Tutorial for LocalNormalization and narrowband data?

[joelshort]

I have found LocalNormalization to work wonderfully with LRGB data, basically with the default parameters.  But with narrowband data it comes out looking pretty horrible.  Is there a tutorial or similar somewhere for using LocalNormalization with narrowband data?

[-Amenophis-]

Hello,

How do you use the process?

Should images be corrected at the gradient level by DBE for example?

Should the reference image have a gradient indentation as well?

Thank you in advance for your answer.

[joelshort]

My process is as follows FOR LRGB images (not narrowband):
Calibrate lights with master dark, master flat and master bias (for CCD, no bias for CMOS)
CosmeticCorrection on calibrated lights
StarAlignment
SubframeSelector and Blink to choose a reference image
LocalNormalization
ImageIntegration to create master LRGB channels (using LocalNormalization data)
DynamicCrop
DBE and/or AutomaticBackgroundExtractor
Small amount of linear noise reduction with MultiscaleLinearTransform
Deconvolution on the LUM channel
LinearFit the RGB masters
ChannelCombination to create an RGB color image.

At this point is where I would stretch the images. 

As you can see, LocalNormalization takes place after choosing the best sub and before stacking.  This means that there is no background subtraction etc that is applied on the reference image before LocalNormalization or stacking.  However, I have found it to be beneficial to choose a reference image that does have a relatively even background.  So if there are two images that are similar, but one has slightly less SNR but the background is even across the image, I will choose the one with the even background vs the one with slightly higher SNR for the reference image.  But most of the time the image with the highest SNR and tightest stars is the reference image regardless of the background.

[-Amenophis-]

Thank you very much

I try LN in few days.