RE-process M33 with PI

[steviemac500]

So I've just got PI and sat through Harry's tutorials tonight with my M33 data and I am really pleased with it. It's all a bit confusing but when I compare this to my previous image I think this beats it hands down. Any tips, especially for a newbie on PI would be appreciated.

[Dark Horizon]

Nice image, Stevei. I can't advise on PI as i've never used it, but it looks like you're doing something right. 

[wimvb]

I am looking at your image on my mobile phone, so I may be wrong, but it seems you have some horzontal gradients. You can apply dbe with some samples along the top, sides and bottom. Then use subtract as correction.

Hope this helps

[steviemac500]

17 hours ago, wimvb said:

I am looking at your image on my mobile phone, so I may be wrong, but it seems you have some horzontal gradients. You can apply dbe with some samples along the top, sides and bottom. Then use subtract as correction.

Hope this helps

Thanks Wim, it does help a lot. My eye doesn't always see these things until someone points it out and then usually I cant believe i didn't see it!!! I've not really got past following the steps in Harry's tutorials yet and haven't really got a workflow sorted yet but i appreciate any tips i can get along the way.

[wimvb]

For what it's worth, here's a rough description of my workflow. This is for my unmodded dslr, which is very noisy.

1. Calibration and stacking, using cosmetic correction for additional hot pixel removal

2. Cropping to remove sracking artefacts on integrated image

3. Dbe with either manual samples or automated sample placement, depends on target. I spend quite some time to inspect each sample. Also inspect background model and redo if necessary

4. Deconvolution to decrease star size. Added this step recently, since I found it difficult to get right.

5. Tgv denoise in L*a*b mode on just L

6. If there is chromatic noise, I use MMT on chroma only, and only on the background, using a luminance mask.

I have described my noise reduction method in detail in my blog (address is in my signature)

7. Morphological transform using a contour starmask and using morphological selection, decreases star size further

8. Masked stretch, very much default settings, but disable clipping

9. Histogram transformation, adjusting black point and increase contrast

10. Colour saturation with luminance mask, increase colour in the target

11. Inverse luminance mask and decrease colour saturation of background if necessary.

12. Tweaking contrast with various stretches

13. Croma noise reduction if necessary.

I do a lot of noise reduction, mainly because I have an old and very noisy camera. I also spend some time on taming stars, in the linear stage.

The steps that I started with are 1, 2, 3, 5, (8), 9, 10. The other steps were added along the way, as I learned more.

Hope this helps. 

[ollypenrice]

Well first off, when I looked at the image it didn't scream (or even whilsper) 'Pixinsight' so that's good. It's a programme which tempts people into excess but you resisted. It's nice and natural.

Yes, there's a gradient. As Wim says, DBE at the linear stage would sort this. I'd suggest very few markers and keep them well away from the galaxy or you'll get an overcorrection creating a dark outer region.

Some SCNR green helps nearly all my images, too.

You have an optical issue on the left, possibly tilt?

Olly

[steviemac500]

On 10/15/2016 at 10:42, ollypenrice said:

Well first off, when I looked at the image it didn't scream (or even whilsper) 'Pixinsight' so that's good. It's a programme which tempts people into excess but you resisted. It's nice and natural.

Yes, there's a gradient. As Wim says, DBE at the linear stage would sort this. I'd suggest very few markers and keep them well away from the galaxy or you'll get an overcorrection creating a dark outer region.

Some SCNR green helps nearly all my images, too.

You have an optical issue on the left, possibly tilt?

Olly

Thank you both. I appreciate the workflow list Wim, there's a lot on there i haven't looked at yet but ill get round to it. I've had a another play but to be honest i'm not sure this has sorted the gradients at all?

[wimvb]

Gradients can be tricky. A clean background is the foundation on which you build your image. As such you need to spend time on this part of processing. A nice feature of PI is that you can save an images complete processing history, if you save it as a project. This makes it easy to redo a process.

[gnomus]

To my eye, the gradient looks much improved.  Before I was seeing red at the bottom and green at the top.  It may be my eyes, or my iPad, but it looks as if there is still some brightening along the top and bottom edges.  How did you calibrate the images?  I ask because it looks a little like amp glow and you could probably eliminate it with darks, if you've not already done this.

Finally, you may want to think about boosting saturation a little.  You can do this in PI using the curves tool, clicking on the S button, and dragging the curve up a bit.  If you made a mask, you could mask out the background before doing thsi - no point increasing the saturation of any backgrdound colour noise.

Oh.  I forgot to say - nice image.

Steve