Different Length Exposures C6/Hyperstar/Ultrastar/Starlight Live

[Jim Smith]

Back in this thread...

...Martin requested screenshots of different exposures of the same object. So I took these last night of the Double Cluster in Perseus. There is no stacking. I only changed the exposure time from 1 to 2 to 4 to 8 seconds for each one. I'm trying to decide which is the most pleasing view. The 4 second one perhaps?

[Martin Meredith]

Thanks for doing this. It is fascinating to flip through them lined up (clicking on any image and using the arrow key does the trick), and watching the star counts increase.

I guess my favourite is the 16s one that you don't show ;-) But seriously, given that you are still building up stars, it is hard to select the 4s one knowing there is more available...

It would be interesting to see the result of stacking equivalent total durations but using different sub lengths. I find that colour of the brighter stars saturates quickly, and also that in longer exposures tracking/seeing can produce less than pinpoint stars. Against this we have to set increased read noise for shorter subs if the same total exposure duration is used. At some point there must be an optimum.

I tend to use 5s subs for the brighter clusters, and 10-15s for the fainter ones, but even in the 5s case I see some burnout  for clusters like M39 that I was looking at last night. I continue to experiment and will try to do this test myself if the weather holds out for the next few days.

Martin

 

 

[todd8137]

the  bottom one is one of the best i have seen on the double cluster 

pat

[Dom543]

Jim,

You are getting very nice star colors !

Try to use the arsinh option. That helps to prevent, or at least delay, saturation and allow longer exposures. Fully saturated pixels are white, no matter what. That kills the nice subtle yellowish hues. When you switch over to arsinh, you may see an almost entirely white screen and no histogram. The contrast slider needs to be pushed down to very close to zero for a nicely shaped histogram to reappear. Then the black level slider needs to be placed just below the histogram to make the background dark.

One can also try the x^0.25 option. But in my experience this results in more washed out stars that look as if they were out of focus.

Clear Skies!  --Dom

[Paul81]

8s exposure for me too :-)

Also Dom raises a good point - the non linear modes are great way to combat saturation when scaling the 16-bit camera data to the 8-bit data for the display. They are also great for pulling out detail in galaxies and nebula, I very rarely use the linear mode.

[RobertI]

Interesting comparison Jim, thanks for posting. I reckon 2 secs shows the best star colours but 8 secs obviously shows more stars. Its also interesting to see the slight rotation of the frame between each image.

[Jim Smith]

1 hour ago, Paul81 said:

Also Dom raises a good point - the non linear modes are great way to combat saturation

I did briefly flirt with non linear modes...I will try again next time out.

[Martin Meredith]

Interesting point Dom makes about the x^0.25 versus arcsinh modes. I have never checked the difference they make to star colours. Something else to add to the list...

I wonder how APers manage to maintain good star colour in long exposures. Post-processing tricks, or less sensitive sensors, or something else?

Martin 

[DoctorD]

Hi Jim

I made an animated GIF out of your images - hope you don't mind:

 

Shows very clearly the effects of increased exposure.

CS

Paul

[Jim Smith]

1 hour ago, DoctorD said:

I made an animated GIF out of your images - hope you don't mind:

Not at all! The animation shows the differences really well. Jim

[Dom543]

Jim,

If you have saved the FIT files, you can reload them with the  "<path>\StarlightLive.exe - args -load-image-rggb <path>\<filename>.fit"  command line. And try out the non-linear options on them during daytime. It takes a bit of experimenting to get used to the non-linear settings. It's better done during daytime than wasting precious time under the night sky.

Martin,

I believe that in AP, people essentially take and process separate images of nebulae and of the surrounding stars. Then they merge the two layers in the last step. I have seen many images that they call Ha-RGB or NB-RGB. The RGB is the get white, or naturally colored, stars.

I am also adding some times very short extra frames to my NB multi-spectral captures to make stars white. E.g. this is how I got a white Sagittarius Star Cloud in front of the red H-alpha nebulosity. Since stars are so much brighter, a 5 second frame is often sufficient to straighten out their colors. And the 5 seconds do not affect at all the nebula, that may have been captures with 60sec exposures.

Clear Skies!  --Dom

 

Edited October 21, 2016 by Dom543

[Martin Meredith]

Thanks Dom. I'd like to find a decent text on star colour in AP. What I have noticed is that the colours presented on images (e.g. via a google image search for a given cluster) vary wildly e.g. the same stars coming out as blue or yellow...

BTW I did a few experiments last night with nonlinear vs linear modes on star clusters and will post results in a separate thread once I've taken a close look and dug out some reference BV colour data for the clusters. This may take a while.

Martin