Stacking images and properly subtracting darks and bias frames

Hello everyone,

It has been really ugly here as far as astro weather is concerned. Cloudy everyday but 1 in the last 2 weeks. I am getting ready to do some imaging this season. The last few days have been spent simulating an imaging session to make dark and bias frames. I do not have a cooled camera so noise can fluctuate with the ambient temperature. I am making the dark and bias frames in about a 55 degree ambient temperature because that is really close to what I usually image in. I am allowing for 15 second intervals between all my light and dark images. I like the way DSS stacks images but was not too pleased with how it subtracted darks and bias frames from the light frames. I stacked all of the light frames as average and came up with a master light frame. I then stacked all the darks and came up with a master dark and did the same to the bias frames to come up with a master bias frame. I then subtracted the Master bias and Master dark frame from the master light frame in Photoshop. I then adjusted the levels, did a final noise reduction/high pass/luminance sharpening, and a mild contrast enhancement. Anyway, I want to be able to get more detail from some of the fainter objects this year. The cg5 mount is good for about 2 minutes tracking at most if you manual guide with a guide star so I decided to try to formulate something for ISO 6400 maybe ISO 12800 to see how they stack up with 30 images.

ISO 6400; f22; 15 seconds; All in camera NR disabled; 30 images stacked

Here are the images from start to finish
1 file from the stack

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

Wow! I was surprised at the lack of noise in the final resulting stacked image! Thanks for sharing

Troy

Here is the results at ISO 12800. This is pretty noisyISO, but with 45 subs and careful subtraction of the bias and darks frames I thinl it could be useable for some of the real faint objects.

The only difference in these as to the ISO 6400 is of course they are 12800 iso and the f stop is faster

1 image from the stack

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

Well done. Been thinking about trying this too for photography other than astro. You beat me to it.

Be interesting if you could post 100% crops of the above, maybe from darker areas and lighter areas. That would really show up the difference.

Strange...

tmmulske wrote in post #10164711
(...)

1 image from the stack

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO

the stacked light frames

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO

the finished product afetr darks, biases and processing as in the above post

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO

Thanks for checking it out

Todd

...Something does not look right, to me, with respect to the stacked light-frames. There is a drastic change in brightness, plus significant loss of bright chroma values.

When I stack (say) 8-frames, I end with with basically the exact same overall brightness, but almost three-times less noise, in both Luma and Chroma.

The final Stacked-Master Dark/Bias, also shows a dramatic jump of brightness... is this because your ISo12,800ec samples are not showing the true brightness that should be observed at such ISO values, even if amplified "digitally"?

PIX

PIXmantra wrote in post #10201927
...Something does not look right, to me, with respect to the stacked light-frames. There is a drastic change in brightness, plus significant loss of bright chroma values.

When I stack (say) 8-frames, I end with with basically the exact same overall brightness, but almost three-times less noise, in both Luma and Chroma.

The final Stacked-Master Dark/Bias, also shows a dramatic jump of brightness... is this because your ISo12,800ec samples are not showing the true brightness that should be observed at such ISO values, even if amplified "digitally"?

PIX

I am not sure what is up with that. Everytime I stack something in DSS this happens.

Todd

tmmulske wrote in post #10207581
I am not sure what is up with that. Everytime I stack something in DSS this happens.

Todd

...I will try to post something towards the weekend.

I do not use DSS, but a rather simple (yet accurate) blending model that runs well even the oldest versions of PS/CS (pretty much automated, you just feed the light frames and it comes out with the final output).

My blended results show virtually identical, uncompromised inter-light-frames tonality.

Cheers,

PIX