Little Horse experiment

TrevorW · #1 02-01-2009, 05:16 PM

This was done as an experiment, I can't use my guidescope at the moment so I wanted to see how much I could capture with only 40 second exposures and bring out in processing for future reference

total 17 minutes 26 frames ISO800

anthony.tony · #2 02-01-2009, 06:19 PM

Not Bad .

Omaroo (Chris Malikoff) · #3 02-01-2009, 07:05 PM

Not a bad try Trevor. What I can't understand though, is where the noise is coming from. With 40-sec exposures from a 350D there should be absolutely minimal noise at ISO800. "Mean" stacking should take care of the noise to a degree, so I presume you applied additive stacking instead because the horse head is still apparent even with such a short run of exposures...

You're on the right track though!

TrevorW · #4 02-01-2009, 08:43 PM

Noise was probably induced in processing to bring out the horse attached is an unprocessed image as stacked in DSS and cropped. Although last night in Perth it was over 30C when I was taking these frames so this probably didn't help.

Cheers

leinad (Dan) · #5 03-01-2009, 04:33 PM

Nice attempt Trev. Bet you're pulling your hair out waiting for that laptop.
Was that taken through the 120mm or 80mm ?

TrevorW · #6 03-01-2009, 04:52 PM

Leinard I'm going to nick my daughters laptop and load PHD on it so fingers crossed and clouds permitting I'm going to try again tonight as well as try out my cooler unit to see if it makes any difference. All my shots are through the 80ED these days.

Cheers

TrevorW · #7 03-01-2009, 05:00 PM

Quote:

Originally Posted by Omaroo

Not a bad try Trevor. What I can't understand though, is where the noise is coming from. With 40-sec exposures from a 350D there should be absolutely minimal noise at ISO800. "Mean" stacking should take care of the noise to a degree, so I presume you applied additive stacking instead because the horse head is still apparent even with such a short run of exposures...

You're on the right track though!

DSS is set on Kappa Sigma clipping as a default

Cheers

leinad (Dan) · #8 03-01-2009, 05:24 PM

Quote:

Originally Posted by TrevorW

Leinard I'm going to nick my daughters laptop and load PHD on it so fingers crossed and clouds permitting I'm going to try again tonight as well as try out my cooler unit to see if it makes any difference. All my shots are through the 80ED these days.

Cheers

Try without the cooler to start with. Remember longer the exposure, better the S/N. You could also take 10-15 Bias frames to remove any shot noise.
Id be very interested to see the cooler results though. Especially during these 20-28degree nights. Hate humidity !

I think if you set the 350D to Parameter 2, this will apply ICNR. doesnt actually say ICNR, but Parameter 2 will turn this feature on. I might try this tonight also. Drawback is a twice as long exposure, but it would be just the same I gather with taking darks anyway; only spending more time on a target then taking your darks afterwards.

TrevorW · #9 03-01-2009, 05:37 PM

I capture flats while it's still light, then do a series of darks, then the lights, then another series of darks. I'm actually not keen on in camera noise reduction as it doubles the time it takes for a frame.

Cheers

Peter Ward · #10 03-01-2009, 05:57 PM

Quote:

Originally Posted by TrevorW

This was done as an experiment, I can't use my guidescope at the moment so I wanted to see how much I could capture with only 40 second exposures and bring out in processing for future reference

total 17 minutes 26 frames ISO800

I found this really interesting as it highlights many areas of digital image capture and processing that are often not well understood.

For example, shot noise is not removed by bias frames. (shot noise is statistical in nature and related to the flux of the object)

It also highlights a very important aspect of stacking images: you need to capture signal faster than you capture noise for the process to be worthwhile.

For dim objects, the signal level of the object approaches the noise floor
and stacking will only increase the signal marginally faster than the noise...hence the result you posted.

Some better known fixes are:
improve the S/N ratio by moving to a dark site, high contrast optics, and perfect focus
use a larger aperture,
a higher QE sensor,
lower the sensor noise (ie cool it)

BTW I covered much of this in my talk at the CWAS Astrofest in 2008

TrevorW · #11 03-01-2009, 06:53 PM

Thanks Peter as usual your knowledge is tutorius and welcome !!!