Noob questions AUTO STRETCH ? BLINK

[Ron Kramer]

I'm shooting subs with a 1600MMc.

When I use blink, I see LOTS of variation in the images.  I can auto stretch by clicking the icon (it then says it applies it to all images). Problem is, not all images match!  Some are very milky, some are dark. So there is no one auto-stretch that is good for all of them.  My question is why?

I shoot HA and some are good with darker background lots of detail and others are lighter.  (could it be clouds?)
I then view my o3 images which looked good in SGP when shooting and they're very milky light.  If I apply the auto stretch in blink to them, they can look fine.  Should I be stretching "perminently" each image before processing or before stacking?

I did a calibrate, and a cosmetic and then a star align.  I stacked and had ok results. But if I blink through the  C_C_R  files I still see milky and nice visible frames?  shouldn't they all have best histogram settings applied before they're stacked?

Do I leave the mismatch - or do I delete the milky frames?  Thing is - if I do a auto stretch on the milky frame it looks decent and the other frame that looked decent now looks to dark. 

[pfile]

leave the mismatches. the integration tool handles normalizing the frames. don't stretch anything but your integrated light master. all parts of the pre-processing flow must be done on linear images.

rob

[Ron Kramer]

ok thanks - so after it's integrated say "all my HA"  I then take that as a light HA master and that should be stretched?
O3 master stretched?
SII master?   

(or each R G and B?)  Or do you mean integrate all those first then final adjust and stretch the final image? to present?)

Also any idea why my HA subs are all over the place - light and dark?  (haze? clouds)?

[pfile]

hard to say why the backgrounds look like that - could be moonlight?

if you are doing a narrowband image, you'd integrate each filter separately (taking care to register all subs from all filters against the same sub, say the best Ha sub) and then assign them to the channels of an RGB image (say with pixelmath or ChannelCombination) and then process the image as an RGB image.

you might want to look into warren's book, all of this is covered in great detail.

rob