Subframe Selector questions

[SAW]

Hi,

Just purchased the PixInsight after using the trial version. I am using a dslr (Canon 600D) but I would like to learn more about stacking manually rather than using the BPP Script. I have loaded 145 sixty second lights into the subframe selector.

Under system parameters I have entered 2.111 for subframe scale, is this correct for my camera ?
Camera gain as 1, all images were ISO800. Is this correct ?
Under scale unit do I select pixels or arc seconds ? I want to measure my results using SNR Weight.

Once I click measure I get my graph/plot displayed, what do the dotted horizontal lines mean ?

If I am using SNR Weight to select images how do I know what scores to keep and which to drop for my final selection ?

Thanks.

[aworonow]

It is my understanding (maybe wrong, but I don't think so) that you get the same relative ranking of images in Pixel and Arc Seconds mode. In the former you need not enter any of the camera info. In the latter you enter the camera info, but get the same relative ranking whether you enter the correct info or some arbitrary info. You can give it a try: us the info you have and select Arc Seconds, then select Pixel. Other than the numerical values, would your selection of subframes be the same in either case?

Alex

[SAW]

Just re-run the script with just 10 lights, first time using pixels and the default subframe scale (1.000) and default camera gain (1.000)  and second time using arc seconds with the same lights and default settings again. Compared the plots for both and they were identical when I viewed SNRWeight.

Is that correct ?

[John_Gill]

Hi,

With the Sub-Frame Selector the higher the SNR the better, and the lower the Eccentricity (star roundness) and FWHM the better.  Remember to Blink your images as clouds and aircraft/satellite trails can alter the SNR.   I tend to favor SNR, but if I have two or more images of about the same SNR, then the lowest Eccentricity wins.  I don't think the other settings matter if the Lights were all taken at the same exposure and ISO.  The sub-frame scale for your camera is 1.43 arcseconds/pixel and the camera gain is as follows :
ISO        Gain    
--------------------
100        8.79
200        4.40   
400        2.19   
800       1.12   
1600     0.58
3200     0.29   
6400     0.15   
12800   0.07

Look up
John

[aworonow]

SAW
Yes, same result is what I expected. You lose the stuff like true resolution if you don't use the camera parameters, and instead have it in pixels (e.g. FWHM), but big deal as far as subframe selection goes.

Alex

[SAW]

Hi,

With the Sub-Frame Selector the higher the SNR the better, and the lower the Eccentricity (star roundness) and FWHM the better.  Remember to Blink your images as clouds and aircraft/satellite trails can alter the SNR.   I tend to favor SNR, but if I have two or more images of about the same SNR, then the lowest Eccentricity wins.  I don't think the other settings matter if the Lights were all taken at the same exposure and ISO.  The sub-frame scale for your camera is 1.43 arcseconds/pixel and the camera gain is as follows :
ISO        Gain    
--------------------
100        8.79
200        4.40   
400        2.19   
800       1.12   
1600     0.58
3200     0.29   
6400     0.15   
12800   0.07

Look up
John

Thank you. Yes I have already used Blink and scrapped any images with satellites, cloud or didn't look good, before using the subframe selector.

What's the difference between using pixels and arc seconds for the scale units ?

[John_Gill]

Hi,

The pixel is an image measurement, whereas an arcsecond is a sky measurement.  For example you might find that the graph shows a difference of 3 pixels, which is not much, but translated to arcseconds will show a more pronounced difference in the graph.  It really is just the scale.  The same is true when doing autoguiding, I use PHD, and in the pixel scale I am getting is 0.5 pixels up or down, but when looking at arcseconds the graph has more variation so I can adjust or fine-tune the graph to get a flatter line.  ( I hope this makes sense).

Look up
John

[rodmichael]

Does the subframe selector have problems with NB sub-frames. I have not had difficulty with the script previously but this afternoon it seems to be having difficulty with some NB subframes of the Veil Nebula. Both Ha and SII sub-frames seem either to freeze the script or to produce unusable (virtually no) data> OIII sub-frames have produces data (SNR, ECC, etc), but it takes a very long time. I basically have to leave the computer for 15 to 30 minutes for the script to finish with 12 sub-frames. I can usually finish script processing of 40 frames in a few minutes.