119 replies to this topic

[idclimber]

was that for my Sub at 120 seconds ?

 

Bob seems to think thats OK ?

It is on the high end of what I find acceptable if I am trying or concerned about preserving star color. If I plan on substituting stars with different exposures or with my broadband filters then I may go higher. The problem going longer is that SXT sometimes does not remove the clipped stars from the image and then it is a lot more difficult to add them back in from another source. 

 

IMHO the metric to look at is pixels clipped. It is far more consistant with our dark skies. Mean or median ADU was problematic for me when I first got started in imaging. Once I move to clipped, my exposures and images improved. 

 

Robin Glover of SharpCap fame has what is regarded as one of the best presentations on the subject. It is the work that Jon Rista and others have used that I have followed in setting my exposures. Here is a link below. This is a very good presentation and well worth watching. 

 

https://www.youtube....RH93UvP358&t=2s

 

 

Again IMHO trying to do the math to know if you are 4x or 10x read noise is not for those who took math four decades ago and don't use it now. As such this often quoted metric is a lot harder justify in the beginners forum. 

 

Probably the easiest way to know if you are overexposed is to look at the sub in the linear un-stretched state. If it looks black with only a few hints of stars you are not overexposed. You may even be able to go higher, especially if the second check shows only 100 pixels clipped. You image from this post passes this test with flying colors, so much so I thought PixIsnsight was not giving me the correct answer. I opened it in Voyager to confirm my results. 

 

 

That's bad advice. Underexposing will lower the dynamic range for the object you are inaging by bringing the faint detail into the read noise bits. That's bad. As I pointed out, once you're not swamping noise, it takes forever to get enough exposure to bring it back out.
Best advice, as Greg, Gilmour and I pointed out, is to expose so the faint details are above the noise floor. Sometimes that means overexposing the stars a little, which you can correct by combining with short exposures for stars only. A far better way to image than what you suggest.

 

The 2600mc especially with the gain 100 has incredibly low read noise. It really does not take that long of an exposure to swamp it. This gives us a lot of latitude in exposures that were nearly identically on the dim details. 

[WadeH237]

That would put this exposure at the upper limit of that range. 

There is no "upper limit".

 

The 4x to 10x recommendation is an estimate.  Once you get below 4x, you will see some effects from read noise.  Once you get past 10x, any difference is so small as to be "why bother"?

 

Oh, and remember that the context of the advice from Jon is to determine the minimum exposure time to beat read noise.  It has nothing to do with determining the maximum exposure time.

 

Off the top of my head, the factors that go into determining the maximum exposure time are as follows:

 

• Can your mount track accurately enough?
• Are you willing to lose a very long sub in the case of some problem that interrupts things?
• Is your background sky bright enough that it limits effective dynamic range?
• Is there an excessive amount of saturation?

 

I think that it's the 4th one above where we disagree.

 

If I understand your position correctly, you are looking for no more than a handful of saturated pixels, possibly without regard to which pixels are saturated.

 

My position is that the total number of saturated pixels is just one factor.  I think that it's actually more important to consider which pixels are saturated.  For example, if hot pixels are making it through calibration and showing as saturated, then altering exposure time will have little to no effect on them.  Saturated pixels that are in the cores of bright stars are simply not a problem, if they are concentrated at the very center of the star, with enough surrounding halo to bring in star color.  Saturated pixels in the object of interest are pretty much never OK.

 

Note that I have changed my position regarding saturated pixels in star cores.  I used to think that any saturated pixels were bad.  I changed my opinion after going through enough Adam Block tutorials.  Much of his data comes from systems that can produce huge stars with very saturated cores.  Yet I can't think of a single instance where he's done anything to try and either avoid, or correct, those saturated cores.  He  does have some examples of blending short and long exposures to correct for saturation, but only in bright objects, never stars.  He's verbally mentioned many times that the cores are fine.  Naturally that is his opinion, but I tend to trust him.

[dciobota]

Didn't realize I forgot to quote idclimber, I was responding to him.

Yeah but lowering exposure will put some of the faint signal "in the noise". That's where the dynamic range comes in. Star brightness takes up a lot of the dynamic range in an image (as evident in the clipped value vs median adu value) so every bit (pun intended) above the read noise helps, no matter how small that is.
But what you said is fair, lowering the exposure at gain 100 you won't lose a whole lot.

Edited by dciobota, 22 September 2023 - 07:08 PM.

[WadeH237]

Shouldn't the read noise be squared? Not that 2.32 is that different from 1.52...

 

 

Wade, I think you need to subtract the Bias/Offset before converting from ADU to e-. If you do that, the image seems slightly underexposed (what I‘ld expect from Bortle 1 with those settings, actually). i‘m assuming the ZWO standard offset 50 in the driver here:

 

Signal: 162 ADU = 40.8 e-

Shot noise = 6.39 e-

Shot noise/read noise^2 = 2.8:1

 

I try and get to 10:1, but depending on the target you might blow out more stars. 300s seems to be a good, round number.

 

It‘s also what I would have used intuitively in B1 with that chip. In B6, I routinely use 180s exposures (for mono RGB) at F/6.4 - that‘s a bit long, but i prefer not to deal with a myriad of subs and depending on target will just add a few 30s subs for the brightest stars.

Thanks for the math corrections!

[Robert7980]

I’m going to recommend trying 300 seconds again… No harm in trying it… 

 

Looking through all the images I’ve taken with the 2600s it’s the 5 minute subs that stand out as superior in almost every case to my eye… That seems to be the sweet spot for my taste and workflow… The noise is almost non-existent compared to anything I’ve shot at 30 to 120 seconds what’s different is there’s a huge difference in signal… 600 seconds is even better, but it gets impractical going that long for other reasons… 

This was a single unprocessed shot at 10 minutes and it was UNCOOLED!… The camera performance is totally ridiculous lol… I think we are splitting hairs when it comes to the 2600, I don’t think the camera really cares what abuse we put it through… It’s dang hard to really overdo things with this beast… Not that it’s impossible to cook the data, it’s just not an easy thing to do… Had the cooler been on with a little work processing it could almost pass as a project and it’s a single sub… 

 

 

One 300 second sub with the cooler actually ON  … Pretty easy to get to practically ZERO noise when pushing into it’s design threshold exposure range… 

 

 

Be bold, push the limits. Don’t underestimate why this camera and sensor is the most popular hotness in the history of AP… It’s almost cheating, almost  — Push it hard and dig that faint stuff out for free! Once the shutter is open you’ve taken your noise hit, everything after that costs you nothing to keep it open, why stack terabytes of subs if you don’t have to… 

[Zambiadarkskies]

I’m going to recommend trying 300 seconds again… No harm in trying it… 

 

Looking through all the images I’ve taken with the 2600s it’s the 5 minute subs that stand out as superior in almost every case to my eye… That seems to be the sweet spot for my taste and workflow… The noise is almost non-existent compared to anything I’ve shot at 30 to 120 seconds what’s different is there’s a huge difference in signal… 600 seconds is even better, but it gets impractical going that long for other reasons… 

This was a single unprocessed shot at 10 minutes and it was UNCOOLED!… The camera performance is totally ridiculous lol… I think we are splitting hairs when it comes to the 2600, I don’t think the camera really cares what abuse we put it through… It’s dang hard to really overdo things with this beast… Not that it’s impossible to cook the data, it’s just not an easy thing to do… Had the cooler been on with a little work processing it could almost pass as a project and it’s a single sub… 

 

 

One 300 second sub with the cooler actually ON  … Pretty easy to get to practically ZERO noise when pushing into it’s design threshold exposure range… 

 

 

Be bold, push the limits. Don’t underestimate why this camera and sensor is the most popular hotness in the history of AP… It’s almost cheating, almost  — Push it hard and dig that faint stuff out for free! Once the shutter is open you’ve taken your noise hit, everything after that costs you nothing to keep it open, why stack terabytes of subs if you don’t have to… 

 

Very eloquently put!  

[JSTAR0057]

I am in a bortle 4-5ish zone and shoot 300s subs with my asi2600mc-p. I keep my Gain at 101.

 

[dciobota]

I am in a bortle 4-5ish zone and shoot 300s subs with my asi2600mc-p. I keep my Gain at 101.

You can safely use gain 100, since the high gain low read mode kicks in at 100.

I too second and third 300s using the osc version, unfiltered, provided your skies are reasonably dark.

[Oort Cloud]

300s will be too short at f/10, and too long at f/2.

How is that it's "design threshold exposure range", and why haven't I heard this term presented by any of the companies that actually manufactured these cameras?

I'll tell you why, because it's designed to take exposures from 32us to 2000s.

Take a 300s shot of the moon, or a 300s exposure of pretty much anything at f/2.

Overexposure and underexposure are both real things. There is no magic number, it depends on scene brightness and f/ratio, just like in "normal" photography.

[Spaceman 56]

I guess these well intentioned suggestions to shoot 300 second subs, are coming from imagers who regularly shoot with Filters.

 

I am shooting from Bortle 1 without filters. none at all.

 

surely the filters would increase the sub exposure time requirement, as they can cut out 97% of the incoming light.

 

am I missing something ?

[Spaceman 56]

I’m going to recommend trying 300 seconds again… No harm in trying it… 

 

Looking through all the images I’ve taken with the 2600s it’s the 5 minute subs that stand out as superior in almost every case to my eye…

are those 5 minute subs without filters Robert ?

[Rasfahan]

are those 5 minute subs without filters Robert ?

I‘m not Robert, but I suggested you try 300s subs. You‘re using filters - they‘re Bayer filters on your camera chips. They have a bit less transmission than the RGB filters I use for mono - and I regularly shoot 300s RGB subs in Bortle 6.

 

Another argument: The IMX571 chip has a higher full well capacity than the KAF-16809 per area. With the latter, 5 min subs were considered short. 
 

You have dark skies - shoot long subs to take advantage and don‘t drown the faint details in read noise.

Edited by Rasfahan, 23 September 2023 - 05:21 PM.

[Oort Cloud]

I guess these well intentioned suggestions to shoot 300 second subs, are coming from imagers who regularly shoot with Filters.

I am shooting from Bortle 1 without filters. none at all.

surely the filters would increase the sub exposure time requirement, as they can cut out 97% of the incoming light.

am I missing something ?

The biggest factor is optical speed. There is a 25x difference between f/2 and f/10.

And yes, filters.

Edited by Oort Cloud, 23 September 2023 - 09:40 PM.

[Zambiadarkskies]

I guess these well intentioned suggestions to shoot 300 second subs, are coming from imagers who regularly shoot with Filters.

 

I am shooting from Bortle 1 without filters. none at all.

 

surely the filters would increase the sub exposure time requirement, as they can cut out 97% of the incoming light.

 

am I missing something ?

In my first reply to your question at the beginning of this topic I said: 

 

"My thoughts are that you should do some evaluations on the single subs and see how many pixels you are blowing out.  Actually do the experiment.  I would be surprised if you are blowing out much at all.  I routinely shoot 300s subs without filters with a 533 under class 1 skies."  

 

I am not an expert and am not an engineer at all.  But there is probably no one (or maybe a tiny handful) in this entire forum that shoots from a backyard with as similar sky conditions as you and I.  And with my set up and skies I bang out a LOT of imaging time so while I am very much a beginner, I am a beginner with a lot of hours...  Just saying.  

[bobzeq25]

This discussion is no longer at all sound.  In fact it's absurd.  What one person shoots for subexposure most often has NO bearing on what someone else should do.  Subexposure is ALWAYS a site and equipment specific thing.  You don't use what someone else uses, that's silly.  You use what YOUR data from a test frame indicates is correct.  There is no other way to do it properly.

 

Spaceman56s data shows that 120" is good for him.  There is essentially nothing to be gained by changing it.

 

There is EVERYTHING to be gained by shooting more of those 120" subs.

 

The above two sentences are why I keep saying that you don't get more detail with longer subs, you get it with total imaging time.  That's certainly true for Spaceman 56 and 120".   FOR HIM 50X120" and 20X300" would have just about the same amount of dim detail, because he would be collecting the same number of photons.  The significant difference would be that 20X300" would saturate a lot more pixels, killing star color and possibly clipping (blurring) highlights.  20X300" would _clearly_ be inferior to 50X120".  100X120" would _clearly_ be better than 50X120".

 

And it matters not one little bit that someone else gets good results with 300".  NOT AT ALL.  Choosing subexposure based on what someone else uses with different equipment on a different target in different skies is a terrible idea.  It would justifiably be called incompetent.

 

Getting this right is simple.  You shoot a test frame with your equipment on a specific target in your skies.  There are a variety of ways to analyze that test frame.  I like going for a subexposure that saturates a few hundred pixels.  That works, and is very simple. 

 

There are other ways to do the analysis.  I used to look at sky background and read noise, but that doesn't work very well in very dark skies like Spaceman 56's Bortle 1.  If he exposed long enough to swamp read noise with the _very_ small amount of sky noise he has, that would be major overexposure.  Star color would be really damaged.  Highlights would likely be clipped (blurred).

Edited by bobzeq25, 24 September 2023 - 12:31 AM.

[Rasfahan]

This discussion is no longer at all sound.  In fact it's absurd.  What one person shoots for subexposure most often has NO bearing on what someone else should do.  Subexposure is ALWAYS a site and equipment specific thing.  You don't use what someone else uses, that's silly.  You use what YOUR data from a test frame indicates is correct.  There is no other way to do it properly.

 

Spaceman56s data shows that 120" is good for him.  There is essentially nothing to be gained by changing it.

 

There is EVERYTHING to be gained by shooting more of those 120" subs.

 

The above two sentences are why I keep saying that you don't get more detail with longer subs, you get it with total imaging time.  That's certainly true for Spaceman 56 and 120".   FOR HIM 50X120" and 20X300" would have just about the same amount of dim detail, because he would be collecting the same number of photons.  The significant difference would be that 20X300" would saturate a lot more pixels, killing star color and possibly clipping (blurring) highlights.  20X300" would _clearly_ be inferior to 50X120".  100X120" would _clearly_ be better than 50X120".

 

And it matters not one little bit that someone else gets good results with 300".  NOT AT ALL.  Choosing subexposure based on what someone else uses with different equipment on a different target in different skies is a terrible idea.  It would justifiably be called incompetent.

 

Getting this right is simple.  You shoot a test frame with your equipment on a specific target in your skies.  There are a variety of ways to analyze that test frame.  I like going for a subexposure that saturates a few hundred pixels.  That works, and is very simple. 

 

There are other ways to do the analysis.  I used to look at sky background and read noise, but that doesn't work very well in very dark skies like Spaceman 56's Bortle 1.  If he exposed long enough to swamp read noise with the _very_ small amount of sky noise he has, that would be major overexposure.  Star color would be really damaged.  Highlights would likely be clipped (blurred).

I showed above that the OP‘s subs at 120s do not properly swamp read noise, so going longer has benefits. As I don’t routinely blow out too many stars at 300s with comparable speed and far lighter skies (and higher aperture - which is important for point sources) I say he won‘t, either.
 

With dark skies available the OP will easily beat anything that is done from higher Bortle class even with the 120s subs. Doesn‘t mean there‘s no room for optimisation. 5 min is hardly a long exposure.

[Spaceman 56]

thanks for the great suggestions. some good stuff coming in here.  

 

I will continue to shoot at 120 seconds as a base line, as that seems safe and good.

and I will  shoot some 300 second subs as a test.  

 

that seems like the best thing to do next.  

 

I will upload a 300 second sub for evaluation shortly. thanks Spaceman

Edited by Spaceman 56, 25 September 2023 - 03:05 PM.

[Spaceman 56]

I showed above that the OP‘s subs at 120s do not properly swamp read noise, so going longer has benefits.

 

thanks Rasfahan.

 

I will try shooting some 300 second subs and upload one soon.

 

It would be a good test. 

[Spaceman 56]

 What one person shoots for subexposure most often has NO bearing on what someone else should do. 

 

Subexposure is ALWAYS a site and equipment specific thing.  You don't use what someone else uses, that's silly.  You use what YOUR data from a test frame indicates is correct.  There is no other way to do it properly.

 

Spaceman56s data shows that 120" is good for him.  There is essentially nothing to be gained by changing it.

There is EVERYTHING to be gained by shooting more of those 120" subs.

 

And it matters not one little bit that someone else gets good results with 300".  NOT AT ALL.  

Choosing subexposure based on what someone else uses with different equipment on a different target in different skies is a terrible idea.

 

Fair comments and well said. 

 

I will stick with 120 seconds as my baseline, as that works and is good.

 

I will experiment with some longer subs and upload so expert people can evaluate them.  

 

thanks Spaceman

[dhferguson]

Cheers,

 

Given similar OTAs, same model camera, and similar light pollution, actually, best camera settings will be repeatable. Denying this is magical thinking. Recall Lord Kelvin's admonition: “When you can measure what you are speaking about, and express it in numbers, you know something about it, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind; it may be the beginning of knowledge, but you have scarely, in your thoughts advanced to the stage of science.”

 

bobzeq25 has made many, many posts, and I salute him for his energy and usually valid advice. However, I also disagree with him about the desirability of beginners--evidently including the technically savvy ones--always learning AP with a small-aperture frac instead of something with a larger aperture such as an 8" SCT. No, the AP hobby is not difficult if you are accustomed to making scientific or engineering measurements. In fact, I am stunned at how much the OEMs have done to simplify our equipment and spread the hobby. Compare what we have to work with now as opposed to, say, in 1975.

 

Happy observing always,

 

Don