M51 Whirlpool Galaxy (15th Apr 2020, first attempt)

[rob_r]

Hi. Preserving is beginning to pay off a little. Even though I couldn't see anything through the eyepieces, trusting the GoTo on the EQ5 with an improved polar alignment, I got my first glimpse of a galaxy. I set up the polar alignment whilst it was still light with the small inconvenience of the polar reticle rotated on the SW EQ5 which I believe is quite normal, 12 isn't at the top. I may buy the Polar Scope Align app on my mobile to help further. I then one star aligned to Arcturus, keyed in M51 and had a look at various magnifications and nothing. Keyed back in Arcturus and the scope landed pretty much centre on the star. So I swapped out the EP for the camera and adjusted the focus. Test shot on Arcturus, 30 seconds, looked OK, slight trailing (again). Keyed in M51 and just thought just go for it. Quite amazed to see the faint spiral and two bright points for the cores on the exposure, be it slightly off to the top middle of the frame. Hope to revisit when I have more time and skies allow and try for extended exposures at 2 to 3 minutes and I know my scope isn't all that well suited for DSO's given it's focal ratio. But for me, it's progress. If I can get better definition I would be happy. I have a 0.5x reducer but cannot get focus with the camera but I have yet to try this without the diagonal mirror. Open to further advice, feedback and pointers.

Scope: Skywatcher 127 Mak (f/12)
Mount: Skywatcher EQ5 with SynScan GoTo
Camera: Canon 600D (unmodded)
Bortle: 6
Lights: 19x30 seconds at ISO6400 stacked in Siril
No darks, flats or bias. 

Auto stretched (...too flat looking?) and removed green noise.

Very much a work in process but thanks for looking.

Edited April 17, 2020 by rob_r
Added Siril processing detail

[vlaiv]

I think it is very good effort!

Don't think that Mak127 is not suitable just because it is "slow". Just be aware of it's limitations - and that is narrow field of view - which by the way works very nicely with targets like this.

I'm almost certain that above image contains more data and with careful processing, one could pull out a bit more. Processing is also a skill that one can improve on.

Not an expert on DSLRs, but I think that ISO6400 is unnecessarily high. Something like ISO800/ISO1600 would be a sweet spot - depending on camera model (I'm sure someone else will offer correct advice on your camera model).

Would like to encourage you to do a bit more subs - 19x30 is just shy of 10 minutes of total exposure. You can't expect much in that short amount of time.

A way to combat long focal length is to bin your data. That improves SNR, but makes smaller image (same FOV, but less pixels in height and width).

[oymd]

Very nice first attempt!

Wellcome to the club!!

just tread carefully, it becomes very addictive very quickly, and before you know it, you are sucked in!!

 

[rob_r]

3 hours ago, vlaiv said:

I think it is very good effort!

Don't think that Mak127 is not suitable just because it is "slow". Just be aware of it's limitations - and that is narrow field of view - which by the way works very nicely with targets like this.

I'm almost certain that above image contains more data and with careful processing, one could pull out a bit more. Processing is also a skill that one can improve on.

Not an expert on DSLRs, but I think that ISO6400 is unnecessarily high. Something like ISO800/ISO1600 would be a sweet spot - depending on camera model (I'm sure someone else will offer correct advice on your camera model).

Would like to encourage you to do a bit more subs - 19x30 is just shy of 10 minutes of total exposure. You can't expect much in that short amount of time.

A way to combat long focal length is to bin your data. That improves SNR, but makes smaller image (same FOV, but less pixels in height and width).

Thanks vlaiv, your feedback is much appreciated. 

So the 'binning' aspect, is this done at capture or once you have stacked all the subs? I take it as in my mind from a simplistic point of view, it is image reduction but keeping the aspect ratio. Siril can 'resample' using a scale percentage using a bilinear interpolation which I guess is the same as Photoshop's Image Size tool. So if I halve the width and height, would that be a 2x2 bin and halves the focal length?

I will certainly revisit this target with a lower ISO but with far more subs. To be honest, I got better than I was expecting out of the short exposure times. I'm eager to see if I can improve upon it. I'll try a 3-star alignment also to get better centred on target.

[vlaiv]

48 minutes ago, rob_r said:

So the 'binning' aspect, is this done at capture or once you have stacked all the subs? I take it as in my mind from a simplistic point of view, it is image reduction but keeping the aspect ratio. Siril can 'resample' using a scale percentage using a bilinear interpolation which I guess is the same as Photoshop's Image Size tool. So if I halve the width and height, would that be a 2x2 bin and halves the focal length?

You can bin at various stages and simplest one being after you've done stacking image and still in linear stage - prior to processing.

Binning is a form of resampling - so you end up with a smaller image in the same way regular resampling works. There is however difference between different resampling methods in what sort of impact they have on image sharpness and SNR improvement.

Mathematically bilinear resampling and binning are quite similar, in fact binning is like bilinear resampling + half pixel shift in both axis (small translation). This is true only for 2x2 binning (x2 resampling down). You can bin x3, x4, etc ... and it will be the same as resizing down x3, x4 in terms of pixel count, but impact on SNR will be different.

Binning has sort of best properties for signal to noise. It has always predictable improvement of SNR - same as bin factor. x2 binning will improve SNR by factor of x2, x3 by factor of x3 and so on.

It has similar effect as reducing focal length of scope - with one difference - field of view remains the same. You can't widen the field of view by binning. Maybe better way to think of it is "enlarging pixel size" rather than reducing focal length. There is another small difference which has to do with read noise, but those are details.

Given resolution of your image above - you can easily bin it x3 or x4 and still have respectable display size. I'm not sure if Siril can bin (probably can, I just have not used it so I can't tell) - you can upload linear image here and I'll do some preprocessing for you - like binning, background removal and simple color calibration and then post you linear file so you can have another go at processing it?

[rob_r]

1 hour ago, vlaiv said:

Given resolution of your image above - you can easily bin it x3 or x4 and still have respectable display size. I'm not sure if Siril can bin (probably can, I just have not used it so I can't tell) - you can upload linear image here and I'll do some preprocessing for you - like binning, background removal and simple color calibration and then post you linear file so you can have another go at processing it?

The only mention of 'binning' is on here, https://free-astro.org/index.php/Siril:Manual. It is under the 'RGB Compositing' tool. But, it looks like I need to split the image into the separate colour channels first. So, I'll have to figure out how to do that. I am trying to follow this, https://free-astro.org/index.php/Siril:Manual#Process_your_single_images to improve things but finding it a struggle to be honest. Anyhow, the linear file is attached.

result_16-04-20.fit

[vlaiv]

Well, this gave me quite a headache

Not sure if I made an improvement over your original processing - maybe a bit of color now shows.

Lack of calibration was quite a bit of a problem - here is what backgrounds look like in channels:

As you see, there is pattern evident, and I'm guessing it is bias - so dark calibration should remove it. I did some "trickery" and flattened the background. However red channel is too weak and very noisy and this shows. Here is color composition stretched:

I used green channel as luminance (I like LRGB workflow better as it gives me more control over denoising and stretching), so here is luminance channel:

Here I tried to minimize background noise and still show galaxy.

Together - above color and this luminance combined:

Both galaxy and stars show some color, but there is just too much noise in red so it dominates background.

Like I said, I'm not sure if above is any improvement on your original processing. In any case, if you want to have a go at cleaned up and binned data, here it is:

m51-c_00.fits

m51-c_01.fits

m51-c_02.fits