›

Channel Combination & SPCC Issue

Bruce Donzanti
edited March 2023 in PixInsight
I have been using SPCC for a while with no problems. I re-ran data that I have successfully used SPCC in the past but it is now failing. I am getting the error seen in the attachment.

Error: Insufficient data: only 3 sample(s) are available; at least 5 are required.

I don't understand how it worked before but not now and why this particular error. I have plenty of data. I repeated it 3 times today with the same failure. Any ideas what I am doing wrong now?

I also just noticed that if I try to stretch the RGB combined image, the 3 channels are not combined. There are completely separate RGB peaks on the histogram. I've never seen this before.

NOTE: I re-ran a dataset from an object I did a week ago and SPCC did fine.  So, something happened to this dataset.  I started from scratch with WBPP but I got the same failed result. 

Comments

  • BlockHeadBlockHead
    Wait.. you cannot do SPCC on a stretched image. The image must be linear. 
    Can you please confirm you have not made this error?

    Can you please make the image available? I just had a person make another kind of error that is similar.

    Thanks,
    -the Blockhead
  • Bruce Donzanti
    The data are linear.  I thought the same thing at first but the RGB images are not stretched.  
  • Bruce Donzanti
    Here is a screen shot of the 3 linear RGB images (I did a STF auto-stretch of the R image) and the STF autostretch for channel combination.
    Screenshot (75).png
    1920 x 1080 - 2M
  • Bruce Donzanti
    One more screen shot to show that Image Solver works but SPCC gives the error again.  Doesn't make sense as I ran another old data set and it works fine.  I then re-uploaded my original M104 data and re-did WBPP but got the same error.  What is even weirder is that I ran these same data 2 days ago and it worked. 
    Screenshot (76).png
    1920 x 1080 - 1M
  • Bruce Donzanti
    Finally, here is what the RGB image looks like when I begin to stretch it after channel combination and the SPCC failure.  The channels never combine- there is a clear separation.  I have never seen this.
    Screenshot (77).png
    1920 x 1080 - 325K
  • BlockHeadBlockHead
    Please make the indivdual RGB images you are using available to download.
    -the Blockhead
  • Bruce Donzanti
    edited March 2023
    Here are G drive links to download a fit file of each RGB image


  • BlockHeadBlockHead
    edited March 2023
    Ok... I found the answer to your issue.
    The problem is your stars are a bit fuzzy... PixInsight has to use the PSF method of detecting stars... and it doesn't detect many by default with your data. If you set the SNR detection minimum to "0" for your image- you will detect the stars and they will match with the few in the catalog from this field. 

    This results in SPCC working for your image.

    May I make a small video that illustrates this point?

    -the Blockhead
  • Bruce Donzanti
    edited March 2023
    Thanks so much, Adam.  I can make that adjustment but I think such a short video would be very helpful for me and others on a couple levels.

    1- to point out this error, which I've never seen since using SPCC, and I have been using it (albeit not that often) since it became available.  I could be wrong, but I would think most folks would not have identified how to resolve this. 

    2- BB imaging in my LP Bortle 7, high humidity sky is not a fun thing, so I rarely do it- that is why I usually do narrowband with great success.  But my point here, is that a lot of folks have an environment like mine and may experience this same issue, especially with few stars in the FOV. 

    Also, trying a setup with .28 arcsec/pk on bad nights of below average seeing and transparency is not a smart idea.
  • Bruce Donzanti
    Adam

    When I tried the fix you told me it worked immediately. Thanks.  I also just saw your short video on: SPCC Not Enough Samples Error.  Nicely done.  I think what would have helped me immensely to understand the problem would have been how the error was worded.  Your suggested wording would have rung a bell immediately as to the issue.  

    Bruce
Sign In or Register to comment.