Number of subs to get something worthwile with a camera

[frugal]

On Tuesday when it was lovely and clear I had my first go at astrophotography. As I do not have a motorised mount for the telescope I tried some wide field shots with the camera (Canon 60D with a zoom lens set to 18mm f3.6). However the results were not great, I seemed to get less information in the pictures than I could see with the naked eye.

I was trying to take a shot of the milky way (which I could make out with the unaided eye). I took 20 subs at 20 seconds iso800 (and 5 darks), but when I stacked them all I got was the brightest stars. I can only take about 20 seconds per sub before star trails start to appear.

So my question is: with a camera like a 60D and a 18mm f3.6 lens, how many subs do I need to be taking to pull up some of the milky way information behind the bright stars?

Sent from my GT-N7000 using Tapatalk

[foundaplanet]

After stacking them did you run the image through another program like photoshop or GIMP..

[jgs001]

You can get some detail using 10x30 second exposures at ISO800 at 18mm.. however, the stacking process results will need work to pull out the detail. Have a look at http://stargazerslounge.com/topic/36308-basic-widefield-with-a-camera-and-tripod/

[tdmodels]

Could you post the resulting stacked image into something like Dropbox, for us to have a play and see if we can point you in the right direction?

Check this youtube video out...  http://youtu.be/q91_jn4lPYM

[frugal]

After stacking them did you run the image through another program like photoshop or GIMP..

I opened it in Photoshop and tried to bring up the background with curves, but by the time I was able to bring anything up there was a lot of noise but no more information. Hopefully someone with better editing skills can do something more with it.

Sent from my GT-N7000 using Tapatalk

[frugal]

You can get some detail using 10x30 second exposures at ISO800 at 18mm.. however, the stacking process results will need work to pull out the detail. Have a look at http://stargazerslounge.com/topic/36308-basic-widefield-with-a-camera-and-tripod/

It was your excellent article that inspired me to have a go in the first place. I saw your gorgeous photos of the milky way and wanted to have a go.

At the moment mine are less "light and fluffy milky way" and more "bitter value brand dark chocolate"

Sent from my GT-N7000 using Tapatalk

[frugal]

Could you post the resulting stacked image into something like Dropbox, for us to have a play and see if we can point you in the right direction?

I will upload it to dropbox this evening when I get home.

Sent from my GT-N7000 using Tapatalk

[jgs001]

Glad you felt inspired to give it a go . There are other factors that will influence the results... sky conditions, a high level mist for instance... which will hide the fainter stars... And it can cause all sorts of odd results on long exposures

[frugal]

Okay, here we go. https://www.dropbox.com/sh/wmzqclf5051z1p8/3oLaRTND1v

There are three files in the folder:

1. MilkyWay.tif - The original attempt with a bit of extra touching up that has started to show the milky way. I had saved the file from DSS without applying the adjustments, so all of the adjustments were done in Photoshop
2. MilkyWay2.tif - This is the file as saved from DSS with the adjustments saved after playing with them as described in the You Tube video linked to by Tim.
3. MilkyWay3.tif - This is MilkyWay2.tif after I have played with it in Photoshop for a while.

I am sure that there is a lot more that can be pulled up from the images, but I am a lot happier now that I can actually see the milky way

[frugal]

Check this youtube video out...  http://youtu.be/q91_jn4lPYM

Thanks for the link, I was not doing the preprocessing stage in DSS, I had thought that it should all be one in Photoshop. Doing the first stage ot stretching the histogram in DSS made everything pop out.

[tdmodels]

Hi Frugal,  you got some good results there.  My processing of your data couldn't bring more up, but then I am also a beginner with processing too  

Lots more subs should help, and it may be worth having a read up on flats too as there is a lot of vignetting in the corners.  They are easier to master than most people think.  I just click away at a computer screen with a blank notepad window open, set the camera to auto and don't adjust the f ratio, zoom or focussing from when you were imaging.  Snap a handful of those and add them into the DSS pot - will help a lot.

My only other advise would be to image Cygnus - So much more going on in there and the milky way regions are a lot brighter

In the image below, I was using a 50mm lens, and 16 x 5min subs at ISO800.  I took darks, but no flats or bias.  I did have a CLS filter installed and was tracking on my Astrotrack.

Here's wishing you clear skies....






 

[frugal]

Lots more subs should help, and it may be worth having a read up on flats too as there is a lot of vignetting in the corners. 

Tim, 

I am not sure how much of that was vignetting in the corners, and how much was the star field rotating and DSS only having a few frames to stack at the edges. I think if I set DSS to the crop mode of stacking a lot of that will go away.

I also need to figure out how to do gradients in Photoshop as I think that will help me get rid of some of the light pollution bleed in some of my other pictures (now that I can bring out some detail in them).

[jgs001]

Do you have the original stacked file saved with adjustments embedded only ? There's a fair amount of detail showing up, but there's also either a lot of LP or some issue in the editing in DSS. If you don't have a save file with adjustments embedded only, and not applied, could you stack it again, save the output with embedded and post that too please ?

[tdmodels]

Here is the MilyWay2 file stretched hard.  You can see the stacked edges, but at the top there is definitely vignetting.





 

[frugal]

Here is the MilyWay2 file stretched hard.  You can see the stacked edges, but at the top there is definitely vignetting.

Can I take the flats now, or does the camera temperature need to be the same as when I took the originals? I have not changed the focus point, or zoom since I took the lights, so I should be able to take the flats with the same settings.

[frugal]

Do you have the original stacked file saved with adjustments embedded only ? There's a fair amount of detail showing up, but there's also either a lot of LP or some issue in the editing in DSS. If you don't have a save file with adjustments embedded only, and not applied, could you stack it again, save the output with embedded and post that too please ?

I am currently uploading MilkyWay4.tif to the same folder which is a new stack with the adjustments embedded. It should be up in a few minutes.

[jgs001]

Great, thanks... I won't really be able to do much till a lot later... but I'll have a play this evening (hopefully).

[tdmodels]

Can I take the flats now, or does the camera temperature need to be the same as when I took the originals? I have not changed the focus point, or zoom since I took the lights, so I should be able to take the flats with the same settings.

In which case, you should be fine to take them now.  I use a piece of white opaque Perspex over the lens and point to an evenly laminated part of the sky, others use a white t-shirt over the lens and as mentioned before, you could also use a white screen on a laptop or iPad.

[frugal]

I tried to take some flats last night, but it was difficult to get an image that showed the vignetting. I tried to shoot the laptop screen, but it looked a bit weird. I tried to shoot a white piece of paper under a daylight bulb, but there were still too many artifacts from the direction of the light source. So today I went outside to try the t-shirt approach, however I kept seeing artifacts from the weave of the t-shirt. Fortunately the day was overcast, so I was able to point it at various parts of the grey sky to get something half decent. There is still some vignetting, but a lot less than there was. Next time I think I will change the f-stop so that it is not right at the end of the scale.

It certainly looks a lot better. No great, but better. I had real problems getting the light pollution removed from the bottom right hand side of the image. My photoshop skills are lacking, so my attempts at removing the glow from the streetlights with a gradient were a bit of a failure In the end I downloaded trial versions of pretty much all of the astrophotography software that I could find. In the end I used Maxim DL to clean it up and remove the light pollution, and then photoshop to clean up the levels (only because I understand levels in Photoshop).

Thank you to everyone who has given me help and advice to create this image. I am really looking forward to the next clear night; I want to try again, but take a lot more subs, to try and bring out some more details.

[foundaplanet]

l had a bit of a go with your image from .tiff. It's okay, I cropped the black edge out first and didn't find it too bad really, you just need more subs..

[jgs001]

That is an improvement, although it looks a bit blue.. I guess that was the LP removal. Here's my effort at processing it

a pass of Gradient Exterminator (medium/medium) about 4 curves stretches, 3 curves contrast boosts and some deep space noise reduction from http://www.prodigitalsoftware.com/Astronomy_Tools_For_Full_Version.html (having just looked at the website again, there's a lot more tools in there I'll have to use more often )

Looks like you got Andromeda on the top edge too.

[frugal]

Cheers everyone. It is interesting to see how different people have brought out different things.

Am I correct in thinking that the signal to noise ratio works on the square law? So if I want to bring up twice the signal I am going to need to take 4 times as many subs?

Sent from my GT-N7000 using Tapatalk

[glappkaeft]

Cheers everyone. It is interesting to see how different people have brought out different things.

Am I correct in thinking that the signal to noise ratio works on the square law? So if I want to bring up twice the signal I am going to need to take 4 times as many subs?

Yeah, you got that right except that it should read "twice the signal to noise ratio" instead of "twice the signal" in the last sentence. The math behind it (light gathering is a Poisson distributed statistical process) is pretty complicated but the end result is pretty simple:

Noise = square root of the signal

thus

Signal to Noise ratio (SNR) = signal/noise = signal/sqrt(signal) = sqrt(signal)