WBPP 2.5.0 Released

Status
Not open for further replies.
One example is what you've encountered: LPS+LPD is applied to gray images; if they are RGB images, the result may not be correct since the method is designed for gray images,
Naming is another source of confusion your master files will be named gray while they are not.
The required space computation will be wrong since we compute the size of one channel for each image instead of three.

These are the first that comes to my mind.

In general, adding debayered images works if you mark them as gray; this is an old "trick" to use WBPP with debayered images. Of course, we should be aware that something may be wrong and we cannot benefit from the separated channel feature.

By the way, why don't you calibrate/debayer images directly in WBPP? this would remove any problem.
 
robyx said:
One example is what you've encountered: LPS+LPD is applied to gray images; if they are RGB images, the result may not be correct since the method is designed for gray images,
Naming is another source of confusion your master files will be named gray while they are not.
The required space computation will be wrong since we compute the size of one channel for each image instead of three.

These are the first that comes to my mind.
Click to expand...
Interesting. I never knew this! But as you say, it does explain why the new LPS+LPD process may have caused problems.
But I am not seeing naming problems. The integrated results are always called Color and RGB

robyx said:
By the way, why don't you calibrate/debayer images directly in WBPP? this would remove any problem.
Click to expand...
If I am shooting with my own OSC camera (ASI2600) I always let WBPP do everything (including debayer). But often I get images from remote telescopes that are already debayered.

In fact, I have just completed another run with some new colour data (already debayered). I even left the LPS+LPD option checked. And the result looks pretty nice! None of the weird banding I saw previously. No naming problems. So maybe your concerns aren't necessarily a problem? (let me know if you'd like to see the log as I realise I am testing outside of normal expectations here!)
 

Attachments

  • Capture.JPG
    Capture.JPG
    579.8 KB · Views: 121
StuartT said:
In fact, I have just completed another run with some new colour data (already debayered). I even left the LPS+LPD option checked. And the result looks pretty nice! None of the weird banding I saw previously. No naming problems. So maybe your concerns aren't necessarily a problem? (let me know if you'd like to see the log as I realise I am testing outside of normal expectations here!)
Click to expand...
Nice! the naming confusion is about the final "_mono.xisf" which denotes the color space. But it's not a tragedy if you know where it comes from!
 
Minusman said:
Does that mean LDD+LPS doesn't work with CFA images and is therefore skipped?
Click to expand...
No and yes.

It seems to work on some of my colour images (it did a nice job on this one) as long as they are already debayered. But if you are running it on un-debayered OSC data the process is skipped
 
OK, that's a shame. So only the canon banding script remains afterwards. Thanks for the info that some light has come into the many gray.
 
robyx said:
Please check the documentation https://pixinsight.com/tutorials/LDD-LPS/

Applying LDD+LPS to RGB images is nonsense since it means that we work on images where debayer has already interpolated pixel data, so we completely miss the opportunity to generate a proper master integrated reference frame.
Click to expand...
Yes, I suspect I was just lucky on my image. I won't be using LDD+LPS again on colour images. I'm going to stick to doing things properly ;-)
 
Status
Not open for further replies.
Back
Top