Weighted Batch Processing Script

W

WB91

Well-known member
Guys,

Is it possible to add an output save type option for the calibrated, registered and final integration images as they all seem to be defaulted to 32bit xisf? I would like to save all my registered light frames as 16bit TIF or FIT and also have the option to save the final integrated image to either 16 or 32bit at the end of the process. Saving all registered light frames with a 20 or 50mb camera make for massive file storage problems.

Also, when using dark flats instead of bias, it would be nice to see a dedicated feature in here for this. I understand you can add darks flats to the darks and the script recognizes this but I am not sure what other settings need to be set to make this 100% fool proof. DSS and APP have this feature.

Last, is it also possible to have the option added when using a OCS camera to align all channels during the registration process to help reduce chromatic aberrations? This seems to work well in DSS, not so much in APP from my testing. To do this before final immigration seems like it would work the best, even though the processing time would be longer, but the image quality should be superb I would think.

Thanks,
Bill
 
for a while now the various processes underneath WBPP have been using XISF as the output file format. this is because a lot of the advanced features of PI now depend on the features of XISF files. there is probably no way to get WBPP to switch to FITS.

but saving linear images as 16-bit anything is not a good idea. there's not enough dynamic range possible in a 16-bit file to capture the levels in a linear image properly. 16-bit images are really only good for stretched images where the histogram spans a large portion of the 'number space' of a 16-bit file.

i hear you about the massive disk requirements but switching to 16-bit is not a good solution here. one could turn on XISF compression by default but there is obviously a performance tradeoff there, plus it could lead to out-of-memory issues especially with these huge image.

on the other two i can't comment. WBPP was actually forked off of BPP and both were originally written by PI users. Juan does maintain them as well, but i think @robyx (who worked on WBPP) does still add features from time to time.

rob
 
Last, is it also possible to have the option added when using a OCS camera to align all channels during the registration process to help reduce chromatic aberrations? This seems to work well in DSS, not so much in APP from my testing. To do this before final immigration seems like it would work the best, even though the processing time would be longer, but the image quality should be superb I would think.
Click to expand...

You can already do this with WBPP. Just enable distortion corrections for image registration, and include all frames (for all filters) in the same preprocessing operation.

However, in these cases I strongly recommend using WBPP only for calibration (and demosaicing when applicable) by checking the Calibrate only option, then use StarAlignment with the thin plate splines model and distortion corrections enabled to register the entire data set with respect to a single reference image. Effective distortion correction requires good star profiles (excellent focus and tracking).

As for the other questions, I agree with everything that Rob has said. If you really need to use other inferior implementations requiring FITS or TIFF to work, use the BatchFormatConversion script.
 
You must log in or register to reply here.
Back
Top