New Tutorial Video "Linear Combine of Ha-OIII-L-RGB" Cirrus example

[oldwexi]

Hi !
My PixInsight Tutorial Video 
"Linear Ha_OIII_L_RGB Combine" is available now in english also.

http://dl.dropbox.com/u/57910417/Video03_Kombination_OIIIHaLRGB_END_english.wmv
The duration of this video is 29 minutes.
It shows the combination of separate linear
OIII- Ha- L- RGB- images into a linear narrowband RGB Color image
using the Ha-RGBCombination script  for Ha and OIII merging.

Also some stretching methods to get from the linear result to the stretched final image
are in this video.
Hope it helps and gives some ideas
Aloha
Gerald

[sreilly]

Thanks, downloaded and will watch later. Always good to have these as reference.

Thanks Again,

Steve

[oldwexi]

Hi,
i have made available the english version of the Video Tutorial
"Linear Ha_OIII_L_RGB Combine"
 NOW also with a Table of content.
The Zipped File is downloadable from:
http://dl.dropbox.com/u/57910417/Video03_Kombination_OIIIHaLRGB_END_english_TOC.zip

After unzipping the directory you can start the video by clicking on
....TOC.player.html
Clicking on the different lines of the TOC allows direct access in the video to the themes you want.

Thats how the TOC will look like and whats selectable:

Introduction
 2. Overview about used Images
 3. Preprocessing
 4. LinearFit of all Channels
 5. RGB Combine
 6. Combine L into RGB
 7. Ha-LRGB Combine
 8. Merge OIII into Greenchannel of HaLRGB
 9. Merge OIII into Bluechannel of HaLRGB
10. Stretch of OIII-Ha-LRGB with MaskedStretch
11. Stretch with MidtoneTransfer
12. Saturate Stars
13. Show Result

Aloha
Gerald

[ugatza]

First, thanks for your altruistic job, Gerald. Two questions:

1.- Is there any reason to select the green channel as the standard in the "Linear fit" process? Do the same role red or blue?

2.- In your case, you combine OIII images for the green and blue, but if I have SII images too, what will be the process, step to step, to combine OIII for the green and SII for the blue?

In advance, thank you very much.

[Jules]

Hi Gerald

A couple of questions about your video:

1. When applying Linear fit to various images you want to combine, is that the same as colour calibration/back ground neutralisation once you have combined?

2. Why combine RGB then convert to Lab?

Great video!

Thanks

Julian

[Geoff]

Really good video Gerard.  I just watched it in its entirety and realise I've still got lots to learn in PI.  I liked your use of LineraFit to get the colour balance.  I have one minor comment--when you did the DBE, you had points everywhere except on the main part of the nebula.  If there were any faint background nebulosity, you would then flatten it by this application of DBE. RBA has a good discussion of this issue here http://blog.deepskycolors.com/archivo/2010/05/28/removing-gradients-while-preserving-ve.html
Thanks again--I've learned a lot from this video and the pixel math video.
Geoff

[oldwexi]

Hi Geoff!
Thanks for carefully watching this long video and thanks four your comments.
I fully agree that DBE can be used differently.

The video i have just in work is using the VdB150 which is full of faint molucular clouds -
so far i used similar samples in DBE as in Cirrus -
will have a try to use DBE  in VdB150 also in another way.
Thanks
Gerald

[ugatza]

And the preview Jules and mine´s questions? 

[oldwexi]

1.- Is there any reason to select the green channel as the standard in the "Linear fit" process? Do the same role red or blue?
2.- In your case, you combine OIII images for the green and blue, but if I have SII images too, what will be the process, step to step, to combine OIII for the green and SII for the blue?

Hi "Ugatza"!
1.) In the Cirrus case the green channel had the best data of the RGBs and no other channel was saturated 100% so my idea was therefore to use it as reference.
     In the meantime i  had cases where it was better to use the dimmest of the RGB channels and linear fit the other channels down not to loose dynamic.
     So if the median of the blue channel,  which is usually the  lowest, is low i select the blue a s reference.
2.) In the Cirrus case i tried to produce as final product a RGB image which got Ha and OIII improvements. The intent was not to create a false color image.
     If you want to put the the SII into the Blue channel the only difference to the steps in my video is - in HaRGBCombination you select the SII  instead of the OIII
     in the third merge. (first was Ha into red, second was OIII into green, third could be SII instread of OIII into Blue).
     this should keep the star colors nicely however i am not sure that your SII channel is so strong that you really will see it beside the OIII?

Personally  i recognized that i learn best when i do a presentation about that what i want to learn more deeply. Making a usable video brings me under pressure to learn more and deeper details...
And, after its examined by others i get feedback on cases were i was not deep enough. Its not altruistic is simply the will to learn PI better and better and  finding a way
to get deeper into the details by producing videos... For me its a winwin.
Also  after many many years of fiddling around with 5 different programs using 3 different fileformats i became grateful
when i found PI which is in my opinion the complete solution for image processing.

Gerald

[oldwexi]

A couple of questions about your video:
1. When applying Linear fit to various images you want to combine, is that the same as colour calibration/back ground neutralisation once you have combined?
2. Why combine RGB then convert to Lab?
Julian

Hi julian!
1.)In that case the linear fit was enough to skip colour calibration/background neutralisation, because the single channels were already nicely balanced.
    If the color balance would not be good enough after RGB Combine i would insert a colour calibration step.
2.)Because the idea was to seperate the chrominance from the Luminance in linear mode.
    I needed the RGB to extract the chrominace a and b. Than the seperate Luminance exposure sum  which was less noisy was merged together with the a und b.
    Using Lab extract/combine simply i exchange the Luminances and keep the chrominances untouched.

The base idea of this kind of processing was also to show that you can stay in linear mode very long during processing.
What in my opinion is a great advantage because it keeps the dynamic. My personal opinions is, the first stretch brings more contrast but also erases data.
Also my opinion is, you will have difficulties with your colours if you merge stretched RGB with stretched narrowband.

Gerald

[ugatza]

Thank you very much, Gerald, for your reply. I´ll test different combinations.