×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

Optimal Sub-Exposure Calculator help

  • Posts: 139
  • Thank you received: 16
Hi Joseph,

returning to the Bortle scale, I finally had the time to look at your reply in more detail. I shared your struggle in trying to find a definitive conversion. The scale you use is the one quoted by Bruce McEvoy. I also found different versions by Henk Spoelstra (who was the only one to introduce Bortle 4.5), Miguel Martinez-Ledesma et al, and a version of the McEvoy scale extended to Bortle 8 by Attila Danko. There was some difference between these scales, so I took both the mean and median amongst them and assigned a value rounded to the nearest 0.05, based on those. It's not very scientific but, as you said, it's subjective anyway.

The 4.5 is a bit odd, being only introduced by Henk Spoelstra -- but it's his nomogram which is the basis for scale in the Wikipedia page on the subject, and which was featured in the 2009 Dark Skies Awareness project. The Bruce McEvoy scale is the most widely reused, as far as I could see.

In my app, I leave the choice of 4.5 up to the user, which is why I sent the two versions.

But the reason I asked the question was to see if you had a different version.

Thanks,
Richard
2 months 1 week ago #98895

Please Log in or Create an account to join the conversation.

  • Posts: 42
  • Thank you received: 11
Hi Ronald,

Sorry for the slow response, I was travelling on a family matter.

As for measuring Sky Quality via a process with imaging equipment; that is something I'm interested in developing. I had read about the process some time ago, but what I read required human interaction in manual section of a background area of the image. (I instead chose to purchase a Unihedron meter during the development of the exposure calculator.) It would be an interesting challenge to develop a more "hands-off" process for determining a value for sky quality with imaging equipment, but at the moment I am working on a tool for mapping camera sensor read noise. I also found that on-line tools like light pollution map are fairly accurate, but they give an optimistic value (no moon, and no clouds reflecting artificial light). So using on-line tools without adjusting for these factors would tend to produce longer exposures than would be recommended with an accurate value for sky quality.

As for simply choosing a consistent exposure time (like 20 to 30 seconds); when I first proposed developing the calculator for KStars there was strong argument against it. The suggestion was that we should consistently use exposures of 3 to 4 minutes. But my experience suggests that a great deal depends upon both the sky quality and components in the optical train. Before I found the presentation by Dr Glover, I spent several weeks with a trial and error approach to find acceptable exposure times for my equipment at two different sites (my backyard with about SQ 19.4 and a nearby mount range SQ 21.5). After I found his presentation I created a spreadsheet for my gear at these sites. I was astonished by how closely the calculations matched the values I found from my many weeks of effort.
2 months 6 days ago #99053

Please Log in or Create an account to join the conversation.

  • Posts: 396
  • Thank you received: 17
Thank you for your response.

I have to admit that I am an impatient Astrophotographer. If the night is clear and the equipment is all working well, I will start with short exposures of 10 seconds (depending on the target), then 20, 30 60 120 and even 180 seconds. Usually my guiding starts to break down somewhere between 1 to 3 minutes, so I am limited there by my equipment. What I find is that for long exposures I can begin to 'see' the object (Nebula or galaxy) and so I start to take more of those exposures. But what I sometimes discover later is that the sky noise was to high or the stars are blown out and so I have a difficult time processing the data as I have to blend multiple exposures.

So optimizing the exposure time is important and knowing how many subs you will need to take to bring the object out the noise clearly. But, I am not the inclined to spend many multiple hours on a target. Usually what I can get in one evening is going to have to be good enough. I think more of us fall into that category than those who have permeant observatories and can spend multiple nights on a target.

I am trying to do the best I can in a limited time and I am not sure that fits very well with Dr. Glovers approach. But I believe his math. The capability of image processing software is in a constant state of flux. I have only just used the StarNet Star subtraction capability in Siril, so that I can stretch a starless image.

I guess what I am saying is that I understand it is complicated by all these factors. I am looking for something that guides me at the moment, with the object in view, that says this may be a way to proceed. A dogmatic approach that says that 300 ten second images on target will get you an image with a SNR that exceeds a certain value just does not seem to help. First, that is 5 hours of image collection and second if I cannot get those 5 hours have I just wasted my time.

I am not trying to take pictures for APOD, I am just trying to enjoy the hobby with results that I am proud of and amazed that I can produce an image of something that far away.

Sorry, for the digression. I appreciate what your exposure calculator is providing us with and I will have to spend more time with it.

I had a good night and collected a bunch of data. This is the result of simple processing of a half hour of 60 second exposures (I have more images at longer and shorter exposures but I had the most at this exposure)



I will spend some more time with the data to see if I can clean it up better. But I am struggling to decide where the most effort should be applied.

Thanks again
2 months 5 days ago #99060
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 396
  • Thank you received: 17
Sorry, I got off on a bit of a rant on the last post and I really did not mean to, I apologize.

When I look up my location on a Dark Skies Map it says that I am 21.77 sqm, looking vertical with no moon, etc. So I am fortunate that I am in a Bortle 3 location. I use your calculator for my ASI 533 camera and it gives me very reasonable numbers for exposure times (Its dark so I can go long) and very achievable exposure counts, only tens at long exposure. But my equipment does not guide well enough for a 300 second exposure. So how should I use your numbers? Should I calculate the total exposure time (exposure x number) and use that as a goal with my shorter exposures? I do not see anyway to put my own exposure time in and have it give me how many images I should capture. In my actual experience I can see a definite difference in background level between 120 sec and 180 sec exposure (this being brighter). So I am inclined to stay with the shorter exposure and just take more images.

Another, simple, question. What filter bandwidth should I use for an OCS camera? Is it the full bandwidth covered by all RGB or that of each individual color (that is what I would assume).

I need to watch your video again.

thanks,
2 months 4 days ago #99079

Please Log in or Create an account to join the conversation.

  • Posts: 42
  • Thank you received: 11
Hi Ronald, No need to apologize nothing in your post seemed wrong to me.

I agree that the calculator can produce exposure times that are beyond reasonable given limitations in equipment and conditions. With extreme inputs, in very dark sky , a fast optic, and narrow band filters the calculation might be so long that it would only be viable within the arctic circle in Winter, ;-).

To reduce exposure time, my first recommendation is to consider a reduction in gain on the camera. Raising the camera gain, will typically reduce the read-noise, and result in a lower exposure time, (but usually at the cost of losing some dynamic range). However, some cameras have an electronic switch that changes the read mode. In these cameras, you will notice a pronounced step in the exposure time graph. At the bottom of this step, read-noise is low, but dynamic-range would typically jump back up, (because dynamic range is effected by read-noise). So if you have camera that shows a step in the exposure time graph, you might want to push the gain up just slightly to the right of that step.

There is another way to force the calculation to go outside the range recommended by Dr Glover, His calculation uses an input called "Noise Increase %". This input controls the factor that is used to compute the balance between the noise from light pollution and the read noise in the camera (this factor is sometimes referred to as the "swamp factor"). Dr Glover recommended using 5% (or lowering to 2% if the exposure time is very short). But I chose to leave this input relatively unlimited so that users could get results for exposure times that meet their needs. You can think of the "noise increase %" as governing the relative amount of read-noise, so an increase in this input will lower the exposure time, (which results in a relative increase in read-noise in the image). Read noise is constant, so a shorter image will have higher amount of read-noise relative to noise from light pollution). So to reduce an exposure time you just need to raise the input value of the "noise increase %".

As you adjust the noise increase upward the exposure time will drop, but you will also notice that sub-exposure count will climb disproportionately, this is because shorter exposures are relatively more noisy, and so more stacking is required to achieve a similar final result.

As for the filter input for an OSC. With no filter, you would use a value of 300 nm. If you end up using a filter, then some analysis of the filter is required. For example, with my OSC on nebulae I use an Optolong L-enHance two-band filter, for this filter I input a value of 34 nm, (this value is arrived at by adding the widths of the two bands passed by this filter).
2 months 4 days ago #99108

Please Log in or Create an account to join the conversation.

  • Posts: 42
  • Thank you received: 11
Hi Ronald,

Here's a bit more detail with examples:

The ZWO docs for the 533MC show the dynamic range does jump back up after the electronic mode switching (around gain 100, this can be seen in the 1st picture). The DR just above gain 100 is very nearly as good as you would get at gain 0. But the reduced read-noise will result in a lower exposure time. (I do not know your focal ratio, so I used the 5.5 ratio of my refractor for these examples.) Setting the gain a bit above 100 should help to be sure that the exposure is beyond the point of the switch. (To be certain you might want to push gain up to 110).

So this gives a sub-exposure time of about 112 seconds. But also notice that the ratio of exposure time (112) to total noise (8.21) is about 13.6. To get to a time noise ratio of 80 in stacking requires 34 exposures, So a total integration time of 63.5 minutes.

But if you want to further lower the exposure time, the noise increase % is adjusted upward, I pushed it up to 9%, to get the sub-exposure down to about 60 seconds. But this means the sub-exposure time/noise ratio is lower (61 / 6.14 , about 9.9, so it's not as clean an image as the sub-exposure at 112 seconds). The stacking calculation says that 64 images of 61 seconds would be needed to reach the time/noise ratio of 80. So a total integration time of about 65 minutes.

Also keep in mind that in the stacking calculation, the value of 80 is sort of average; it should work well on bright target but if you are trying to get faint details you might need to push that value up. (I've had some feedback from folks that do a lot more integration, some do an equivalent time noise ratio well above 200!).
2 months 4 days ago #99116
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 396
  • Thank you received: 17
Thank you this is all very helpful. I do use a gain of 105 for the ASI533, to be just above the drop as you indicate (I could bump it to 110). I would not want to lower it.

On a good night of seeing I can guide for 120 seconds and achieve round stars, so those numbers are all very doable. Of course, the problem with the longer exposures is that you blow out the stars, so I also shoot shorter exposures. I am using Siril as my processing tool (for now) and it allows for star extraction and then recombination after stretch. I am trying to process with the longer exposures on the dim objects and recombining with stars from a shorter exposure, to see how that works.

I think this tool does give a very good starting point for setting your exposure time. I will have to take some data for my SBIG 8300MM camera to be able to use it with this tool, as I don't see it in the list. I will read back thru this thread to see what data you need for this camera.

It still seems to me that you have exposures, as you are taking them, that contain the information of the target levels and the sky noise levels of the background (as best as you could extract them from the data). And the current processing programs, like Siril (I don't use Pixinsight ) or even a photometry program like Fits Liberator give you access to those values (again if you can interpret them correctly). So that with this information you could make a more informed estimate of what exposure to use and then your calculations would tell you how many you need to take to achieve a certain SNR. Even something simple like knowing how many stars are saturated in an image would help, as you could say I would like less than 5% of the stars saturated (only the very brightest ones). That might be something you can get from the Fits Viewer in Kstars, I am not sure.

thanks again for your contribution.
2 months 4 days ago #99119

Please Log in or Create an account to join the conversation.

  • Posts: 42
  • Thank you received: 11
Hi Ronald,

I found specs on the SBIG 8300 on-line stating it has a read-noise of 9.3.

So I copied and modified the ATIK CCD camera data file for you. You should be able to unzip the attached file and move it into the cameradata folder under KStars. (either into /usr/share/kstars/cameradata or into ~/.local/share/kstars/cameradata).
2 months 3 days ago #99138
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 42
  • Thank you received: 11
Hi Ronald,

I found specs on the SBIG 8300 on-line stating it has a read-noise of 9.3.

So I copied and modified the ATIK CCD camera data file for you. You should be able to unzip the attached file and move it into the cameradata folder under KStars. (either into /usr/share/kstars/cameradata or into ~/.local/share/kstars/cameradata).
2 months 3 days ago #99139

Please Log in or Create an account to join the conversation.

  • Posts: 139
  • Thank you received: 16
Hi Joseph,

would you mind making the appropriate XML files for the four camera I uploaded and maybe put them on the Google drive?

I checked that my version of 3.6.9 beta does not included them

Thanks,
Richard
2 months 3 days ago #99142

Please Log in or Create an account to join the conversation.

  • Posts: 42
  • Thank you received: 11
Hi Richard,

Sorry for the delay, I had a family matter that required my attention and took me on a long trip over the past week. I had also had some enhancements to implement in my tool, and found a defect in my calculations to resolve, (it turned out to have a negligible effect on the result). The attached pdf shows a re-run of your data.

I've created exposure calculator files for your cameras and copied them to your folder on the Google drive. So you should be able to copy these to a KStars cameradata folder.

But for the Moravian C5A I've created two versions. Let me explain; the camera data files support 3 "modes" for sensors:


"FIXED" handles true CCD cameras, so there is essentially only one read-noise value in the file.

"ISO_DISCRETE" was intended to handle DSLR cameras where the selection of gain/iso is from a list of discreet values.

"NORMAL" handles cameras that have gain values that variable over a broad range: the camera data files include a table of values, but the exposure calculator will interpolate between these gain values to produce a read-noise value.

When I first built the calculator, I had created a file for a Moravian C3 in a "Normal" mode. But from what you have described in a prior communication, I now suspect that was not correct. From what you described, its seems that that the Moravian cameras are more like DSLRs in that they only allow a discrete value for gain selection.

So for your Moravian C5A, you will find a version with a file name that ends in "-D" (for discrete e.i. "DSLR" mode), and version that ends in "I" (for interpolated e.i. "Normal" mode). When running the calculator with the "-D" camera data you will find a dropdown that lists just the 0, and 2750 gain values, (But the label on the calculator will show "ISO"). when running the "-I" file the calculator will allow you to select values within that range of 0 to 2750, (but I as I said I think this is incorrect),

Let me know if my understanding on the Morivain cameras is correct.

Thanks
2 months 2 days ago #99147

Please Log in or Create an account to join the conversation.

  • Posts: 396
  • Thank you received: 17
Thank you for the file for the SBIG I will load it up next time I start up the equipment.
2 months 2 days ago #99150

Please Log in or Create an account to join the conversation.

Time to create page: 1.248 seconds