Stuart Taylor:
I'm going to re-shoot my dark libraries. Many poeple suggest that the cover on the 2600 isn't necessarily light proof, so they suggest putting the camera in a black bag to shoot the darks. But I need to cool it to my usual working temperature to match the darks, so won't putting it in a bag interfere with the cooling? I am worried the camera will overheat if I put it in a bag.
What do you do?
Pedro A. Sampaio:
I'm not suggesting it doesn't happen to other people, however I've never had any issues with light leaks in my dark frames. I mostly shoot darks with the camera cooled to the desired temperature, in a natural lit room during the day. Never had any issues with my darks.
I'd suggest you try during the day, as normal, and see first if there's a light leak - by stretching the darks, for instance.
If there's no problem, then you don't have to "solve it" beforehand.
CS
Stuart Taylor:Pedro A. Sampaio:
I'm not suggesting it doesn't happen to other people, however I've never had any issues with light leaks in my dark frames. I mostly shoot darks with the camera cooled to the desired temperature, in a natural lit room during the day. Never had any issues with my darks.
I'd suggest you try during the day, as normal, and see first if there's a light leak - by stretching the darks, for instance.
If there's no problem, then you don't have to "solve it" beforehand.
CS
Thanks. My master dark has these stats (with an offset of 50). Does it look like there is any light leak, would you say? The cap on the ASI2600 looks pretty tight as it's a screw in with an O ring
count (%) 100.00000
count (px) 26091648
mean 500.328
median 500.280
avgDev 1.092
MAD 0.920
minimum 468.100
maximum 23832.271
Pedro A. Sampaio:Stuart Taylor:Pedro A. Sampaio:
I'm not suggesting it doesn't happen to other people, however I've never had any issues with light leaks in my dark frames. I mostly shoot darks with the camera cooled to the desired temperature, in a natural lit room during the day. Never had any issues with my darks.
I'd suggest you try during the day, as normal, and see first if there's a light leak - by stretching the darks, for instance.
If there's no problem, then you don't have to "solve it" beforehand.
CS
Thanks. My master dark has these stats (with an offset of 50). Does it look like there is any light leak, would you say? The cap on the ASI2600 looks pretty tight as it's a screw in with an O ring
count (%) 100.00000
count (px) 26091648
mean 500.328
median 500.280
avgDev 1.092
MAD 0.920
minimum 468.100
maximum 23832.271
It's difficult to judge just from the statistics themselves... That being said, although I'm not an expert, I don't think there'se anything wrong with it. You have most of your pixels in the very lower end of the histogram, without being zero (ie clipped). You do have some max/near max pixels, but that again, most cameras have hot pixels that register like that, nohting out of the ordinary in my opinion.
You should try to take a dark frame, and stretch it with PI or PS just enough that the dark pixels are visible, and see if there's anything weird - a proper dark frame stretched shoud be relatively homogeneous, with no bright sides or streaks. If you are unsure, try to upload one so we can check for ourselves.
Andy Wray:
I usually end up doing darks at the target cooling temperature inside after I go to bed with all the lights switched off inside and remembering to turn off the outside security lights.
FWIW: I'm surprised your ASI2600 16 bit ADC darks have a median of 500 with an offset of 50; that does seem high.
My ASI1600MM pro is only a 12 bit ADC and my darks have a median of about 815 at the same 50 offset (300 sec exposure). Given I have only a 12 bit ADC, I would expect my best median to be 800 with an offset of 50 ( i.e. (50/4096)*65536 = 800 ), so I am happy with my darkness.
I'm probably comparing apples with pears here and actually, at the end of the day, you do have darker darks than I do ;)
One last thought: you didn't say what gain you were operating at. Looking at the numbers you posted, you may be able to reduce the offset as your minimum values are relatively high compared to your median.
)Pedro A. Sampaio:
IMO, you should not have only 50 (off-set value) as the mean. Remember, darks record read noise and thermal noise, besides some incident radiation (i.e. cosmic rays). If the mean expected value was only 50, it would mean you had 0 read noise, which we know is not the case. I find it highly unlikely that a light leak would turn up to be an homogeneous distribution of light. Most likely, it would result in an asymmetric illumination of the sensor, something akin to Amp glow like we see in other cameras
Stuart Taylor:Pedro A. Sampaio:
IMO, you should not have only 50 (off-set value) as the mean. Remember, darks record read noise and thermal noise, besides some incident radiation (i.e. cosmic rays). If the mean expected value was only 50, it would mean you had 0 read noise, which we know is not the case. I find it highly unlikely that a light leak would turn up to be an homogeneous distribution of light. Most likely, it would result in an asymmetric illumination of the sensor, something akin to Amp glow like we see in other cameras
I don't understand what you are saying. My mean is not 50. It is 500. (see my post above with the stats)
Carl Newberg:
I use a small ice chest with some cooling packs (hard sided so they can't block the camera vents). I run the power cord and USB cable out and collect away for as long as needed during the day.
As an aside, this discussion was interesting - I found that when I calibrated my Ha data with my last set of darks, most of the signal went with it. Ended up not using darks on that data set - wondering if I need an offset as mentioned above????
Pedro A. Sampaio:Carl Newberg:
I use a small ice chest with some cooling packs (hard sided so they can't block the camera vents). I run the power cord and USB cable out and collect away for as long as needed during the day.
As an aside, this discussion was interesting - I found that when I calibrated my Ha data with my last set of darks, most of the signal went with it. Ended up not using darks on that data set - wondering if I need an offset as mentioned above????
Did you use a pedestal during the pre-processing? That usuaaly resolves the issue.
Specially with CMOS NB imaging, some pixel values in light frames can be very low, even with a small offset. During calibration, Dark frames (actually the Master Dark) pixel values are subtracted from all light frames in order to remove shot noise from the lights. That means that some values can reach 0, wich is very bad and can greatly degrade the image. Because of this, it is common to use a "pedestal" (something like 100-200) for each light frame pixel before darks are subtracted.
For anyone interested in the subject matter I highly recommended Adam Blocks EXCELLENT video series on the pre processing in pixinsight.
Pedro A. Sampaio:Stuart Taylor:Pedro A. Sampaio:
IMO, you should not have only 50 (off-set value) as the mean. Remember, darks record read noise and thermal noise, besides some incident radiation (i.e. cosmic rays). If the mean expected value was only 50, it would mean you had 0 read noise, which we know is not the case. I find it highly unlikely that a light leak would turn up to be an homogeneous distribution of light. Most likely, it would result in an asymmetric illumination of the sensor, something akin to Amp glow like we see in other cameras
I don't understand what you are saying. My mean is not 50. It is 500. (see my post above with the stats)
Sorry about the confusion. I was actually referring to the comment Andy posted regarding the Mean pixel level being 500 with your offset as 50.
My point is that even with an offset of 50, a mean value of 500 per pixel (out of 65535) isn't high at all, it's quite normal for a proper dark frame. So, in my view, I think you're fine taking darks the way you've been taking them until now, and shouldn't worry about it.
-- MORE IN DEPTH EXPLANATION (or the way I understand it) --
Since we're talking about a 16bit depth camera (ASI2600), that means that each pixel can have a value from 0 to 65535 (or 2^16, that's why it's 16 bit). A value of 0 for a pixel is realy bad for editing, because most image processing software use multiplication and similar operations to "work" on the image. 0 times any number is always 0. So a good way to remedy this is to use an offset, i.e. add 50 or whatever other number you want to ALL pixels, in order to make all 0 level pixels a number higher than 0. Those 0 level pixels can be dead pixels or simply pixels that didn't receive enough photons. In any case, an offset would add to the "mean" value of pixels because it is additive to all pixels of the sensor. 50 for all practical purposes still 99,9% pure black, but since it's different than 0, it's not clipped and can be better used by editing software.
But darks don't capture just the offset. The purpose of dark frames is to capture shot noise, read noise, thermal noise and all noise related to the eletronics of the sensor. That's where a mean value of 500 would come from.
