Narrowband integration time ?

[knobby]

Hi all, I'm sure I've read this before somewhere but still unsure of the theory ...

 

Am I correct in thinking that when stacking, 1 to 2 Half's the noise, then 2 to 4 Half's again, then 4 to 8 ... 8 to 16 etc.

 

So going past 16 really needs a massive investment in time ?

 

Also where does sub length come into this ? I use a ZWO 1600 pro and seem to get good results in my bortle 7 / 8 zone with 300 seconds at unity gain.

 

Also, now sure that (as I've been told but wanted to see for myself 😉) that only Ha cuts through the Moon glow.

 

Due to time restraints I've been aiming for an hour per filter but this equates to 12 subs so I guess I need to grab an extra 4 to see the benefit ?

 

For completeness and a recap ...

 

300 seconds / gain 139 / ZWO 1600 pro / f4.9 ZWO narrowband filters.

[Adam1234]

The more subs the merrier, I try to get at least 5 hours per filter, I'm not exactly sure the point of finishing returns is bit I think it is way above 16 subs. 

This video may be of some help to 

 

 

There is also a thread on cloudy nights regarding optimum subs length, if you Google asi1600 exposure times

 

[ollypenrice]

The sub length question has been re-answered by CMOS cameras which work well with short subs. High read-noise CCDs do not. CCDs do, from experience rather than theory, benefit from long subs. Plenty of CMOS imagers are finding 300secs at unity gain to be effective. In the end the key number is the total integration time and even with the fastest setup known to man you will be scratching the surface with half an hour per filter. The fastest setup currently appearing regularly on SGL is probably Gorann's RASA 8 and CMOS rig. Check out his posts and integration times.

Noise increases as the square root of the signal. What's less clear is what this really means in terms of a particular photo project. We don't stretch each project in the same way. Sometimes we want to wring the neck out of faint signal, sometimes we don't. Sometimes heavy handed noise reduction on certain parts of the image will be unobtrusive and acceptable, sometimes it won't. (How much structure is there in this faint signal? The more there is, the less acceptable NR routines will be.)

The theory will only take you so far. You'll meet the reality when you start to work on your project but half an hour per filter will not produce a great image with present technology..

Olly

[newbie alert]

I've been told from someone else about a certain amount of subs with diminishing returns,  don't you think that the 16 you want will get you off the noise floor... more to capture than just getting you out of the noise band.. remember with noise it's your camera creating it , you will get bored way before you reach that area of diminishing returns 

[JSeaman]

I have been experimenting with this over the last few nights with a similar set up, well, near identical actually. I also have a 300 second exposure with unity gain (139) on a ZWO 1600 pro at f4.9 but I am using Baader filters and a 12 inch scope. That said, the differences are outweighed by the similarities. I have experimented with sub quantity, binning and duration and I am settled (on this camera/aperture) at 300 second subs.

The attached video gives you an example of what stacking does, it doesn't halve the noise every time and it is diminishing returns but the invesment lets you do more with the data. I would say you need at least a couple of hours worth of data on each filter and up to 6 is great. 

When you zoom right in on the pictures you can see the noise reducing but the difference between ~40 pics (3.5 hrs) and 80 pics (~6.5 hrs) isn't huge.

 

[knobby]

Thanks gents, interesting replies, I've been stubborn in sticking with 300 sec / 139 gain ... might have a dabble ( and new darks library 🙂 ) going down to 180 seconds as @Blazar has done the maths from Robin Glovers video which implies 170 ish is optimal for my situation.

Probably get less subs lost by wind  / satellites etc 

[Blazar]

29 minutes ago, knobby said:

Thanks gents, interesting replies, I've been stubborn in sticking with 300 sec / 139 gain ... might have a dabble ( and new darks library 🙂 ) going down to 180 seconds as @Blazar has done the maths from Robin Glovers video which implies 170 ish is optimal for my situation.

Probably get less subs lost by wind  / satellites etc 

Well its a start 😉 . I never said I was any good at math of course 🙂 kind of makes sense though looking through it all.

[souls33k3r]

I've always been in the 300s sub exposure with unity gain kinda guy. I could've used the SharpCap smart assistance or whatever that tool is or use Robins formula. I was in a bit of a rush so got a mate of mine to punch in some numbers using my F/5.5 in Bortle 8 skies and that came out to be 245s so using 300s subs, I wasn't too far off. But there was this itch I just had to scratch so I went with a 600s sub. Tbh I can see so much more nebulosity with a 600s sub (of course it's double to that of a single 300s sub) but when I have that mate 2x300s subs and compare that against a single 600s sub, the stars were just a tad bit bigger but only if you're pixel peeping, noise was almost the same tbh (single subs show the single 600s sub having less noise) but the Signal was evidently better. 

I might start doing 600s subs from now on and like you said there's always that risk of risk of losing a sub because of whatever reason but truth be told I've hardly ever discarded a sub because of my gear, more to do with clouds. 

Hope this helps matey. 

[Davey-T]

15 minutes ago, souls33k3r said:

I might start doing 600s subs from now on and like you said there's always that risk of risk of losing a sub because of whatever reason but truth be told I've hardly ever discarded a sub because of my gear, more to do with clouds. 

All other things being equal its worth remembering given the file size of bigger sensor CMOS cameras that a 300sec' and 600sec' sub have the same file size.

Dave

[souls33k3r]

8 hours ago, Davey-T said:

All other things being equal its worth remembering given the file size of bigger sensor CMOS cameras that a 300sec' and 600sec' sub have the same file size.

Dave

Absolutely but truth be told, the file size has never really bothered me much, only because I've not known a life in astrophotography with a smaller file size. I remember seeing Alex's face when he asked me the file size of my camera because he's used to around 1MB of a file size 😂 

What does bother me is the 294MM file size 47MB (bin 2x2) and 97MB (bin1) and I suppose the 2600MM is going to be even more crazy 😁

[Davey-T]

 

3 hours ago, souls33k3r said:

What does bother me is the 294MM file size 47MB (bin 2x2) and 97MB (bin1) and I suppose the 2600MM is going to be even more crazy

Atik 314L 709kb ZWO2600MM 50971kb

Dave

[JSeaman]

Having just moved from a 314l to the 1600 I can tell you the file size is justified by the image size and resolution, the 1600 is so much better!

[Davey-T]

8 minutes ago, JSeaman said:

Having just moved from a 314l to the 1600 I can tell you the file size is justified by the image size and resolution, the 1600 is so much better!

Depends on the attached scope, I've had Atik 314L on the 10"SCT for years and no intention of changing it,  got ZWO2600MM on Sharpstar 150HNT f/2.8.

Dave

[JSeaman]

Scope is irrelevant to resolution improvement, the 1600 is a better camera

[Davey-T]

48 minutes ago, JSeaman said:

Scope is irrelevant to resolution improvement, the 1600 is a better camera

Not really, SCT is already oversampled with unbinned 314l. not sure in what way it's a "better" camera.

Dave

[JSeaman]

Sensitivity, sensor size, read noise, current, ASCOM integration to name a few

Anyway, side tracked into a different conversation, regarding the file size, you are sending more data so it's a bigger file, completely expected

[souls33k3r]

1 hour ago, Davey-T said:

 

Atik 314L 709kb ZWO2600MM 50971kb

Dave

LOL yeah according to my calculations it is around 52MB as well which is crazy unless you have a super fast machine that can handle the processing of such huge file size. I'm sure it won't be an issue for many. 

 

Any way, I'm not going to distract the Ops topic any further

Edited March 3, 2021 by souls33k3r

[Sp@ce_d]

314L+ is a super little camera & for many of us our first mono CCD.. doesn't really need darks & quick to preprocess. Ended up with 3 of them. One lived on my RC10 for over 5 years!  Even though I've not used them for a while now I still can't bring myself to let them go

I was experimenting with first light on my ASI 2600MC under the full Moon with an IDAS-NBX & decided 180 subs was best compromise although I'll probably go 300 when there's no Moon about.

However, when it comes to processing.. wow.. I reckon I'll need to add a week onto the time for calibration, noise reduction etc. Gone are the days of a quick process with a few hours of each channel with the 314L+ & QSI683 files. It's almost a case of kick it of & come back tomorrow.. & that's on a 10 core i9 with 128GB & the fans running full pelt!

[JSeaman]

Are you guys really finding things that slow?

I just ran a test on 90 subs, 16 darks, 21 flats, 21 dark flat and 51 bias from scratch and it took a little under 4.5 minutes with 50/60% CPU usage
 

[CraigT82]

6 hours ago, JSeaman said:

Scope is irrelevant to resolution improvement, the 1600 is a better camera

Depends on what you mean by resolution. 

[JSeaman]

Again I don't really want to be hijacking the thread but can you give any definition of resolution where a 314 beats a 1600?

[knobby]

Feel free to hijack 😃 I love a bit of banter