by that? Are those dead/hot pixels visible on the body lcd or image?

[distudio]

When you guys mentioned dead/hot pixels, what do you really mean by that? Are those dead/hot pixels visible on the body lcd or image?

In the image, hot/stuck pixels can contribute to image noise and/or show as permanent specs of colour in an image file.

[Marc Sabatella]

It's plain to see the degradation of the sensor over this period in the form of additional noise.

No, what's plain to see is additional hot pixels. Furthermore, most of these are probably not true hot pixels in the sense of being permanently stuck on, but rather, ones that become evident only in the extremely artificial conditions of the test (these appear to be shots take with the lens cap on in a 30 second exposure). And indeed, temperature and battery condition may well be playing a role there.

As I expected, noise in the classic sense appears no worse - if anything, perhaps slightly better - in the more recent shot. To the extent one wants to judge noise in such an artificial test anyhow.

[Marc Sabatella]

When you guys mentioned dead/hot pixels, what do you really mean by that? Are those dead/hot pixels visible on the body lcd or image?

In the image. A true hot pixel is one that is *always* on. You'll see a one pixel red, green, or blue spot in exactly the same position on every photo. The above example is slightly different - these are pixels that are not permanently "hot", but which can become so temporarily as the sensor overheats.

[distudio]

In the image. A true hot pixel is one that is *always* on. You'll see a one pixel red, green, or blue spot in exactly the same position on every photo. The above example is slightly different - these are pixels that are not permanently "hot", but which can become so temporarily as the sensor overheats.

Marc a pixels that's always on is a stuck pixel, a hot pixel generally has a state which varies with exposure time and a dead pixel gathers no charge. The noise in the images that I presented has far less to do with overheating than aging of the sensor and the latent hot pixels combined with the normal noise of the sensor will decrease the sensors overall signal to noise ratio, I don't know why this concept is so difficult to accept. 30 seconds at ISO 200 isn't outside the scope of my general photographic needs either.