Warning: No correlation between the master dark and target frames - Channel 0

Adam - One other person on this forum is seeing black-point issues in 1.8.8-9 along with a local astro buddy who stepped back to 1.8.8-8.
 
First things first. You need to cross WBPP off of the list of issues. The data that is being outputted is perfectly fine.
The issue is now with the way you are using BN and ColorCalib and/or PCC.

Unfortunately I cannot process your full data set- but I am confident that your integrated image (before you do anything else) is NOT black clipped.

So if you want to continue. My statement above needs to be verified. Do you agree that your "master light" is not black clipped?

Then we can proceed with your other issues.

-adam
 
Good... now we are on the same page.
So, do you agree that this is NOT a WBPP problem?

If so, we can move on to the next step you did. I assume this is DBE?
If not... then maybe you just did BN and Color Calibration.

Whatever the next step was is where we continue....

-adam
 
I can't prove that this is a WBPP 2.2 issue. When these steps are done below the histogram curve is pushed to the far left and almost out of view and the image can't be corrected. There have been PI posts on this issue. This has never been the case in the 7+ years I have been using PI...! I can provide screen shots it you like.
  • WBPP or BPP
  • Image Integration
  • STF
  • Automatic Background Extractor -OR- Dynamic Background Extractor
  • Any noise processing or or anything else
  • Histogram Transformation -
  • ...etc
 
I may have found the issue. I created a list of process icons that I have carried forward for 3 PI releases. This icon is on my PC desktop and I click on it to start PI. This loads about a dozen icons on my PI desktop. If I load a 'Process' from the PI header and not select one of these process icons on the desktop the Histogram Transformation works. I need to do more testing later and will get back to you.
 
It sounds like something you have saved in your Histogram Transformation process icons isn't valid for the particular data you are working on.
It would be surprising to me to codify any data settings in HT since it is entirely dependent on the input data (i.e. every image's histogram is different).

But I am not certain what you are doing..the above is just a guess.


-adam
 
I have not modified any setting in these Icons. They are on the desktop for quick use only. None of these issues occur in 1.8.8-7.
 
as adam just said, HT is really not something that can ever be blindly applied to a given image... that's probably your problem. if you want to do a quick stretch of a new image, it is best to use STF and then transfer that over to HT and apply HT to an image.

if it worked in -7 it was an accident, or you only think it worked consistently, because really, the HT settings for a proper black point setting are never the same fro image to image 99.999% of the time.

rob
 
So as I mentioned all of this worked in 1.8.8-8 and -9 until a series of patches were pushed down. I have been following the same process for 7+ years going back with each PI release. I don't 'blindly' apply HT. I turn on Track View, select the image and drag the STF triangle onto HT. The resultant plot is either pushed far to the left -or- when you adjust the sliders to set a reasonable black-point you destroy half your data. I have 1.8.8-7, -8 and -9 on VM's on my server and I can jump from one to the other and test things. Perhaps its best if I fall back to -7 for now...
 
What you are describing doesn't make sense with my experience.
A sure fire method (you know I am going to suggest this!) is make a screen recording (a video) and show your work!

You can also do the slightly more tedious thing and create screen shots that show your settings.

How can we ever figure things like this out? Falling back on an older version in the case of what you are describing just doesn't seem like a good or advisable approach to problem solving! There is a reason for what you are observing- and the information/answer is in the data and displayed in the processes in some form.

For example- I demonstrated that your data is not black clipped. With the readout preview you can see the value of the sky. You can adjust the STF to be something less than this value so the image does not appear black clipped using the leftmost adjuster.
Now when you drag your STF to the HistogramTransformation process the value that you chose will be in the Shadows field of HT. If it is NOT, *then* there is a problem. Then when you apply HT to the image *and reset your STF* your image should appear exactly as it was before you permanently stretched it which should not be black clipped.

I have just outlined how you can prove there is a real problem. You can do this quickly in 2 minutes. You can circle in red where I am wrong... Where I am an idiot and you are right!

-adam
 
so upthread you implied that you have a saved HT icon which is what adam and i are objecting to. if you are using STF+HT every time then it should work fine. one thing though is that trying to adjust the transformation in STF is quite difficult so if you are doing anything more than AutoSTF with STF you might want to use the HT tool directly to manipulate the stretch.

given that you're apparently doing it right, i think a screen recording is the only way forward, so someone can see if they can spot what's going wrong.
 
I am using Auto STF then later the HT as described above.. Next time I will do a screen recording when I am trying to address an issue or PI bug.
Unfortunately I am getting ready to go out of town to work on a remote joint observatory and I don't have time to do a recording.
 
You must log in or register to reply here.
Back
Top