Hi Dave,
The FIRST thing that you HAVE to keep in mind is that the image that you see on the screen in front of you (which will also be the source of any image that you ever choose to print out) is ONLY constructed from Red, Green and Blue data.
In other words it is an R-G-B image.
It is NEVER an -L- RGB image. Ever.
3 colour channels. 3 colour planes. 3 different phosphor dots on a CRT. 3 different liquid crystals behind 3 different colour filters on your LCD.
THREE
And LRGB is FOUR
So, where the h3ll does the 'fourth' channel go?
It has no other choice than to be 'blended' into the original three channels (Red, Green and Blue, naturally).
Same argument goes for incorporating an 'extra' Ha channel. Where the heck is the data going to go
The solution is to use the 'brightness' information in the Lu channel as extra 'detail' information, to 'brighten' the underlying RGB channels (and, for an axtra Ha channel, what often happens is that the Ha channel is used to brighten 'only' the Rd channel, thus 'enhancing' the Rd data with respect to the other channels).
But, you can only 'brighten' the data in a channel 'so much', before you start 'running out of headroom' - i.e. the data in that channel starts to 'saturate', and the extra detail that you hoped you were adding simply results in a car-crash.
So, instead, the Lu data is 'scaled back' during the LRGB combination process - to a point where the effort was still worthwhile, but before the car-crash point.
At the same time, what can also be done to help mitigate at the 'loss of headroom' when adding the Lu channel, is to 'scale-up' the RGB channels beforehand, then add the Lu channel, and then let PixInsights automatic rescaling feature get all the data back into the 0.0 - 1.0 range.
This RGB 'scaling-up' can either be done 'globally' as a 'Saturation' boost, or it can be done using the L, R, G and B sliders.
There is no particular 'right' or 'wrong' way. You may not have to touch the sliders or selector buttons at all - whether you do actually depends on the histogram distribution of your Lu data in the first place. If you have worked the Lu image very hard - to the point where it is a stunning image in its's own right (probably the 'typical' approach) then the Lu data may actually totally 'overpower' the RGB data, giving the image lots of detail, for sure, but at the expense of all the colour saturation you actually had in the first place.
This is where (as I understand things, I don't have 'binning' options on my $400 camera) the art of Unbinned L + Binned RGB comes into its own. You bin the RGB, to save exposure and acquisition time, but still maintain 'depth' of colour because the 'binned super-pixels' are 'super-sensitive'. So you maintain nice colour depth. And the detail that you no longer have is then subsequently replaced by the Lu , un-binned, high(er) resolution data - 'layered' on top of the low(er) resolution, well-saturated, colour data.
But, you have to be 'gentle' and 'skillful' when 'blending' in the Lu data again - you CANNOT just 'pour it on' (think children and wax crayons here - give an artist the same crayons, and a much nicer result should ensue - unless that 'artist' has the same skill as me
)
That is why it can help to either concentrate on Lu (ALONE) for a while, or RGB (ALONE) for a while - until you can 'control' that data. Then, you can try to massage detail out of your 'combined' images once you can see how the process is going to behave.
Juan does have a point in that if both your Lu and RGB data has the same spatial resolution (eg from Lu, Rd, Gn and Bu filters on a single camera), then there may actually only be a tiny advantage in bringing that Lu data 'into play' - at least as an Lu channel. That is why it can be 'more fun' blending in an Ha channel - the results are more obvious, and hence the effort put in makes the smiles bigger
Basically, LRGB is just NOT 'essential' - get the RGB side done, and done right, and you may not need the Lu data at all.
Cheers,
