M33 Triangulum Galaxy Post

Below  my new image of M33. Much longer Integration time of nearly 4 hours brings out more detail. Still much to learn about Pixinsight. For noise reduction found the EZ_Noise script which is part of the  EZ Processing Suite Plug in for Pixinsight works really well on Linear images. So much so that this time I didn't need more noise reduction after the stretch.


M33 Triangulum Galaxy

Max

Many Thanks for the tip, Andrew. After last nights excellent meeting I downloaded a trial of PixInsight. Below is my original data but now processed in PixInsight including PhotometricColourCallibration. Colour balance is still a bit off, but I can now see more detail I think. The image has a lot of red noise which I dont know is due to clouds when I took the frames or noise from the DSLR.  

Nigel, 

Great improvement, interesting to see how Andrews suggestion helps.
Most guides suggest a core process that starts off with DynamicCrop to get rid of edge noise caused by stacking errors etc.  Then BacgroundModeliization (ABE or DB)  that will get rid of most gradients before doing PCC, that could well get rid of your "Red Noise"

I notice you say you are new to Pixinsight, it can be daunting to learn.  The most commonly quoted Guide is "Inside Pixinsight by Keller.  Its really a reference book and unhelpfully, the best place to start it is Chapter 21, worked examples. This contains the following helpful process chart for OSC cameras (including DSLRs)
Max

Nigel,

I'm very impressed that you only started the PixInsight trial after Monday's meeting, and are already producing good results with it. Even with the benefit of Colin's excellent week long training course, it was months before I felt moderately confident and competent about what I was doing, and I still feel I'm fairly low down on the learning curve. But getting results I quite like.

Max has mentioned the EZ Denoise script, which is an add-on to PixInsight. I've had very good experiences using it. It seems to do quite magical things with my DSLR data.  I use it for linear noise reduction, after background extraction and colour correction (which nearly always is using PCC). I just press the button and let it do what it thinks best. It produces much better results than I have managed to achieve with any of the other noise reduction methods. Which might be a reflection of my skill level with the other methods of course!

Keep up the good work!

Tim

Nigel
why not share a Dropbox link to your original stacked etc fts file and see what the group can do with it?
Also perhaps post image in your account with acquisition details.
Max

Went back to the start. reviewed the individual subs and discarded the marginal & bad frames. Re-stacked in DSS and generated a 32 bit image which I then opened in Pix Insight. And did the following workflow:

• Dynamic Crop
• ABE
• Background Neutralisation
• PCC
• Histogram Transformation
• EZ Decon
• EZ DeNoise
• SCNR
• Saturation

There's still a lot of red noise in the image which becomes more evident as you do a curves stretch. 

Original 32bit Stacked image out of DSS is linked - Would love to hear from anyone able to remove the red noise 

Hi Nigel

I have had a go at processing your image from the linked original data you provided.



This was processed in PixInsight then Photoshop.

The PixInsight work flow was: DynamicBackgroundExtract, BackgroundNeuralization, PhotometricColorCalibration, EZ Denoise, MultiscaleLinearTransform (further denoise), MaskedStretch, HDRMultiscaleTransform (on masked galaxy), LocalHistogramEqualization (on masked galaxy), Curves (RGB), Curves (saturation on masked galaxy), MorphologicalTransformation (reduce stars), and SCNR (remove green).

The MaskedStretch actually appeared to make the most difference for some reason. The blues were noticeably bluer and the red noise areas noticeably diminished after this step.

In Photoshop I used the CameraRaw filter to make a range of small tweaks but nothing drastic.

Your original data is not bad at all. Though I notice that there is quite a lot of spherical distortion as you move out towards the corners. 

Cheers
Andrew

That’s a pretty large improvement on where the image started. Great work.
Max

Hi,
took me a while to process, this is my M33 final result. Processed in Pixinsight, approx 180 minutes of data. Captured on an APM 107/700 with 0.75 FR on a NEQ6 Pro mount and auto guided. Camera is a CDS Pro II cooled DSLR (basically a heavily modded Canon 6700D. Capture software is EKOS.

IO really enjoyed this challenge, I am sure my processing skills can improve further but I tried some things in PI that I haven't tried before and have learnt from the process. This group is just what I need to keep motivated.

Jonathan