Hot pixels at 30D

[localchap]

Hello all, I've never found similar thread so decided to start off a new one here.
There's been a lot of talks about dead/hot pixels at camera sensor. I've tested mine brand new 30D and...(for God sake I regreted I've done it 'cos it made me really upset)
SO, I put kit lens 18-55 on, ISO 100, RAW file, manual mode, shutter speed, 1sec-nothing, 1/30, 1/60, 1/150 and up to 1/600 with the same ISO level, nothing. But when I went for longer 10 and 13 seconds speed, I've found (while watching at PC at 1:1scale) a white little tinny dot and another blue one which appeared at one shout and gone at second shot. Its not seen at any of pictures taken since I've bought it
I'd like to hear some experienced opinion, how crucial it is, is there need to apply to Canon service for sensor remmaping (its s bit complicated for me I've bought it in States now back home to Europe). Besides, like I heard remapping could make sensor working worse so you never know.
Also, I heard that every sensor had at least one dead/hot pixels, is that true.
All opinions and sharing thoughts at the point are really welcome!

[Ronald S. Jr.]

As I understood it (I don't do long exposures much), hot pixels in very long exposures are fairly common.

[localchap]

I don't do long exposures myself just decided to have it tested an at 10 then 13 sec those dots appeared.

[Jon]

One or two hot/dead pixels out of 8.2 million at exposures of over a couple of seconds aren't worth worrying about. It's when you have several hundred at 1/500 that you need to think about worrying.

[JBF]

Canon's tech support told me that at long exposures and high ISO speeds acceptable hot pixels would be less than .01 percent. If that is the case then on an 8 megapixel camera if you had 8,000 hot pixels it would be in the acceptable range. However I tested mine and at 30 sec and ISO 1600, I had 144 hot pixels, so I am not too upset. And even at that type of shot I can't see the hot pixels unless I zoom in at 100% or higher. Don't sweat the small stuff.

[SteveO]

I heard a rumor that with the 30D, if you do a "sensor clean ", the camera automatically maps out hot pixels. Any truth to this?

Steve

[Jon]

Nope. That rumour first surfaced with the 5D. It seems to have a strong link to people forgetting they've turned automatic noise reduction on.

[Keiffer]

I don't do long exposures myself just decided to have it tested an at 10 then 13 sec those dots appeared. So if you don't normally do 10-13 sec exposures you'll never notice this again:-) Isn't that a happy thought.
That's the bad thing about looking for trouble or problems, you normally find them.

[calicokat]

I wouldn't worry about it :)

[Tdragone]

This 30 minute exposure has more than 30 pixels of various colors in it:

http://www.pbase.com/tdragone/image/60647209.jpg but none of them show up after Post Processing (Noise reduction) or when printing the original below 8.5X11 (Where 2 are viewable with NO post processing)

Hope it helps calm the nerves :)

[Richard7481]

They can be removed in PP, so why worry.

[calicokat]

This 30 minute exposure has more than 30 pixels of various colors in it:

http://www.pbase.com/tdragone/image/60647209.jpg but none of them show up after Post Processing (Noise reduction) or when printing the original below 8.5X11 (Where 2 are viewable with NO post processing)

Hope it helps calm the nerves :)

thats a great photo

[roli_bark]

.... However I tested mine and at 30 sec and ISO 1600, I had 144 hot pixels, so I am not too upset. ....

Are you serious ? How could you count exactly 144 hot pixels ? Or, for that matter, how can Hot Pixels be counted at all ?

Am I missing something ?

[adas]

Roli, he's using a Dead Pixel Test program, wich will count it for him.
Another way is to look at the numbers given by the histogram in Photoshop.

[prep]

High dark current pixels can be partly corrected by subtracting the data number that accumulateds between the clear out read and the read down of the array. This will be longer than the shutter open time, and possibly the exposure time if it is a gated array.

I woud expect Canon to have done it right.

BTW, you may want to stack several shorter exposures, you get better S/N that way.

[localchap]

Lads, I've found just brilliant program that counts dead/hot pixels when you upload your picture with lens cap on and done with various setting, I'm not sure if I'm allowed here to place a link for it so if moderator wouldn't mind I will, quite interesting program BTW.

[Jon]

If you're talking about DeadPixelTest.exe, the link's posted on here somewhere. Post the link to whatever you found by all means.

[brconflict]

All, I bought a 30D which exhibited a single white dead pixel in nearly every image, even at ISO 100. Canon replaced the sensor assembly for me. The main reason for Canon going with a CMOS sensor, to my research is to make sensor manufacturing less expensive. At first, CMOS sensors were horrible in quality, and you may notice that Nikon is only now starting to make cameras with them as well. Where I think Canon gets away with this technology is through the DIGICII processor that rids the CMOS sensor of it's short-comings. In any case, Canon saves money with the CMOS (barring R&D costs), so they can afford to give us fewer dead pixels. Bear in mind, that in my experience with the 30D and the 350D, yes, everyone is right that low-light shots at high ISO settings, and long exposures will certainly produce hot pixels. With as fast as dSLRs have improved in the market, you honestly can't complain too much.

My opinion is that Canon is all about mass-market, and lower cost. The more they make, the cheaper the manufacturing process, which allows us better, cheaper products, but where they lack is good support (Canon cannot deny their poor support), and higher rate of defects.

[JaertX]

brconflict....I hate to tell you this, but, um, all companies are all about lower cost...profit is what keeps you in business. maybe you've had bad luck, but many of us have had exceptional customer service from Canon. In fact, product support is a major reason I'll stay with them.

anyway, so you people are using some type of software to count your hot/dead pixels...that's horrible. Cheating, in fact! You have to count them all by hand or you're not a real photographer. Oh wait, that was a different thread.

[wasbridge]

not to hijack the thread or anything like that but... as you can see I have a 10D, and I like night shots where I take anywhere from 30second to 15 minute exposures and I get a bunch (50-100) pixels or blocks of strange colors. Generally at least 10 are viewable at no, or very little zooming in and the rest are pretty inconsequential, is this a broken camera, or just a symptom of an old camera? also, would something newer fix this?

[Jon]

All, I bought a 30D which exhibited a single white dead pixel in nearly every image, even at ISO 100. Canon replaced the sensor assembly for me. The main reason for Canon going with a CMOS sensor, to my research is to make sensor manufacturing less expensive. At first, CMOS sensors were horrible in quality, and you may notice that Nikon is only now starting to make cameras with them as well. Where I think Canon gets away with this technology is through the DIGICII processor that rids the CMOS sensor of it's short-comings. In any case, Canon saves money with the CMOS (barring R&D costs), so they can afford to give us fewer dead pixels. Bear in mind, that in my experience with the 30D and the 350D, yes, everyone is right that low-light shots at high ISO settings, and long exposures will certainly produce hot pixels. With as fast as dSLRs have improved in the market, you honestly can't complain too much.

My opinion is that Canon is all about mass-market, and lower cost. The more they make, the cheaper the manufacturing process, which allows us better, cheaper products, but where they lack is good support (Canon cannot deny their poor support), and higher rate of defects.Less lower cost than lower power consumption and room for technical growth. Your "barring R&D costs" says it all. Canon developed their own sensors so they actually have R&D costs for their sensors. Nikon buys sensors someone else has developed. Now that's cutting costs. But it means you're stuck with what your supplier is willing to make.