Jump to content

SkySurveyBanner.jpg.21855908fce40597655603b6c9af720d.jpg

1st m31 attempt. How much difference will dark frames make?


Recommended Posts

This is my first go at m31 the result of stacking three 1 min subs. Obviously I need a lot more data but how much difference would dark frames make to the camera noise on this image or is it insignificant on this amount of data? Any ideas what the ring artefact is to the right of the frame?

Taken with canon 1000d & skymax 127 on alt/az tracking mount.

3 min ISO 1600 :

post-4016-0-09557500-1411824190_thumb.pn

Link to comment
Share on other sites

  • Replies 36
  • Created
  • Last Reply

Definitely need darks.

On your image you have banding like on a old telly. The darks will reduce this some what. They are caused by the sensor temp.

You can create a dark library in the fridge by matching the temps.

Must say you are going to struggle to get a good image at f12 with a alt az

A much better recommend is to get a focal reducer (like a celestron 0.63) to get you down to f7.

And better still a eq mount such as a eq3 pro or eq5 pro.

Good luck.

Link to comment
Share on other sites

Never had a dark make any difference to horizontal banding on my 1000D. Biases I do take though, and these help with some it. However, the trouble is it is not always consistent from sub to sub, so you can't really subtract it all off. The only solution is lots of subs to try and average it out. It is more of a problem with alt-az as the pattern rotates with time -  on an EQ mount the bands are always in the same direction, so you can try debanding software on the stack.

NigelM

Link to comment
Share on other sites

Thanks Nigel I did take a couple of darks to try and they made no difference as you say. Will be interesting to see what happens when I take a decent amount of sub's. Just need some clear sky ahem lol

maybe 30 sec sub's would be better for the banding issue as the cam would stay cooler?

Link to comment
Share on other sites

In theory darks will help, but In practise I'm not sure how useful they are with DSLRs. The difficulty is knowing how temperature critical they are, if you're carrying out a long imaging run the sensor temperature may change significantly over its course. If the temperature profile doesn't match this could introduce artifacts.

If it's a choice between taking darks and more lights I'd be inclined to go for the lights - that will always improve the signal-to-noise ratio. I've given up on darks for the time being, pending further investigation. Here's an example of a 30 minute image with no darks or flats applied, noise isn't a problem. (However, if i was using slower optics or imaging a fainter target I'd have needed to stretch the data more which would have also multiplied any noise, making controlling it more critical.) 

Link to comment
Share on other sites

Been thinking about this and have changed my mind about the darks.

At short exposure lengths such as yours darks will make little difference.

Bias frames will help somewhat, but not that much. You going to need hours of data.

I would take many short frames, over at least 3hrs, if possible at longest exposure possible without star trails. (+- 45sec?) Might be necessary to nudge with handset every 10min to ensure tracking is stable.

Now stack these in DSS, using cappa sigma stack and 30+ bias frames.

This should provide some good data to play with processing wise.

To get good galaxy data with my mak I sometimes gather data up to 7hrs. But my mount is tracking with guiding, so I can sleep :)

Link to comment
Share on other sites

Thanks for your advice guys its much appreciated! Got just under an hour's data last night though the inconsistent tracking means DSS will only stack 30 mins at the most.

Just wondered if someone could tell me how you re-set DSS to the default settings as something has gone weird and it wont stack more than 17 mins with the resulting histo being pin thin producing a washed out image so I must have changed one of the settings somewhere.

Link to comment
Share on other sites

Here is a new version now with 27 mins of data, no darks but did use bias frames. Banding is still evident but signal to noise ratio is obviously much better. Colour balance not great. Curves and levels in CS2 then frequency domain filter and noise reduction used in image analyser.

post-4016-0-32179200-1412371029_thumb.pn

Link to comment
Share on other sites

It migth be worth changing the ISO. I read, a long time ago, that the 1000D uses different electronics for ISO1600 than for lower ISOs, Mine is signifcantly worse at ISO400/800 (I get vertical striping as well as horizontal at these ISOs!), but your camera might act differently.

NigelM

Link to comment
Share on other sites

Tried the 10" Dob last night as I've been trying to improve the tracking. Only got a a couple of subs before the cloud (and rain!) came over. This image is a 20 and 30 sec stacked in DSS and what strikes me is how much less prominent the banding is compared to the skymax 127 images.  The moon washed out the contrast (did not have a shroud on OTA) and as can be seen by the stars the tracking still needs attention.

52secs ISO 1600 10" Dob:

post-4016-0-05304500-1412505655_thumb.pn

Link to comment
Share on other sites

This image is a 20 and 30 sec stacked in DSS and what strikes me is how much less prominent the banding is compared to the skymax 127 images.

That could be because the dob is a much faster scope. With more signal in the subs you don't have to stretch the data as much, so any noise or artifacts aren't multiplied as much.

Link to comment
Share on other sites

That could be because the dob is a much faster scope. With more signal in the subs you don't have to stretch the data as much, so any noise or artifacts aren't multiplied as much.

Could also be temp, the noise halves for every 5C drop in temp.

Last night it was 8C around here, which means the sensor would have been running at 15C.

Link to comment
Share on other sites

The analysis below, gives you a clue on the two sensors used by canon . It does not include the pro series cameras such as the 6D

(600D, 650D, 60D etc vs 1100D, 450D, 550D) 

What this shows is that the dark noise in a cannon appears as a multiplier of both heat and time of exposure.

Thus, on a normal night where the sensor (+7c above ambient) is running at 20C, dark noise becomes a problem after 80min.

Analysis on Temp, Exp seconds before dark noise appears.

SN°C   I800

0°        1747

5°        779

10°      357

15°      169

20°      82

25°      40

Excludes read noise (Bias) which is always present.

See this great article http://www.christoph-kreher.de/Astro/DC.pdf

See page 7 and 8.

Link to comment
Share on other sites

That's interesting thanks, although some of that goes over my head. (Btw, I think you meant 80 seconds rather than 80 minutes). If I'm understanding correctly I wouldn't be getting much benefit from darks at the moment as I rarely use longer than 2 minute subs, but I should always be taking dark bias frames.

As I said above I also have some concerns about temperature matching between subs and calibration frames, but its difficult to say how critical that is. However, if dark current doubles with a five degree rise in temperature that sounds quite critical to me - how much does sensor temperature fluctuate over an imaging run? I'm going to continue to ignore darks for the time being, there are more important things for me to sort out. However, I'm going to make more effort to take dark bias frames, I've heard 30 is a good number. I will try taking them between switching targets, for better temperature matching.

Link to comment
Share on other sites

That's interesting thanks, although some of that goes over my head. (Btw, I think you meant 80 seconds rather than 80 minutes). If I'm understanding correctly I wouldn't be getting much benefit from darks at the moment as I rarely use longer than 2 minute subs, but I should always be taking dark bias frames.

As I said above I also have some concerns about temperature matching between subs and calibration frames, but its difficult to say how critical that is. However, if dark current doubles with a five degree rise in temperature that sounds quite critical to me - how much does sensor temperature fluctuate over an imaging run? I'm going to continue to ignore darks for the time being, there are more important things for me to sort out. However, I'm going to make more effort to take dark bias frames, I've heard 30 is a good number. I will try taking them between switching targets, for better temperature matching.

Yes, 80 seconds :)

If you are using DSS to stack it does have a tick box, to match the dark frames to the light frames. (actually creates a master and then does a stretch.)

I have created a dark and bias library (in the garage and the fridge) of +- 81 subs each. (1/9th the noise should remain.)

These are for my regular imaging times 180s and 300s for ISO800 and ISO1600, at 5C ambient and 15C ambient. 

This allows me to mix and match my master darks and bias to what I am imaging.

Ie.

@15C with Sensor temp +-7C, = 22C, this means I need 4 sets. 

@05C with Sensor temp +-9C, = 14C, this means I need 3 sets.

The 180s at ISO800, are (just) safe without darks and therefore, just bias needed.

edit - just had a look at my library and I have a 300s @ 0c for ISO1600 as well for deep winter. :)

Having said that on a short run of less than 2 hours, I use dithering between subs and cappa sigma stack.

When doing this I don't use darks, only bias and the cappa sigma will remove the noise.

However, after multiple runs of the same target, with 6 to 8 hrs of data, I use the darks and HDR stacking.

This tends to bring out more detail for me and the noise is subdued, by the sheer number of lights.

Anyway, hope this helps you in your thinking.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.