LRGB with standard RGB and HD L-chan?

[dmcclain]

I'm finding that attempting to combine one set of RGB data with an HD stack of L-channel data, causes my colors to become strongly saturated. I also found that unless I first convert the RGB channels to double-precision floats before using LRGBCombination, I would end up with a bizarre palette of colors in the combined image.

Any helpful suggestions about this kind of processing?

[mmirot]

Linear or non linear images?

[vicent_peris]

I'm finding that attempting to combine one set of RGB data with an HD stack of L-channel data, causes my colors to become strongly saturated. I also found that unless I first convert the RGB channels to double-precision floats before using LRGBCombination, I would end up with a bizarre palette of colors in the combined image.

Any helpful suggestions about this kind of processing?

Hi,

Before attempting any LRGB composite you should first stretch the RGB and the L images.


Best regards,
Vicent.

[mmirot]

Vicent.
Is there a good way to  LRGB combine linear images?

 I tried to linear fit  the  L to the RGB but it did not work well.

Max

[dmcclain]

Right... I just stumbled onto the fact that images needed to be stretched, else the colors become very saturated and bizarre. However, now that I have stretched the individual channels and produced an LRGB composite, how do I color calibrate the image? Isn't it a bit too late in the nonlinear domain?

Perhaps that is what is intended by the channel scaling in LRGBCombination? And if so, how does one derive a proper set of scale parameters?

As it happens, with an LRGB dataset at hand, the colors come out nearly reasonable, but with a bit of green noise. Seems like a lucky outcome, but perhaps it was designed to be that way when using "matched" filter sets?

The above comments seem to apply only to HD data, not to standard precision. In the case of standard precision data sets, the LRGBCombination seems to produce very nice results that can then be color calibrated in the usual manner. Something unique about double- versus single-precision float data.
------
Hmm... interesting... I can perform normal color calibration with BGNeut and ColorCal against the LRGBCombo of stretched HD frames. It seems to work just fine.

[vicent_peris]

Hi,

I recommend you to watch the two video tutorials we have in this website. Some notes to the workflow:

- Color correction (DBE, BackgroundNeutralization or ColorCalibration) should be applied always while the images are linear.

- The LRGB composite should be assembled always after stretching both the color image and the L.

- In most cases, you'll get better results if you perform a dynamic range compression to both the color image and the L before attempting the LRGB composite. If you compress the dynamic range with HDRMT, you should apply it always after stretching.


Best regards,
Vicent.

[mmirot]

Vicent any room open for the advanced conference next week in Tucson ?

 I live near by and have some time off.

Max Mirot

[vicent_peris]

Yes, there's one seat. You're welcome.

V.

[mmirot]

Great.

Thanks

Max

[mmirot]

The PayPal button for next weeks conference is deactivated.   Should I use the this weeks button for payment and registration?

I plan to be there on Monday so save me a seat. 

I was at the first US conference at the Adler Planetarium in Chicago.  I look forward to meeting with you again.

[vicent_peris]

Hi Max,

Don't worry, pay with that button, it's the same price. I remember you, I'm happy to meet you again! :-)


Best regards,
Vicent.

[mmirot]

Great!