Hot pixels on long exposures

smr

smr

Suspended / Banned
Messages
2,047
Name
Joel
Edit My Images
No
I think there may be a problem with my camera, I bought a Lee 10 stopper and tried it out yesterday.

This image is a 322 second exposure and you can see what happens at correct exposure, the first is 0 exposure and the second is pushed to +1 stop. You may have to click on both to see them fuller and the hot pixels - I took some 30 second exposures as well and they came out ok with none of these hot pixels but I bought the 10 stopper hoping to use minutes long exposures. Is this a problem with my camera sensor and does it need repairing?



#

I also went for a very long exposure at 910 seconds on this shot below (not correctly exposed i know), and zoomed in you can see the hot pixels, on all of these images noise reduction was on so it spent as long as the exposure took to reduce the noise - but obviously this looks like a problem with my camera as there surely shouldn't be this many hot pixels? I've seen just as many of them on other 5 or so minute exposures..





interestingly before both of these shots here is a 211 second exposure with no hot pixels... is it maybe then that I didnt allow the sensor to cool down sufficiently before the subsequent long exposures?

 
Last edited:
You can't repair a sensor for hot pixels - if they always appear in the same place then they can be mapped out (at least on Nikons, no reason why Canon can't). 910 secs is pushing it tbh, particularly in the current heatwave! With noise reduction on you are also asking the camera to work harder too and @ iso 100 there should be no need for any noise reduction. And why are you at f13 on a crop sensor at a focal length of 13mm? f8 would have been a better bet imo/e. If you were shooting jpeg, was there any lens corrections etc turned on? All of these things add heat to the mix, and you are already asking the sensor to work hard without adding to it.

I do a fair bit of long exposure work, but I turn off any unnecessary adjustments in camera (lens corrections, ca corrections etc) and shoot in raw, all decent pp progs can deal with ca and lens profiles. Let the camera rest between shots too, to dissipate heat where it can.
 
Last edited:
  • Like
Reactions: smr
Yes, Canon has an unofficial method....I had to look it up again as last time I did was on my 40D some years back.

I found this mention of what I recall was broadly what I think I followed https://www.dpreview.com/forums/post/27957337

HTH but as noted by Paul I think at the exposure periods you are doing you are pushing the envelope too far!
 
  • Like
Reactions: smr
Thanks guys will look at that link too. I think you're right, I have optimism now that I've seen the 211 second shot without any hot pixels showing. So it maybe down, contrary to what I said, that my exposure on the other shots wasn't right etc. and also yes I guess a 13 minute exposure is pushing things a bit too far!

In any case if there are no discernible hot pixels in my 211 second exposure, which is around 3 minutes, I am guessing, hopefully, that the reason for the hot pixels was due to incorrect exposure or something like not allowing the sensor sufficient time to cool before those other longer exposures, by which time the sensor had already got quite warm.

If the hot pixels do become a problem I'll try the mapping method.
 
smr said:
My photos are fine at other exposure times though, anything under 30 seconds I definitely don't see any hot pixels, so is it the same thing? Or is this method more for hot / dead pixels which show at any shutter length?
Click to expand...
There is a clue in the name "Hot pixels" - the sensor heats up as it works, so with a long exposure it is heating up all the time the shutter is open. If that is in adverse conditions or an inordinately long time then it can start to overheat and the pixels (photosites) get too hot and you get the end result you have shown. The simple solution is to shoot for shorter durations and/or at cooler times of the year.

It is something that can, and does, affect just about all digital cameras - some are less prone than others, but it can still happen.

Don't confuse hot pixels with dead pixels, a dead pixel is just that, one that will no longer work ever! A hot pixel can, and usually does recover once it has cooled.
 
  • Like
Reactions: smr
smr said:
My photos are fine at other exposure times though, anything under 30 seconds I definitely don't see any hot pixels, so is it the same thing? Or is this method more for hot / dead pixels which show at any shutter length?
Click to expand...

Sorry I can't answer the specifics......as has been sometime since I did the mapping process but thinking back I was not doing long exposure so...........!?
 
  • Like
Reactions: smr
Just thinking about this a bit more... a couple of months ago, on a not too hot of an evening (although iirc it was warmish) I took my camera out with my camdiox 10 stopper - i thought (for some stupid reason) that the filter must be the reason why I was getting all these specks as I knew them then...

This is a 500 second exposure. The camera had not been used all day until this shot, so the sensor wasn't hot or warm before hand... so I am again wondering if I have a defective camera, basically if I want to do ultra long exposure photography I can't... I think on this occassion I didn't have long exposure noise reduction enabeld in camera - but even so I wouldn't have thought that I should be getting this many hot pixels ......... ? I've uploaded this one to my flickr so it's easier to see, img uploader seems to be a bit of a faff for viewing pictures properly...

Original:

camera test by Joel Spencer, on Flickr

Cropped in where you can see loads of hot pixels...

camera test 2 by Joel Spencer, on Flickr
 
I'd say that's noise rather than hot pixels, and well within perfectly acceptable levels too. Nothing wrong in my opinion.
 
I'm not seeing any hot pixels in your images. Hot pixels show as red, green or blue (and pixel sized, funnily enough). Is that what you're seeing? I may not be seeing anything because you've zoomed right in and know where they are and I can't zoom enough. I do astrophotography with an old Canon 350D and unless I dark frame (take the same length exposure at the same temperature with the body or lens cap on) I get loads of hot pixels. It's one of several reasons why a lot of astrophotography is done with cooled cameras. I found the LENR applied in camera wasn't effective at getting rid of them (and it doubles the length of every exposure). Noise that you get from underexposure is something completely different. My 550D is terrible for that.
 
Thanks for the replies... right, those photos with the boats in... I thought that was my first long exposure that evening but it wasn't ... the first one was actually this one....

111 seconds, and no hot pixels. All good... shadows boosted to 80 percent as photo was underexposed..

IMG_9734 by Joel Spencer, on Flickr

The next one is 168 seconds, hot pixels showing... no processing done.

IMG_9737 by Joel Spencer, on Flickr

The last photo is 365 seconds.. no processing done. You can clearly see blue red and green hot pixels now.

IMG_9738 by Joel Spencer, on Flickr

Image 1 was definitely the first long exposure that day/evening.

I still don't think that I should be getting _that_ many hot pixels ? What if I want to shoot star trails which take 20-30 minute exposures?
 
Last edited:
There does seem to be a lot of hot pixels in the last one.

I’ve never had a camera that didn’t have hot pixels but not enough to worry about.

If your cameras under warranty then it might be worth seeing if you can get the sensor changed. If not then you have two options, try and map them out or don’t take such long exposures.

As for shooting star trails, you wouldn’t shoot a 20-30 minute exposure anyway. You’d shoot multiple 30 second exposures and then stack them together. Shooting one long exposure means that any flaw in the exposure, such as car headlights passing briefly will spoil the shot. By using multiple short exposures you can either drop the frame that the lights are in or use a different frame without them and blend it in.
 
Gaz J said:
There does seem to be a lot of hot pixels in the last one.

I’ve never had a camera that didn’t have hot pixels but not enough to worry about.

If your cameras under warranty then it might be worth seeing if you can get the sensor changed. If not then you have two options, try and map them out or don’t take such long exposures.

As for shooting star trails, you wouldn’t shoot a 20-30 minute exposure anyway. You’d shoot multiple 30 second exposures and then stack them together. Shooting one long exposure means that any flaw in the exposure, such as car headlights passing briefly will spoil the shot. By using multiple short exposures you can either drop the frame that the lights are in or use a different frame without them and blend it in.
Click to expand...

I think the remapping wouldn't apply though as I only see the hot pixels when doing minutes long exposures.
 
You could also get the lee little stopper to save on using such a long exposure.
 
You must log in or register to reply here.
Back
Top