Hints for DSO exposure/gain/offset using ASI1600MM-C

[iapa]

I've moved from a Canon 600D to ASI 1600mm cooled and was wondering about exposure and gain settings.

My most frequent targets are M27/51/81/82 and I was using 600s ISO800 with the DSLR

However, I often seen comments that the ASI would be better at more, shorter exposures.

Save wasting time, hey, why repeat work already done? :), I was wondering if anyone had suggestions for these targets (using LRGB at the moment).

 

 

[vlaiv]

Here is what I currently use (only had a first light with it, but these settings worked rather well on a first try):

60s exposure

Unity gain - 139 (13.9db), at -20C, offset 25

x64 dark frames (will continue to add on subsequent sessions, because I usually leave my gear the following day in basement where it's cooler to gather darks for hour or two).

x256 Flats / x256 Flat darks

Note: I image from red zone (~18 mag skies) with use of Hutech IDAS LPS P2.

This is L only, no RGB filters yet

HTH

[iapa]

@vlaiv thanks of that. I started at 600, but with the lack of dark I was getting 3 of 4 images a night

[vlaiv]

5 hours ago, iapa said:

@vlaiv thanks of that. I started at 600, but with the lack of dark I was getting 3 of 4 images a night

Glad you found my post useful, in general, going higher count of shorter subs rather than fewer longer ones will not benefit you in terms of signal to noise ratio, so you are stuck with total exposure per night dictated by astro dark. On my location (45 degrees), it is something like less than 3 hours at shortest period, but I guess it is still less than that up north.

What it does help with is following: given 12bit ADC and unity gain setting you are limited to 4096 - offset "full well" capacity, and depending on LP and target you might start saturating stars real quick, so short exposure is good in handling this. Low read noise of camera helps with not loosing much in terms of SNR over long subs - if read noise was 0 (and everything else was linear with integration time, and we can assume that it is close at lease) it would not matter how many subs you take - it would be the same as single sub of total duration (bar full well and things like that), so having low read noise does affect this but much less than CCD cameras having 5-10 times more read noise than ASI1600. It also helps with stacking (sigma clip stacking likes many subs), less integration time will be discarded in case of bad guiding / wind / whatever mechanical glitch, and one can take much more calibration frames for masters (helps with reduction of noise, you can take 64 darks in one hour, and that will lower noise injected back into frames when calibrating by x8 for example).

Down side of having large number of frames is that it takes huge amount of space (multiple gigs of data), and one really needs to do processing in at least 32bits (DSS is out of the question here due to precision and memory management issues).

[iapa]

8 hours ago, vlaiv said:

Glad you found my post useful, in general, going higher count of shorter subs rather than fewer longer ones will not benefit you in terms of signal to noise ratio, so you are stuck with total exposure per night dictated by astro dark. On my location (45 degrees), it is something like less than 3 hours at shortest period, but I guess it is still less than that up north.

 

 0hrs astro dark, in fact, 0hrs nautical dark right now.

To help with noise, I also dither every frame (600s) but if I drop to, say, 60s, I may change that to dithering every 10 frames otherwise I could spend more time waiting for the scope and guiding to settle than imaging.

[Andyb90]

How is the light pollution where you're imaging from? Also what f-ratio are you imaging at?

Are you using Baader filters?

I'm in an Orange/Yellow area and find my L subs get over exposed quite quickly. I've used an LP filter in the past, but recently new streetlights were put in and the filter doesn't seem to be as effective.

Next time I try my L filter I'm going to have a go with gain 0 or 75 and 60 second exposures.

I image at f5.5, but from what I understand the faster your f-ratio the shorter your exposures would be.

Andy.

[vlaiv]

3 hours ago, iapa said:

 0hrs astro dark, in fact, 0hrs nautical dark right now.

To help with noise, I also dither every frame (600s) but if I drop to, say, 60s, I may change that to dithering every 10 frames otherwise I could spend more time waiting for the scope and guiding to settle than imaging.

I didn't realize things get that bad when one moves just a bit up north (or maybe it is "just a bit" only in my mind :D, guess difference is more like 10 degrees?).

For dithering, yes I suppose you can let it dither every 5 - 10 frames, don't know how much time each move takes, I only dithered "naturally" before - less then perfect guiding / differential flex - it caused drift of 10-20 pixels over the course of couple of hours - not noticeable on a single frame when exposure is in range 1-4m. I'll probably start dithering myself now since I'm switching to OAG for my long fl scope (yet to have proper first light, I only managed to find focus position and setup everything on first night - now just waiting for good time to try it out).

I think that number of subs between dither moves has relation to parameters of sigma / kappa-sigma stacking. It helps remove hot/cold pixels and their artifacts from image. So if you have 5 frames between dither move you want to setup stacking algorithm to remove 5 outliers - thus excluding 5 hot pixels that happen to fall on a certain position when stacking. If you shoot 10 frames between moves - you need to make sure it will remove all 10 instances of hot pixel on a given position. Well you don't have to be this precise because of total number of frames increases significantly with short exposures so if you miss a few of the hot pixels with given settings it will be offset by number of frames that go into average (for example you dither on 8 frames, your kappa sigma settings remove 6 outliers, so two remain and they end up in a stack of 120 frames - that is less than 2%)

[iapa]

16 minutes ago, vlaiv said:

I didn't realize things get that bad when one moves just a bit up north (or maybe it is "just a bit" only in my mind :D, guess difference is more like 10 degrees?).

this was last month

16 minutes ago, vlaiv said:

For dithering, yes I suppose you can let it dither every 5 - 10 frames, don't know how much time each move takes, I only dithered "naturally" before - less then perfect guiding / differential flex - it caused drift of 10-20 pixels over the course of couple of hours - not noticeable on a single frame when exposure is in range 1-4m. I'll probably start dithering myself now since I'm switching to OAG for my long fl scope (yet to have proper first light, I only managed to find focus position and setup everything on first night - now just waiting for good time to try it out).

I think that number of subs between dither moves has relation to parameters of sigma / kappa-sigma stacking. It helps remove hot/cold pixels and their artifacts from image. So if you have 5 frames between dither move you want to setup stacking algorithm to remove 5 outliers - thus excluding 5 hot pixels that happen to fall on a certain position when stacking. If you shoot 10 frames between moves - you need to make sure it will remove all 10 instances of hot pixel on a given position. Well you don't have to be this precise because of total number of frames increases significantly with short exposures so if you miss a few of the hot pixels with given settings it will be offset by number of frames that go into average (for example you dither on 8 frames, your kappa sigma settings remove 6 outliers, so two remain and they end up in a stack of 120 frames - that is less than 2%)

Of course, it I am down under 120s subs, I could just turn guiding off and let it drift naturally