what have i done wrong with this image?

[lookus]

this is a shot of the horsehead. it is my first time with flats and bias frames. this has been stacked in DSS and processed in PS. i am learning to drive DSS and PS.

6x 15 min exposures
6x 15min darks( in know i need more)
10x flat
10x dark flats
10x bias

i experimented with different settings in deep sky stacker. but did not keep a very good track on all the settings. the thing is that this image has tonnes less noise than the other stack settings except it seems to have this overlapped rectangular frames at various angles. any ideas what may have caused this, as i am very happy with the image(compared to the others which were way noisier and seemed to suffer vignetting despite the flats) or what DSS settings to use.

i stacked with a mosiac setting
\median kappa sigma clipping
dark,flat and bias left on average
alignment on automatic

[[1ponders]]

Ok first issue is stacking with mosaic. Stick to General first. I'll do my HH and get the settings and get back to you with a sample image. It looks almost the same FOV and orientation and same exposure.

How long are your flats? If they are only a sec or 2 leave the Dark flats out of the equation for the time being

[[1ponders]]

For the time being set all your stacking to median. You can leave the lights at average if you like, but try all median to start with

[lookus]

Hi Paul. Thanks for looking.

the length of flats were 3mins. Is this OK? i wasn't sure but i thought i heard someone else once say they were doing 3 min flats. tell me if this is incorrect.

the image was taken with a 450d and a ed80 pro.

[[1ponders]]

You can take your flats as short as you want. I set my DSLR onto P and let the camera decide. Or set for 1 sec and adjust the brightness of your flat souce. The important thing is same ISO and even flat surface. If you only take a second then you don't need darks.

[[1ponders]]

Ok this is about 12/300sec shots Through a WO72FD with FR using a modded 20D. I used ICNR and 1 sec flats. The lights and flats were median combined and then the lights stacked using the general setting.

All the image has had is a crop (included an ofset image i shouldn't have) and 2 iterations of levels and curves.

with your data you should be able to achieve similar this but not as red.

[[1ponders]]

What did you shoot this through? You have quite a bit of curvature at the corners. Looks like it needs a field flattener. It looks like your camera orientation was either way out or you didn't line up closely.

[[1ponders]]

There is nothing wrong with the detail in the center. You can do much more with this image. Find you which images and how many are out of alignment. If there aren't too many keep them out of the stack process. But median combine and General stack parameters will probably help a bit. Also at this stage leave the bias out and see how it goes. The Dark frames already have the bias built into them so they aren't really needed for DSLR shots. JMO

[lookus]

thanks for all your advice. that is a fantastic photo. i clearly have no idea what i am doing in photoshop. (or DSS it seems)will try again.

the images were taken with a 450d and an ed80 pro.

i will restack the images with the settings you suggested and have another go.

will also look for a field flattener that will suit my scope.

[[1ponders]]

the WO P Flat II is the one you want for the ED80. It works a treat. There is a review of it in the review section.

[[1ponders]]

The initial part of PS isn't difficult. Crop out any truly offset stacking, use levels to raise the black point just to the left of the histogram and then using curves raise the black end and curve flatly to the white point. I think there is a how to in the resourses as well

[TrevorW]

Stick to darks and flats only forget bias frames

Also such things, as did you take from a dark site, moon, LP, type of filter. weather, is the camera modified etc will have an effect on image quality.

Cheers

[spearo]

I know someone who uses DSS with shots from a 20Da and he uses the cosmetic function in DSS 1pix/1%.
ok that was me...
stopped using darks because multiple tests with/without and wide rande of settings showed i got better results without darks, while good lights were stacked in Average (no interpolation) and using the cosmetics section which removes individual weird pixels (the noise.
might be worth a try.
I know...its heresy...
frank

[multiweb]

Looks like your subs weren't normalised prior to data stretching and stacking. It's quite hard to get right but using mean instead of sum or median when stacking minimise gradients a bit. You could also use a PS plugin called gradient Xterminator and do selective masking to get rid of the seams. It's a very nice shot. I reckon it's worth fixing. :thumbsup:

[lookus]

do bias frames complicate things? they were very easy to do so i thought i would do them as i thought it must be better than not doing them.

am i better off not doing them do you think?

[TrevorW]

Personally I don't use them and have read elsewhere they are not necessary

[Bassnut]

Stretched, I can see massive differences in sub alignment (as you say "overlapped rectangular frames at various angles"). You have some serious movement in your image train between subs, almost as if the camera is loose and flopping around (rotation and position), but only between subs, not during exposure. Are you bumping the scope or moving stuff between exposures?.

And, what was the ISO?, 15mins is a long exposure for a DSLR, yes low noise, but the bright stars look saturated.

[lookus]

Thanks for the ongoing help guys. no i did not bump or move the scope and all 6 light frames look good.

let me clarify. the difference in sub alignment is only apparant in this stack in which i had certain DSS settings(not sure what they were now as i was experimenting with several settings). it is not at all evident in other stacks using the exact same data only with different DSS settings.

the thing is that the stack where the sub alignment is out is the stack which has much less noise.

if you want a quick look here are 3 stacked images straight out of DSS. all using the exact same data, 2 of which show sub mis-alignment but with considerably less noise than the other one which is well aligned.

how do i get the low noise of the 2 mis aligned stackes combined with the proper alignment of the 3rd stack?

it must have something to do with the dss settings, however i have not been able to reproduce the error.

also-iso was 800.

[Bassnut]

OK, DSS is not aligning properly then and mistakenly offsetting a sub(s). Is there a manual mode where you can specify reference stars in each sub?. The "correct" stack maybe noisy because DSS has just rejected subs that cant be aligned rather than place them incorrectly as in the misaligned stack?.

[lookus]

here is a new go at the image using the same data, only no bias frames and DSS settings as paul suggested. it still has more noise than the mis-aligned one even though i used noise-ninja to reduce noise, but at least it is better aligned.do i need to darken the sky background more? how do i do this without clipping?

[TrevorW]

Try stretching it some more is PS