1 year on, first proper photos!

[Andyk93]

Just finished my re-process, I'm unsure as to if it's better because I am missing parts of the galaxy it seems, I think more subs are needed to get the outer edge fainter stuff. 

[wimvb]

7 hours ago, Astrosurf said:

Lovely images! Andromeda doesn't have much in the way of colour so I would just concentrate on detail.

Alexxx

Have you seen Paddy's M31 widelfield? Plenty of colour in that image.

17 minutes ago, Andyk93 said:

Just finished my re-process, I'm unsure as to if it's better because I am missing parts of the galaxy it seems, I think more subs are needed to get the outer edge fainter stuff. 

Great image. You managed to get detail quite a bit into the core. The dark areas in the bottom and to the left suggest that you placed some samples over stars during DBE. To get DBE right you need to do two things:

1. examine each sample individually and make sure they don't cover stars or target

2. don't discard the background model. Examine it, and adjust the samples if the background model isn't smooth.

To get to the faint outer dust of this target, you mainly need longer exposures to increase the signal, and lots of them to reduce the noise.

[Andyk93]

Thanks @wimvb I have done some reprocessing. I think I have improved Orion massively. I'm unsure on andromeda but at 10 minute exposures and only doing 8 I'm thinking I need loads more! DBE's extracted background looks flat with darker corners when it's finished. Please ignore the stars in Orion, I hadn't noticed the scope was out slightly with collimation.

 

[wimvb]

Great Orion. You managed to bring out the really faint dust between M42 and the running man. Stars are under control and the core of M42 is resolved. Superb image.

I think you can safely darken the background a bit. This will give the image a little more punch.

M27 is in a rich star field. I think that this image has the same problem as your previous Andromeda (M31); missplacement of DBE samples causing dark patches in certain areas. Sample placement is more difficult in this image, due to the sheer number of stars. If you examine each sample individually, you can decrease the size of the ones that touch stars. Later on, before stretching, you can decrease the intensity of stars by using Morphological Transformation with a star mask. If you choose the mask such that smaller stars will be dimmed, but larger stars not, this will give a more natural effect, and bring out the larger stars better.

 

Good luck

[Andyk93]

Thanks @wimvb, really appreciate it. I'm going to give it another go tonight, DBE seems to be extremely critical of where you place those samples because I am checking each one individually.

I am wondering what is going wrong with my m31. every time I try to process it I am making an awful job of it. if anybody has some time to process it to see if its me or the data (or both) that would be greatly appreciated. I know its not going to be award winning but it looks extremely fake when I process it. thanks!

This is 3hr30mins of 10 minute subs over 2 nights stacked and fresh out of DSS.

andromeda 3hr30.TIF

[wimvb]

I downloaded the andromeda tif, and had a look at it. Unfortunately, I don't have the time at the moment to process it properly. It seems to me though, that you pushed the image too hard, and start to stretch and saturate noise. The colour in the original is very faint, and saturating it as much as you did, you inevitably also saturate the background.

The crucial point with processing in PI is the ability to create masks. For your image, standard luminance (lightness) masks will help a lot. Using a lum mask you can selectively increase the colour of the galaxy and the stars, while keeping the background under control.

It is possible to push the data further than I did, but it will take some tweaking and correcting stars and the galaxy core, to get proper results.

This is definitely an image you'd return to, as your processing skills increase.

As for background correction. I applied DBE with a sample size of 11 pixels and 10 pixels per row. I removed samples that were over the galaxy, and moved samples that covered stars. With the settings that I used, the corners weren't covered, so I placed a few extra samples here (but was a bit careless with the bottom right corner). Since the main gradient was due to vignetting, I used division as the correction method.

Since I kept the samples away from the galaxy, not all vignetting was removed. It is possible to destroy the faint "halo" surrounding M31, by placing samples closer to it. At some point you will start to destroy valid data.

There are severaly other issues with my processing of this image: stars are donut shaped, the galaxy core is starting to turn pink, the background level should be adjusted. And of course the lower right corner still shows vignetting.

Proper processing of this image would include going back and forth between increasing contrast and battling chromatic noise, with appropriate masks applied.

I would also pay closer attention to the stars and maybe reduce them  using Deconvolution and MorphologicalTransformation.

Cheers,

[Andyk93]

I have now managed to get my guiding good enough for 30min subs. I did 2 targets last night and processed them today, they were both 6x10min each with 20 bias and 20 flats processed in PI, I seem to struggle with noise but I kind of expect that with a Nikon d3300. These would definitely benefit from more data but I'm pretty sure with better processing they would look better. 

Bubble nebula

NGC7000 North American Nebula

All advice welcome

 

[tomato]

Hello Andy,

I cannot offer much advice other than to keep going as your images are really great, your raw data looks to be of very good quality.

Can I ask if you are imaging from a reasonably dark site?

[wimvb]

Nice images, both. The Cygnus wall (NA nebula) seems upside down. I think it would also benefit from SCNR to get rid of some of the green.

I like the stars in these images.

[Andyk93]

4 minutes ago, tomato said:

Hello Andy,

I cannot offer much advice other than to keep going as your images are really great, your raw data looks to be of very good quality.

Can I ask if you are imaging from a reasonably dark site?

Thanks very much, I do all my imaging on the outskirts of a small town. I can when dark adapted see mag 5 stars so I don't think I have too bad light pollution 

[Andyk93]

4 minutes ago, wimvb said:

Nice images, both. The Cygnus wall (NA nebula) seems upside down. I think it would also benefit from SCNR to get rid of some of the green.

I like the stars in these images.

It does seem a different orientation to others but that's how it came out the camera so I didn't think to change it. I did SCNR once so I'll process it again but with 2 runs of it. I do struggle to create masks for it because it's such an extended nebula. Thanks for the comments.

[wimvb]

I generally use SCNR whenever I find that the image shows a greenish colour cast. This can be several times during processing.

Masks are difficult to get right, but crucial to getting anywhere with PixInsight.

 

Good luck, and Merry Christmas,

[Andyk93]

Tried plate solving in sgpro last night, it worked for an hour to the meridian but after it flipped i didn't center back on it so as a result I lost 2 hours of 5min subs. I now have a modified canon 1000d but I'm getting banding that isn't being taken out by PI's banding script. 

Here's 10 x 30 seconds. I'm struggling with noise. 

[wimvb]

For only 10 x 30 secs, that's quite a remarkable result. The best noise reduction is always: more data. To get more detail in the background, you need longer exposures. I hope you get the plate solving working.

Cheers,

[Andyk93]

Here is 3 x 600 seconds of the eastern veil. I did have a lot more data but had to delete a lot of it due to a weird pattern showing up after stacking.

I like to add stuff on here as it shows my progress.