I have been using ICNR for a while as I was having difficulty with dark farmes not removing amp glow or handling warm/cool pixels well. I have attached a couple of frames to show you what I mean, they are both at iso 800 with a Canon 20d. The ICNR is a stack of 11*300s (total of 110 mins of "sky" time), stacked and aligned in IP 2.8 then I used Auto in the DD option - I realise this overcooks it but wanted to show up the noise, finally loaded into PS, scaled down to 25% and saved as a JPEG at about 190Kb ... it looks a bit grainy but not bad....
The second shot is 19*360s (114 mins of "sky" time). Processing steps taken were identical, amp glow, band of red, streaks in the nebulosity. In short pretty horrid.
Is this normal?
My (16) darks were taken (outside) at the end of the imaging session, of course the temp was not identical but I cannot control temp so is there a better way to process these? I have read about adaptive dark frame matching but am uncertain as to how this is perfomed in IP.
Pretty conclusive from my pov - for longer (> 3 mins) exposures with a 20d ICNR is better.
From my previous work I believe I might be better with darks at <180s but I cannot prove this. Overall therefore I suspect my best option is 1600 iso, sticking to 3 mins exposures and stacking more frames perhaps shooting darks at the meridian flip or when switching targets (ie in the middle of a session?).
John, ICNR, is the only way to go it will assure you that your darks are matched as close as possible to your light frames.
As for that band of red across your second image, I always thought the amp glow appeared on the right hand edge of the captured frame, please correct me if i am wrong.
I do ICNR as well but really would like to be able to take darks separately. Many people do it so I'm sure its possible. I guess I need to take the darks in-between the lights, instead of at the end (for example).
Maybe 1 dark every 2 lights. The problem with that is, it doesn't allow you to do a completely automated run, because you need to be there to put the lens cap on, setup the new exposures, etc.
At the moment with ICNR, I can set up a 2 hour run and go to bed, set the alarm for 2 hours later and point at a new target.
John, ICNR, is the only way to go it will assure you that your darks are matched as close as possible to your light frames.
As for that band of red across your second image, I always thought the amp glow appeared on the right hand edge of the captured frame, please correct me if i am wrong.
Leon
Leon, you are quite right, the amp glow is the red doughnut shape on the bottom RHS of the frame, the band in the middle is the infamous "dark river" for which there is no cure outside of ICNR...except perhaps a different camera.
At the moment with ICNR, I can set up a 2 hour run and go to bed, set the alarm for 2 hours later and point at a new target.
That is my approach also, just wanted to check I was not mad - some folks do use darks with good effect, maybe thay all use modded cameras and do not have to push the data so hard...
One more observation on darks taken after an imaging run. The .CR2 files increase in size quite quickly and then seem to stabilise. This (I assume) is the effect of sensor temp. My thought is that during the considerable gap between the imaging run and the darks the camera sensor cools off, then during imaging it heats up. Again...here is an eg:
As you can see both the lights and darks increase in size over the run - in the darks the grouth is 1Mb - I assume all noise so the problem I have is I normally shoot less darks than lights so the average dark temp is much less than the average light temp. I think using the last 3 or 4 darks might give a better result in this case however it looks like it takes about 16 shots for the temp to get close to stable thus best results would be with an equal number of lights and darks or failing that use the last 1/3 of the dark run frames (but not less than 8)? Any image processing theorists out there care to comment?
A second idea would be to increase the interval between shots to prevent heat buildup - however I expect that interval would "waste" as much time as ICNR.
Final thought - an external (usb?) cooling fan - not active cooling but just just an attempt to xfer the excess heat away from the sensor so that there is less heat build up over the runs.
I think the later model DSLRs are immune to this issue (amp is off for long exposures on the 40d - is that correct? Which other models?) I even think I recall seeing an amp off mod for the earlier canons somewhere on the net...
