Rule of thumb or personal experience?

[+Stu]

Hi folks. 

I had a go imaging M33 last night. I ended up going for 30x 240s ISO800. I saw this as a safe bet as I didn't want to over expose and thought better under than over. 

Is there any guide on what settings to try, taking into account aperture, target magnitude, how dark it is etc.... Or is it just down to experience with your own setup and finding what works?

[Bottletopburly]

More the merrier with data I always think you need a minimum 4hrs of data to start  getting anything decent , aperture is king though on faint targets.

[Clive]

I expose individual images for as long as possible with my local sky background so that the individual image background score given by DSS does not exceed about 10% (so that's just 180 sec for my setup from my garden). Then I just get as many of those exposures as I can in the night (I used to be limited to 2 batteries worth or about 4 hours total or 80 images, but have since purchased an AC/DC adaptor for my DSLR). Having said that, I was quite happy with my only M33 image so far with just 28 x 180s exposures

 

 

[+Stu]

I'll check out that percentage thing on DSS and see what my images show 👍

[+Stu]

On 25/09/2022 at 17:39, Clive said:

I expose individual images for as long as possible with my local sky background so that the individual image background score given by DSS does not exceed about 10% (so that's just 180 sec for my setup from my garden). Then I just get as many of those exposures as I can in the night (I used to be limited to 2 batteries worth or about 4 hours total or 80 images, but have since purchased an AC/DC adaptor for my DSLR). Having said that, I was quite happy with my only M33 image so far with just 28 x 180s exposures

 

 

I checked the lights and they range from 2.4 - 4.3%. So I maybe could've gone with ISO1600 instead of 800?

[Clive]

The best ISO for your camera may be different to that for mine so keeping it at ISO 800 may be the best option for you. A 2% to 4% background level is fine. You won't loose or gain any more of the faint stuff by increasing it, just the bright stuff will saturate quicker. At the end of the day, as previously mentioned, it's all about getting as much exposure time as possible so you can extract the stuff you want to see from the noise created by the background sky pollution.

[+Stu]

Ok cool. I've had a couple of gos at processing it and I'm starting to get my head round it now, I think. I'll post the result when I'm happy with it. 

[+Stu]

 

This is what DSS churned out. I've had a play with it in GIMP but can't really make it what I want. Am I flogging a dead horse or more practice with processing?

 

*Posted wrong image. 

 

 

Edited September 27, 2022 by Stu

[+Stu]

It was the right one, doh 🤦

It was 30x lights, 20x darks, 24x flats, 50x bias. 

[Clive]

It thats just the output from DSS then there's plenty to work on. I don't know GIMP but a simple ASINH stretch should pull out the details and colours.  And lots of other simple processing techniques will enhance the image even further.

[+Stu]

Cool, I've been following Making Every Photon Count, step by step. Is that the "initial histogram stretch" it refers to?

[+Stu]

It also mentions Digital Development Processing. Is that a seperate piece of software? Or something within GIMP/Photoshop?

[Clive]

19 minutes ago, Stu said:

Cool, I've been following Making Every Photon Count, step by step. Is that the "initial histogram stretch" it refers to?

I'm not sure, I tend to make the ASINH stretch the last thing I do after all the stuff done on the linear image data. So certainly my cropping, gradient removal and basic colour correction is all done on the linear data before applying any curves (of which ASINH is a special type that preserves the original unstretched colour unlike just applying a common set of curves to all the colour layers)

[+Stu]

1 hour ago, Clive said:

I'm not sure, I tend to make the ASINH stretch the last thing I do after all the stuff done on the linear image data. So certainly my cropping, gradient removal and basic colour correction is all done on the linear data before applying any curves (of which ASINH is a special type that preserves the original unstretched colour unlike just applying a common set of curves to all the colour layers)

Been doing some searching and there's a lot of recommendations for SIRIL for that stretch. I'll have a nose later. 

[Clive]

22 minutes ago, Stu said:

Been doing some searching and there's a lot of recommendations for SIRIL for that stretch. I'll have a nose later. 

It's a stretch method that's been around for years. I would be surprised if any of todays astro processing software didn't include it in some form (but I wonder if some of the newer software just consider it to be too 'old fashioned' and don't include it?). My 2010 version of IRIS (the predecessor to SIRIL - IRIS backwards + L for Linux) implements it and I always use it to some extent.

[+Stu]

10 minutes ago, Clive said:

It's a stretch method that's been around for years. I would be surprised if any of todays astro processing software didn't include it in some form (but I wonder if some of the newer software just consider it to be too 'old fashioned' and don't include it?). My 2010 version of IRIS (the predecessor to SIRIL - IRIS backwards + L for Linux) implements it and I always use it to some extent.

I had a quick look but will experiment more later. It's similar to the levels adjustment I was doing in GIMP. It makes it brighter and the background becomes much lighter. Is that the desired effect?

[Clive]

I performs a non-linear muliplication to the data (just like a logarithmic or gamma curve stretch) but in such a way that it preserves the ratios of the linear RGB data (so the resulting colour remains the same) whilst still enhancing the faint detail in the image.

[+Stu]

That's all a foreign language to me to be honest, it's only the 2nd image I've had a proper go at processing. I'll have a more in depth attempt in a bit.