DSLR - optimising imaging

luka · #1 06-04-2016, 04:20 PM

Hi

I finally got guiding working and now I am thinking how to optimise the capturing process. I live close to the CBD with lots of light pollution, VLMag is about 4.5.

1. For my IR modded 1200D the optimum ISO is 1600 (or 3200). That was easy to find here.

2. How to set exposure time to maximise the signal to noise ratio considering the light pollution?
I am finding lots of conflicting information about this. Some argue that histogram should be starting 10% off from the dark side, then I read it should be about 1/3 of the way or 1/2 of the way and then there is Expose to the Right method.
This covers pretty much the whole histogram range.

How do the more experienced images here pick their exposure time? From a dark site I am guessing that longer is better but what about from light-polluted places?

Thanks in advance
Luka

glend (Glen) · #2 06-04-2016, 04:39 PM

There is an element of trial and error in working this out for your location. Your scope f ratio is 7.5 so It's not particulary fast. But I think you have picked the right ISO, I use ISO1600 with my f8 RC08. Without going into all the jargon, pick a bright object like Eta Carinae and run some test subs at various times and check the histogram. I'd suggest between 180" - 240" to start with but if your location over powers that you might have to reduce it and shoot more shorter subs. You could try a CLS filter but you will need to increase your sub lengths to compensate. Have fun.

luka · #3 06-04-2016, 04:46 PM

Thanks for your reply Glen.
Where should I aim for the peak of the histogram to be? Middle of the range?

glend (Glen) · #4 06-04-2016, 05:07 PM

I would suggest about 25% from the left but you want to avoid saturation. In my case the camera is cooled so I don't have to worry about heat buildup and noise as a result. You are going to be building up heat so keep that in mind. As I said everyone's light pollution situation is different. I am certainly not an expert but just worked out what suits my location the best.

traveller (Bo) · #5 06-04-2016, 06:02 PM

Luka, I experiment with my DSLR settings. I set ISO at 800 for brighter objects like 47 Tuc and M42, but for fainter objects, I go for 1600.
I have done 5 mins at 1600, but they are very noisy (sensor running at 30c) and darks were not that useful.
I have experimented at around 3 mins at 1600. I found that for fainter (mag 7 or there abouts) DSOs, that will get histograms around the middle or just off to the left (2/5-3/5 of the histogram), but that's still quiet noisy (sensor around 20-23c). I will need to shoot matching darks to see if they can be toned down a bit more.
I think I will stick to this setting as the cooler months set in and see how I go.
On the other hand, DIY cooler for the DSLR....

glend (Glen) · #6 06-04-2016, 07:39 PM

Quote:

Originally Posted by traveller

Luka, I experiment with my DSLR settings. I set ISO at 800 for brighter objects like 47 Tuc and M42, but for fainter objects, I go for 1600.
I have done 5 mins at 1600, but they are very noisy (sensor running at 30c) ....

Bo where are you getting that sensor temperature data from? The CMOS sensor, does not contain a temperature sensor. The temp data you see in the EXIF is actually the DIGIC processor temperature which is on the separate main board. The processor is likely the hotest part of the internals but can be used as a proxy for the CMOS although it will actually be a little cooler. The processor is heating up the other components but not directly. If you have a multiprocessor camera, and some later mdels have two DIGIC chips then more heat is generated. You can try to reduce processor workload, and heat, by turning off functions that require the processor - like internal noise reduction, camera dark generation, etc. All those auto functions generate more heat. If you shoot separate darks, and bias frames, you don't need the internal dark generation or noise reduction processing. Of course cooling the CMOS sensor directly is the only way to get frames and darks to match at any sub duration. In that situation, where bias and darks match (which is around 0C in a 450D) you can stop doing darks, or just shoot a group at 0C and then use the Master Dark and Master Bias for all your stacking.

traveller (Bo) · #7 06-04-2016, 08:29 PM

Thanks Glen,
I am simply reading the temp from BYE display, didn't realise difference between sensor v processor, so learnt something new today.

Bo

luka · #8 06-04-2016, 09:54 PM

Thank you for your replies.

As always, nothing is simple. I thought that once I have guiding all my problems will be solved and that I can do 10+ minute exposures. Last night I realised that, while the mount can do it, the light pollution severely limits the imaging time. And now I know how bad the noise is as well.

This is funny, last night I took darks for 5min@ISO1600 and when I looked at them today I initially thought I forgot to put the cover on the scope. It took me a while to realise it was noise as the stars "did not look right"

traveller (Bo) · #9 06-04-2016, 11:26 PM

Do you use a LP filter? I use the Baader UHCs which helps a bit (a bit too much filtering IMHO) but it's a trade off between hot pixels and exposure time for me.

luka · #10 07-04-2016, 01:01 AM

No LP filter. I was hoping to pick up one from the classifieds section. It may be a long wait, they tend to get sold very quickly..

glend (Glen) · #11 07-04-2016, 01:59 AM

Many factors have to come together to get 10+ minute guided images, experience is the best teacher. Sadly your camera is probably not going to help as the noise is going to limit your sub times. Perhaps in the middle of winter you can run longer subs. Even with my cooled full spectrum DSLR I am only running six minute subs in good conditions at f8 but I run longer subs with my mono narrowband cooled DSLR. I prefer shooting colour with my newt at f5 because I cwn run shorter subs. Heat and noise are not a problem as the sensor sits at just above 0C, but tonight the wind came up and ruined my guide graph so I had to shutdown. My system sits in an observatory but wind can mess up good guiding pretty quickly.
Don't rush it, stick with brighter DSOs and nebulas to start with where you can run shorter subs, say under five minutes. You have the processing side to learn as well. I sort of see an 8" newt as ideal to learn on, they are fast, give true colour, and reasonable focal length to let you see some detail, they are also cheap to buy, light weight, and can sit on an HEQ5.

ZeroID (Brent) · #12 07-04-2016, 12:20 PM

Hey Luka,
I'm running anything from 30 to 120 seconds @ ISO 1600 on my Lunt at F7 for best results. I aim to get the histogram 1/3 of the way across from left in BYE, the first line on the screen I think.
Actual time will depend which part of the sky and the sky condition.
East over the CBD is the 30 sec normally but west towards the hills can get me up to 120 secs. Cloud, haze, seeing, turbulence dependent of course.
If the noise is too much at longer exposures take more shorter subs is the answer.
My first positioning frame is about 10 secs to see if I'm on target, then correct framing and time and upscale appropriately. BYE is good for that. Every night is a little different.