Sharpless 2-199

[Anweniel]

First go at this target and still experimenting with imaging hydrogen alpha on an unmodded Canon 1000D

Lights 9 x 10 mins

Darks 23 x

Flats 51x

Bias 31x

@ISO 800

WO Megrez 90, Celestron CG5, Canon 1000D, Baader 7nm Ha

From last night betwixt ~22.30 and 01.00

Stacked in DSS.

Still not sure how to approach stacking and processing of Ha in this way but making some small progress with time, any tips are greatly welcomed!

[davew]

Hi mate,

I'm relying on memory here so forgive me. I think the Canon filter only allows around 15% or may be 20% or so of the Ha line. If that's the case then your 10 minutes of exposure will equal about a minute and a half. Then if your not using a reducer you'll be at almost f7 further reducing the signal.

I think you'll either have to give longer exposures or higher ISO. Neither will keep the noise down but you may see more signal ?

DSLR imagers usually go for the modded cameras for Ha although one or two have managed a bit.

Nice to see you're using a decent amount of darks and flats.

Dave.

[Anweniel]

Hi Dave,

Thanks, and yes you are right about my setup not allowing the best of the photons but I can only work with what I have. Modding the camera is not possible as it belongs to my partner. I've had a play imaging the heart nebula with various exposure lengths and ISO settings to get used to what may be best, a few limiting factors though including the temperature of the sensor during the spate of warm weather we had as I tried these settings out and also the exposures at 30 minutes were a bit too noisy at the ISO's 800 and 1600.

I think we discussed the Atik and Moravian KAF8300 cameras when you did a talk for us at Walsall astro a few weeks back, so I am just trying to get what I can with what I have until I can splash out on a ccd

Thanks for the advice, I was thinking of maybe dropping ISO and increasing exposure length to get a better balance of signal to noise??

[Anweniel]

Hopefully this demonstrates what I was experimenting with, to the right is 9 x 10 mins @ ISO 800 and to the left 3 x 30 mins @ ISO 800.

Again temperature would play a huge part as the sensor was tipping 30 degrees when I was doing these initial tests so the noise problem likely accentuated but I also appreciate its probably not quite a fair fight as one side has a stack of 9 and one of 3 weighting the stacking in the benefit of 10 minute subs (which is what I have stuck to for now) but I thought of it in terms of imaging time - so in an hour and half I could attain this for example.

Hope this makes sense!

[davew]

Actually the right hand image looks quite reasonable. One trick a friend tried, in this area as it happens, was to do the Ha like you and strip out the red channel. Then do a straight RGB and blend the Ha with the red channel of the RGB. It worked quite well but you do need to be careful about the noise.

CCDs are very expensive when you buy all the stuff that goes with them. If a second hand Atik 383l comes up it may be worth a look. You'll only need 1.25mm filters too.

As the night time temperatures have recently dropped then it may be a good idea just to bang away at it and get more subs. As you've indicated, experiment and see,

Dave.

Just thought..... Those images demonstrate to some degree the difference between long and short subs of equal over all imaging time in a DSLR. It was recently discussed on another thread.

[Shibby]

Very interesting comparison. It does seem to reveal that 30mins is pushing it too far with a DSLR in the summer. (Bearing in mind that your SNR is affected by being unmodded) Subscribers to this recent thread would find this interesting, too.

As for processing, make sure you stack in "super pixel mode", then once you import into Photoshop, immediately delete the G+B channels. 2x Drizzle stack will also help reduce noise and increase the resolution that you have lost out on due to the bayer matrix.

[Anweniel]

Thanks Lewis, ill try 2 x drizzle although usually my pc cant cope!

[Shibby]

It's not too bad with super pixel mode because you are back to the original resolution

[Anweniel]

Used the 2x drizzle and super pixel mode as you suggested Lewis.

Wasn't certain about how to go about removing colour channels despite a quick google search! Hopefully I did a sufficient job.

1) I aligned the RGB channels in DSS with a 20% colour saturation and saved and applied the settings to a 32 bit TIFF file

2) I then cropped the dark frame that DSS adds and reset the levels histogram so the black point almost touched the left side of the gradient.

3) Next i dragged the green and blue channels and threw them all the way to the right

4) Then converted to grey scale, flattened - rough gradient removal and quick stretch in curves.

Not 100% sure its a massive improvement but it feels like its going somewhere better??