Masks Again

[marekc]

Hi Steve,

I fear I may be going a bit off-topic, but if you want a good `icebreaker' with ATWT, Juan's post on ATWT-based noise reduction is very good:

http://pixinsight.com/forum/index.php?topic=3184.msg21961#msg21961

In that post he uses ATWT to reduce noise on some data that Harry shot. It's a relatively `gentle' introduction to ATWT. I followed it last night, when I was processing a B&W H-alpha image of M16. I ended up using even *less* aggressive settings than Juan does in the tutorial, and I was able to reduce noise pretty nicely.

- Marek

[RBA]

I missed the tutorial Marek just pointed out. Nice stuff yes.

I couldn't but smile when I saw this sentence from Juan:

Now this is a busy background (I read this sentence somewhere and nobody understood him --I did/do).

Don't know if Juan read that quote from someone else, but I recall having a rather pointless debate with Ron Wodaski on his group about "busy backgrounds", so reading that gave me a chuckle.
I no longer write in Ron's group by the way. It got too old way too quickly and got tired of the lecturing and shepherding mode that dominates in that group from the same usual suspects...

[RobF2]

I've learned a lot but have so far to go.......

Amen!!   

This thread is/has been fantastic.  I think it's to the detriment of PI that this sort of detailed info isn't readily available from existing sources (apologies Harry if you have covered and I've missed it all).  I'm embarrased say I hadn't really thought to manipulate my source image much more than cloning and hard HT before Starmask, and certainly hadn't thought about using ATWT (or anything else really) on a mask afterwards.  My eyes open just a little more   

[sreilly]

While at lunch the other day with my wife I was explaining some of the things about masks that I did know, or at least think I know, when it occurred to me that what I had seen in Harry's video masksat.wmv is maybe similar or the same thing as using the "to lightness" or "lightness mask"  option in the HDRMultiscaleTransform process. What Harry's video shows is extracting the luminance from a color image and then using that luminance data for a luminance mask. Is this correct or totally different. The option "to lightness" says that HDRMT will only be applied to the CIE L* component of color while "lightness mask" says it protects the dark background regions. I'm not sure I understand the difference unless it's only that the "to lightness" option is only used on a color image while "lightness mask" can be used on either but it sounds as if they do the same thing. Is this correct?

It would be wonderful if there was a live preview window for creating masks so you could see in real time what changing settings creating the mask will result in much like the real time preview in the histogram transformation tool. That would certainly go a long way in seeing the effects of setting changes while creating the mask.

[RBA]

Not the same thing. A lightness mask is well, a mask built from the lightness (extract lightness and use it as a mask - in the case of a mono image, that'd be a duplicate of the image). "To lightness" means that the process - with or without mask - will be applied to the CIE*L component only, that is, if it's a color image, it'll extract the CIE*L component and apply the process to THAT component only, then it'll rebuild the image adding the untouched 'a' and 'b' (color) components to the modified L. In a mono image, "to lightness" doesn't do anything and it doesn't need to be used.

Don't know if this makes it any clearer, but pretty much each option does what it says.

[RBA]

As for "live previews" of masks, well, if you have an image that you're going to use as a mask, and you're modifying it (the mask), depending on what you're doing, you may be able to see the changes in a live preview window (LPW for short from now) -- for example, if you're modifying the mask with the HistogramTransform tool, just open the LPW... What you can NOT see in real time is how those changes will affect the image being masked, you need to apply the changes to the mask to see the results in the masked image - and then apply whatever you want to apply to the image of course, otherwise you'd only modified the mask, which shouldn't change the way the masked image looks like.

[Juan Conejero]

Now this is a busy background ... Don't know if Juan read that quote ...

I read a lot, Rogelio ...

[RBA]

Now this is a busy background ... Don't know if Juan read that quote ...

I read a lot, Rogelio ...

Right