M101 - am I pushing the limits of my setup or more to learn?

[SiD the Turtle]

Hi folks, I've been practising, practising, practising on M101 with my setup below to try and hone my capture workflow before I start scanning the sky for more targets. Let alone the image workflow- my trials for Photoshop and Pixinsight have now expired so I'm now down to Paint.net until I decide which one I prefer and can afford!

This image is very lazily stretched and curves applied, I haven't spent any time with star reduction, setting the blackpoint, reducing noise etc, I just wanted to get input at this point:

The images are still very noisy which I think is down to imaging on very warm evenings (sensor was between 25-35 degrees Celsius). I know in PixInsight there is a de-banding filter that will get rid of the Canon banding at least.

I'm tempted to switch out the OTA to something faster but I do like the ability to get up close with the galaxies for now. Thoughts? I have ~£200 to spend now that next in line is either going towards Pixinsight or some dew heaters, USB hub etc, or I can put it towards a new, faster OTA. My reticence is (if I understand correctly) the increased FoV on the faster scopes would mean smaller galaxies, and I'm really interested in galaxy photography for now, before I move on to nebula etc.

Setup

• OTA: Celestron 127 SLT
• Mount: Skywatcher HEQ5
• Camera: Canon 600D unmodified
• Guiding: SkyWatcher Evoguide 50ED + ZWO 120MC-S + PHD2
• Computer Hardware and Software: INDI/ASCOM/EKOS on a Raspberry PI4
• Imaging Software: Deep Sky Stacker, Paint.net (crop, levels and curves).

Images

• Lights: 62 x 300s
• Darks: 53 x 300s
• Bias: 431 (for comedy value, I took 200, lost them, then took another 200-odd then found the original. On another forum someone suggested this number of bias frames could be creating noise)
• Flats: 60 (missed flats on the final night)
• ISO: 1600
• Bortle class 5
• Across 3 nights of imaging.

Edited August 4, 2020 by SiD the Turtle

[MarkAR]

Some good detail showing. I think there will always be more to learn and as for the next step I would consider a cooled dedicated camera.

Using this 

https://astronomy.tools/calculators/ccd_suitability

it looks like you are slightly oversampling. A camera with larger pixel size (6+ um) would be better for your scope.

[alacant]

Excellent effort. If you like galaxies, I'd stay exactly where you are. It looks miserable when you've a small blob in the middle of a vast expanse of sky.

With a dslr, best to lose the darks and take bias along with flats. Dither between light frames and stack with a clipping algoithm. Darks will introduce more noise, as you're finding. The dither will control the noise considerably.

FWIW, I was relieved when my trial of PI expired;)

Cheers and look forward to some more galaxy shots.

Edited August 4, 2020 by alacant

[vlaiv]

Just a couple of thoughts that you may find helpful:

1. M101 is rather faint object and this is only ~5 hours of data

2. Mak127 has rather long focal length and paired with your camera - you are over sampling. This means that both dark noise (non cooled camera) and read noise will have more impact. It is good that you decided to go with longer exposures of 5 minutes, but do consider binning your data as a processing step.

3. Give Gimp 2.10 a try over Paint.net and others - it is very good software for image processing once stacking is done and it's free

4. Don't think in terms of focal length and field of view when thinking high resolution imaging. Think in terms of what is realistically best resolution that I can achieve and then what is the largest aperture that you can achieve it given limitations of your camera and mount and all of that (do consider different camera as well - sometimes investing in camera makes the most difference) - also, think binning as well - even if software as a part of equation of resolution / sampling rate.

[SiD the Turtle]

Hi all, just wanted to say thanks for all the input and advice. Especially regarding over-sampling which I hadn't considered in the past.

One question- can I bin using a DSLR? Surely as it's a colour camera with a bayer filter if I bin i'd be mashing together a random combination of R, G and B pixels?

[alacant]

2 hours ago, SiD the Turtle said:

can I bin using a DSLR?

Hi

Yes. There's an excellent software bin algorithm in StarTools. It's saved me on many occasions;)

Cheers and HTH

[ollypenrice]

On 04/08/2020 at 15:12, vlaiv said:

 

4. Don't think in terms of focal length and field of view when thinking high resolution imaging. Think in terms of what is realistically best resolution that I can achieve and then what is the largest aperture that you can achieve it given limitations of your camera and mount and all of that (do consider different camera as well - sometimes investing in camera makes the most difference) - also, think binning as well - even if software as a part of equation of resolution / sampling rate.

Vlaiv's point above hits the nail on the head. Forget about how much sky you have around your target galaxy. This has absolutely nothing to do with with the 'natural size' of the galaxy's image. By its natural size I mean its size on screen at 1:1. That's one camera pixel being given one screen pixel. If you take two cameras with the same sized pixels in the same scope, one camera having a huge full frame chip and the other a small astro-camera chip, the galaxy will be exactly the same size and have the same resolution in both images. You'll just have to crop out the spare sky if you don't want it there! Another way to think about it is to ask yourself how many pixels you are gong to put under the galaxy's projected image from the scope.

In order to keep it simple and accurate think strictly in terms of arcseconds per pixel. This is the real unit of resolution and arises from focal length and pixel size. Your guiding, your seeing and your optical resolution will conspire to set a limit on this. You need some idea of what that limit might be, though it varies considerably with the seeing. The FWHM value (essentially a measure of star sizes on the chip) can easily double on a bad night. Personally I find that I'm usually over and out at roughly an arcsecond per pixel though I actually sample at 0.9"PP. That gives a satisfying scale on galaxies. Your guide RMS needs to be about 0.5 arcsecs for this. Good HEQ5 mounts can deliver this with careful setting up.

I think your M101 is good. Stars are tight and round. With a DSLR I would suggest a large dither, about 12 pixels, to combat noise. This, combined with a large number of subs stacked using a sigma reject algorithm, will give you very nice data to work with.

Olly

[wimvb]

With a budget of about £200, and your current setup, I would start looking at processing software. Test the ones available, PS, PI, APP, startools, then go for the one that suits you and your hardware best. I found PS subscription model to become too expensive in the long run. PI offered all I needed; startools and app were not an option at that time, but are so nowadays.

[SiD the Turtle]

On 09/08/2020 at 09:50, ollypenrice said:

...

Thanks for this, working out this stuff melts my brain. Have I got this right though, that because of the 127 MAK's focal ratio compared to a faster scope, the field of view is reduced, the galaxy takes up a larger amount of space on the camera sensor and therefore you get more pixels making up the galaxy? However, that's the light spread across more pixels and therefore more, longer exposures are required. With a faster scope the galaxy is spread across a smaller set of pixels therefore you need less subs, but the size of the galaxy is much reduced.

Correct?

[ollypenrice]

1 hour ago, SiD the Turtle said:

Thanks for this, working out this stuff melts my brain. Have I got this right though, that because of the 127 MAK's focal ratio compared to a faster scope, the field of view is reduced, the galaxy takes up a larger amount of space on the camera sensor and therefore you get more pixels making up the galaxy? However, that's the light spread across more pixels and therefore more, longer exposures are required. With a faster scope the galaxy is spread across a smaller set of pixels therefore you need less subs, but the size of the galaxy is much reduced.

Correct?

Almost, apart from, 'because of the 127 MAK's focal ratio compared to a faster scope, the field of view is reduced...'  The field of view is reduced because the focal length is longer, not because the focal ratio is slower. My advice would be to remember that focal ratio is what's called a derived variable. It describes the relationship between two primary qualities of a telescope. These are its aperture and its focal length and they are hard, physical properties which never deceive. 

In AP we have to be wary of focal ratio because it can be deceptive. We cannot just say 'F2 is twenty times faster than F10' (as is claimed for the Hyperstar focal reducer, for example) because the aperture is unaltered, so the amount of light from the galaxy is unaltered, so how can the exposure time be reduced? By making a tiny image of the galaxy rather than a big one is the answer. Not very impressive after all!

To start from the beginning, you want a certain resolution in your galaxy images. Say you choose 1 arcsec per pixel. You need a scope and camera combination which gives about that. So your pixel size is now matched to your focal length. Now, how fast do you want it to be? That will depend on aperture. (Yes, more aperture will mean a faster F ratio but don't be fooled. The 'active ingredient' here is aperture. More light. The 127 Mak has rather a small aperture for its focal length. The same focal length with more aperture would be preferable for deep sky imaging. (I'm putting it this way rather than saying 'The Mak is a bit slow for DS imaging' because I want to keep focal length and aperture up front in the explanation given the confusion brought by F ratio.

Olly

[vlaiv]

1 hour ago, ollypenrice said:

Say you choose 1 arcsec per pixel.

I would say don't

1 hour ago, ollypenrice said:

The 127 Mak has rather a small aperture for its focal length

Maybe even better statement would be - Mak 127 has rather small fully corrected and illuminated field for its aperture?

[Ouroboros]

Pretty good data and resulting image given the quick process. I bet you could extract more from that data given more time. 

Just a thought about Canon banding.  What works for me, and is recommended, is to apply it to all the light frames after calibration, cosmetic correction and debayering but before alignment and integration. This is done with batch processing and the image container script. 

[ollypenrice]

2 hours ago, vlaiv said:

I would say don't

 

Oooh! OOOHHH!!!  But that is not what you said when I posted a comparison of 0.6 and 0.9 "PP images of M101 and M51! You said you saw more detail in the 0.6"PP versions. I said I didn't think it was worth bothering with!!

lly

By the way (on more serious matters) I had to copy and paste your green grin icon since I don't have it offered any more. I like that icon. I feel it resembles me to perfection! 

 

[tooth_dr]

 : and D

22 minutes ago, ollypenrice said:

Oooh! OOOHHH!!!  But that is not what you said when I posted a comparison of 0.6 and 0.9 "PP images of M101 and M51! You said you saw more detail in the 0.6"PP versions. I said I didn't think it was worth bothering with!!

lly

By the way (on more serious matters) I had to copy and paste your green grin icon since I don't have it offered any more. I like that icon. I feel it resembles me to perfection! 

 

 

[vlaiv]

24 minutes ago, ollypenrice said:

Oooh! OOOHHH!!!  But that is not what you said when I posted a comparison of 0.6 and 0.9 "PP images of M101 and M51! You said you saw more detail in the 0.6"PP versions. I said I didn't think it was worth bothering with!!

Indeed I did but that was in relation to a bit more serious equipment if my memory serves me well.

Something like 12 or more inch newtonian and probably serious mount?

Here we are talking about 5" scope mounted on HEQ5 mount.

Over the time, I think I developed rule of thumb for appropriate lower sampling limit (or upper if you will - never now what would be appropriate term - lower in number, but higher in "magnification"):

<80mm 2-3"/px

80-100 1.7-2"/px

100-120 1.5-1.7"/px

120-150 1.3-1.5"/px

and so on.

I would not go for 1"/px unless I had at least 8" aperture and very good mount. In fact - I do have 8" aperture and less then serious mount and I have system that samples at 1"/px and I believe that 95% of time it is over sampling.

[SiD the Turtle]

4 hours ago, ollypenrice said:

Olly

Thanks Olly, great explanation.

[ollypenrice]

10 hours ago, vlaiv said:

Indeed I did but that was in relation to a bit more serious equipment if my memory serves me well.

Something like 12 or more inch newtonian and probably serious mount?

Here we are talking about 5" scope mounted on HEQ5 mount.

Over the time, I think I developed rule of thumb for appropriate lower sampling limit (or upper if you will - never now what would be appropriate term - lower in number, but higher in "magnification"):

<80mm 2-3"/px

80-100 1.7-2"/px

100-120 1.5-1.7"/px

120-150 1.3-1.5"/px

and so on.

I would not go for 1"/px unless I had at least 8" aperture and very good mount. In fact - I do have 8" aperture and less then serious mount and I have system that samples at 1"/px and I believe that 95% of time it is over sampling.

I'm using 140mm at 0.9"PP and like the result. (The comparison in our earlier conversation was between this and a 14 inch ODK at 6.2"PP which I felt was oversampled.)

I agree that the OP would be pushing it going for 1"PP and your guide above agrees broadly with my experience.

Olly