Author Topic: Using ATrousWaveletTransform for Sharpening  (Read 3412 times)

Offline sreilly

  • PixInsight Padawan
  • ****
  • Posts: 791
    • Imaging at Dogwood Ridge Observatory
Using ATrousWaveletTransform for Sharpening
« on: 2014 May 30 18:24:58 »
Several times I've heard of someone using this process for sharpening an image but look as I may I can't find anything on using this process. I'd like to get some idea as to the use of this tool if anyone knows of some instructions to do so.

Thanks,
Steve
www.astral-imaging.com
AP1200
OGS 12.5" RC
Tak FSQ-106ED
ST10XME/CFW8/AO8
STL-11000M/FW8/AO-L
Pyxis 3" Rotator
Baader LRGBHa Filters
PixInsight/MaxIm/ACP/Registar/Mira AP/PS CS5

Offline Alejandro Tombolini

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1267
    • Próxima Sur
Re: Using ATrousWaveletTransform for Sharpening
« Reply #1 on: 2014 May 30 18:47:12 »
Hi Steve,
Try increasing Bias in different layers.
You can see some uses of this tool for sharpening in this processing example, also here, and here for chrominance as target. Another example here.

Saludos, Alejandro. 

Offline JoLo

  • Newcomer
  • Posts: 46
Re: Using ATrousWaveletTransform for Sharpening
« Reply #2 on: 2014 June 04 08:22:45 »
Although i have only been using PI for a couple of months, I think I have sharpening with MMT (basically same process for ATWT) down pretty good, at least for my needs.  I use MMT instead of ATWT because of the ringing issue - I have used MMT for layers 1 and 2 and ATWT for 3 and 4 on some images, as I have read that ATWT does better with the larger scales, although I have noticed little difference with using MMT for all.

Good masks are critical to this process (and to noise reduction for the same reasons) as I like to apply stronger sharpening to the high SNR areas of the DSO, which is typically too much for stars and dimmer areas of the DSO.  I generally use a luminance mask and use HT to crank up the brightness and crank down the background, usually 2-3 iterations in small steps.  I then make a copy of the mask and use CloneStamp to get rid of the stars on one, and the DSO on the other.

I apply the DSO mask and are usually around these bias settings for most images:

1:  0 - 1 (lot of noise in layer 1, need to be careful)
2:  1
3:  0.025 - 0.050
4:  0.025

3 and 4 are really strong and need to be toned down on galaxies and nebula. 

I then apply the luminance star mask and are usually around these settings:

1:  0 - 1
2:  1
3:  0.05
4:  0

Both 1 and 4 don't seem to have too much affect on the stars, but 2 and 3 will sharpen dimmer/brighter stars effectively.

For the DSO it is a dance between sharpening and noise reduction; I just keep playing with each until I achieve a balance that works for me.  Sometimes I have to reopen the mask to fine tune it, reapply and resharpen / NR till I am happy.  I use MMT for both sharpening and noise reduction, but always in separate steps.

Like I said, this works well for me, your results may vary.  As with everything PI, trial and error is the key.

Joe
Orion Atlas
AT106
AT65
Canon 450D
Atik 314L+

Orion 12" Dob