OIII/ SII stretching versus HA in nebulae. Art or science? Other · Tim Hawkes · ... · 4 · 460 · 0

Philosophical musings on image processing...

There are just so many ways to process exactly the same data set  that it is quite possible to produce a whole set of perfectly valid interpretations of the same data that end up almost unrecognizably different from each other on the basis of decisions made during processing.   The tools get ever more powerful and the options wiider.  We probably all just end up with our own predominant 'style'.    Just interested here in exploring how much we vary , the main drivers for decision making and whether there is any mainstream concensus at all?

For my own part ..decision making is probably about 70% science (to the level that I think that  I understand it)  and about 30 % artistic effect.   So for example ..

#  I like to use the OOH palette alongside RGB stars for many HII regions because  it is more 'naturalistic'   - but nevertheless acknowledge that this is still an artifact because the nebulae are, in reality, too diffuse for the colours to be seen.  I will stretch the OIII signal up for better effect and to make it obvious where it is --- but almost feel guilty taking this too far because it distorts  the depiction of the  relative ratios of HA and OIII light.  

#  I also like to use the SHO palette  particularly on some type II nebula.   Although certainly highly unfaithful in depicting the relative levels of NB emissions  this palette can provide almost 3 D information on the nebula shape with the O3 usually locating the hotter core region and the SII  some of the ionization edges  (nearest the viewer)  -  so there is real scientific value as well as artistry in the depiction

However how far to take the stretching up of SII and OIII?   It is easy - no decision to make - with some of the Wolf Rayet nebulae such as the Lion and the Crescent  because the energies involved make the OIII relatively bright.  But for some of the H II regions the ratio between the strength of the HA  versus OIII or SII emissions is really quite extreme --- so it takes relatively a lot of imaging and time to get sufficient SNR using these two filters to even get enough signal  to get much 'blue' and 'orange/ brown' into the image at all.  Good 'current season'  examples might be the Cave nebula and the Bubble nebula  where there are striking SHO  as well as many predominantly red  images .       

I can't help but think that with images seen side by side the totality  can all seem misleading --- i.e. the same often predominantly brown/ orange/ blue depiction of a vast range of objects that are really very different but that as we depict them might appear to be constituted almost the same.  

Obviously no right answers on any of this -- just interested to know what other folk think 

TimH

andrea tasselli:
The one thing I  don't get is: what is a type II nebula? F

You are quite right .  It was a 'typo'  :-)   .  I meant HII region of course.    Thanks for your trenchant  opinion.   Of course enforcement may be difficult :-)

Tim

Wei-Hao Wang:

Tim Hawkes:
#  I like to use the OOH palette alongside RGB stars for many HII regions because  it is more 'naturalistic'   - but nevertheless acknowledge that this is still an artifact because the nebulae are, in reality, too diffuse for the colours to be seen.  I will stretch the OIII signal up for better effect and to make it obvious where it is --- but almost feel guilty taking this too far because it distorts  the depiction of the  relative ratios of HA and OIII light.

This is science. The idea is to show the audience where OIII comes from and contrast it against H-alpha.  (Science never demands the OIII/Ha pixel brightness ratio in a picture to be identical to the true intensity ratio.  Never.)  And if it looks pretty, it's even better.  

So, don't worry about it.

(I have no comments on the SII part.)

I  have tried  equal total exposure time and gain for OIII and HA, applied an equal stretch, then adjusted to allow for the relative  spectral response of the camera at the two wavelengths and then adjusted again for the  relative response of the human eye.  It  doesn't look so spectacular-- just a fading of the dominant red of the HA to a  pink/ whitish colour in some cases. ...  uglier certainly but you could argue it is more realistic  and better defined.