M31 Just a blurry dot

[montezuma110]

When I was out last night i captured this image. Its 55 lights frames at 30 seconds each @ f6.3, iso 800. Taken by a 40d on a alt az go to mount, with a 250mm lens.

What do i need to do differently to differently in order to get an image of the full galaxy not just its core.

I cant really increase the shutter speed as above 30 field rotation becomes a big problem.

If any one has any advice it would be greatly appreciated

Thanks

[atmenterprises]

You could try upping the ISO to 1600. Aside from that you're probably going to need to increase your exposure time. Many of the great shots I see of the full galaxy are five or 10 minute exposures. In fact, saw a single 600-second exposure on here recently that showed a full galaxy.

[rfdesigner]

You need to 'process' the image, you have a lot more than just the core.

First take some "flats".. that is some frames of a plain white object, ideally held very near to the lens while the camera is still focussed on infinity.

Then use this to 'flatten' the image, having done that then you can process the image, compress the contrast so as to reveal the dim outer edges of the galaxy.

I've tried to flatten and compress this but as I'm starting with a JPEG most of the data has been lost, if you do the same with a RAW frame it will start to look much better, I have a colour cast in this that is due to the 8-bits in JPEGs, with the RAW data you won't get that.

Derek

[dph1nm]

I agree - there should be a reasonable image lurking in there somewhere, given the exposure time and focal length. Changing ISO won't help.

On my web page you can see M31 , 110min in 30 subs at 250mm FL with a Canon 1000D, F5.6

http://www.dur.ac.uk/nigel.metcalfe/astro/canon1000d_250.html

which might give you some idea of what to expect. I do have a lot of light pollution, so you could do better from a dark site.

NIgelM

[montezuma110]

Thanks for the advice.

I have done some flats and bias and reprocessed in dss.

Derek, what exactly do you mean by compressing the contrast?

[tingting44]

Thanks for the advice.

I have done some flats and bias and reprocessed in dss.

Derek, what exactly do you mean by compressing the contrast?

cant wait to see the out come, ive never bothered with flats or bias, so will be good to see if they do make an improvement

[rfdesigner]

In photoshop it is called 'curves' in DSS, I'm not sure. what you are doing is applying a conversion so dim bits of the image are brightened a lot while the bright bits are changed very little.

I use IRIS and in that you get a "contrast adjustment" tool, although I also use the "log" tool which applies a mathematical log function to the whole image.

Derek

[montezuma110]

Ah, ok I know how to use curves.

Will post the new one tonight.

Will a composite of 100 30 second exposures be similar to 50 60 second exposures?

[rfdesigner]

Similar, probably a little more noisy though.

With 100 x 30 seconds you have 100 lots of readout noise, with 50 x 60 seconds you only have 50 lots of readout noise.

Ideally you want the skyglow to swamp the readout noise. If your camera has readout noise of 8e (common on decent DSLRs) then you want to have your skyglow at roughly 10x(8e)^2 or 640e, this will nicely swamp the readout noise without having excessively long frames. The gain of the camera will determin what reading you get in your image, but the 40D at ISO400 give a ADU/e of 0.85 so it's fair to guess that at ISO400 you want a background glow of 600 ADUs and at ISO800 you probably want to be around 1200 ADUs. Your camera will be a little different but these figures could be a good starting point (~ +double/-half)

Derek

[montezuma110]

Still struggling to get a good image, guessing there just isn't enough data there.

New image with flat and bias.

Any suggestions?

Thanks

[rfdesigner]

How did you take the flat and bias?, and how did you apply them. They've certainly helped, but it looks like they're not quite right.. the flats need to be taken under conditions as nearly identical to the lights as possibl. Also the lens/scope needs to see plain white in all directions from the front lens.

The sum for making the flat fielding work is:

(light - dark ) / (flat - bias)

if you don't have darks then you may substitute bias frames.. if you just do: light / flat or light / (flat - bias) then it won't quite work.

Also ALL images ALL the way through the processing must be RAW or full dynamic range fits / tiffs etc.. i.e. NO JPEGS

Process order is: (not definitive.. but this will hopefully clear up most of the big issues)

1. RAW capture

2. subtract bias / dark

3. divide by flat field

4. demosaic

5. adjust the background (black level) of each colour by adding or subtracting to each RGB channel to make the backgrounds equal.

6. adjust the foreground level of each colour by multiplying or dividing to each RGB channel.

7. play with sharpening etc.

8. apply curves and noise reduction.

you are getting there, it takes a while to get your head round the whole process.

Derek

[montezuma110]

Which steps of this process are performed in DSS and which in photoshop?

All my frames are RAW, and my flats were same iso, focus and focal length.

I just load all my frames into DSS and then edit the tiff in photoshop.

Is this the correct procedure?

[ollypenrice]

cant wait to see the out come, ive never bothered with flats or bias, so will be good to see if they do make an improvement

If you use darks, as you should, then the bias is contained within that but a master bias can be used as a 'dark for flats'.

Do flats make any difference? Yes, they make all the difference. The image has a bright inner circle and dark corners, an effect known as vignetting. Flats will cure this at a stroke and allow you to separate faint galaxy from the uneven illumination.

Olly

[tingting44]

thanks for clearing that up olly wish i could have a go tonight, only cloud about for me