I have just finished reprocessing the image. I think it is now a more accurate colour rendition in terms of the relativities between Ha and OIII. Below you see small versions of before (image 1) and after (image 2) the reprocess. Sharp transitions are removed and more nuanced Ha is present - particularly in the lower left quadrant. The original PBase image has been replaced.
Combining OIII and Ha with RGB is quite complicated!
Last edited by marc4darkskies; 16-03-2016 at 10:46 AM.
Yeah technically better, as you say, better transitions between the Ha and OIII and rings around stars are fixed, a good old fashioned Sidonioing there
Yeah technically better, as you say, better transitions between the Ha and OIII and rings around stars are fixed, a good old fashioned Sidonioing there
Mike
Ta Mike.
Quote:
Originally Posted by gregbradley
That looks great Marcus. A nice repro of what was really only a minor point. But still at this level its those little things that bug us.
Greg.
Thanks Greg. I did the repro because I realized I didn't do the NB layering right, leading to harsh transitions and missing Ha. I let the OIII overwhelm the Ha in the initial version. Worth noting that both signals broadly occupy much the same space. Anyway, this is my first go at using OIII as well as Ha in an RGB image ... always learning!
Either image looks ok to me Marcus. I prefer the saturation levels of the first image but the smoother transitions of the second. Maybe some more saturation might make the second image pop just a little more. Still very nice images.
your latest version looks the best for me! Very well done, i like that object!
Markus
Thanks very much Markus!
Quote:
Originally Posted by Paul Haese
Either image looks ok to me Marcus. I prefer the saturation levels of the first image but the smoother transitions of the second. Maybe some more saturation might make the second image pop just a little more. Still very nice images.
Cheers Paul. I boosted the PBase version a little earlier this morning with a simple curve to give it more pop but the saturation is the same as before. The sharp transitions were a symptom of less than perfect layering of Ha and OIII and the consequential loss of the fainter Ha information in the first image. Essentially the relativities of OIII and HA in the first version were not representative of the raw data. The OIII in first image also had a slightly too blue look. The second image presents a more nuanced and accurate rendition of colour IMO (while resisting the temptation to go fluro!! ).
I have also found blending Ha that you can lose some fine detail. My recent SH2-308 bubble image had lots of fine tendrils in the O111 data that mostly were lost in the blend.
I have also found blending Ha that you can lose some fine detail. My recent SH2-308 bubble image had lots of fine tendrils in the O111 data that mostly were lost in the blend.
Anyone with suggestions to combat that?
Greg.
I don't know how you process but I can't see how you can lose detail blending Ha or OIII unless the seeing wasn't so good when you gathered the NB data. I had less than average seeing with this image for my Ha and OIII so they didn't significantly add to the fine detail already present in the Lum. Having said that you have to make sure you have a combined LHaOIII (as luminance) such that it represents all three channel's signals well and to which you have applied all your noise reduction and sharpening. I would usually process each separately (L, Ha and OIII) and then layer them for a composite "Super Lum" layer that you will subsequently apply to the colour. However you build the RGBHaOIII colour layer (which is where my problem was with the first version of Thor's), this Super Lum should contain all the detail you need. Hope that makes sense!
I have just finished reprocessing the image. I think it is now a more accurate colour rendition in terms of the relativities between Ha and OIII. Below you see small versions of before (image 1) and after (image 2) the reprocess. Sharp transitions are removed and more nuanced Ha is present - particularly in the lower left quadrant. The original PBase image has been replaced.
Combining OIII and Ha with RGB is quite complicated!
Yes Marcus,
more detail is showing through -
it's almost as if every image is so unique that you have to learn how to process it - without a standard workflow.
Hi Marcus,
the processing of images is so complicated that the first try
usually falls far short of the full potential that the data can reveal.
I often have 5 or 6 attempts before I'm satisfied.
Hi Marcus,
the processing of images is so complicated that the first try
usually falls far short of the full potential that the data can reveal.
I often have 5 or 6 attempts before I'm satisfied.
cheers
Allan
Okay, I understand. You're right, it IS complicated and anyone who claims to get it right first go is probably pulling your leg OR under processing their data! The fact is, if you've been in the game a while you have a standard workflow but even then, one size doesn't fit all and each object requires it's own TLC and workflow tweaks. My standard workflow broke with the addition of OIII. All is right with the world now though!
I have just finished reprocessing the image. I think it is now a more accurate colour rendition in terms of the relativities between Ha and OIII. Below you see small versions of before (image 1) and after (image 2) the reprocess. Sharp transitions are removed and more nuanced Ha is present - particularly in the lower left quadrant. The original PBase image has been replaced.
Combining OIII and Ha with RGB is quite complicated!
Love them both Marcus but image 2 is probably marginally ahead.
Just had another gander at this Marcus and while posting full res always causes me to shrink it in my browser for a good overall look.... I did spend a bit of time surfing the full res, very much fun
).
for a good overall look.... I did spend a bit of time surfing the full res, very much fun 
