Pinxinsight File Size

[groberts]

I have an as yet unresolved problem using Star Alignment on the thread below.  But whilst looking at this it's drawn my attention to the file sizes generated using PI.  In my case the LRGBHa 32 bit FITS stacks are each 250Mb - is this normal or a problem? 

Graham

 

    

 

[wornish]

Its normal they are big! 

Once I have finished the post processing I tend to delete all the intermediate files and just keep a high quality JPG of the final image.  I do keep all the original raw data files as well.

I you keep all the intermediate files you really do eat up huge amounts of storage.

[discardedastro]

I'd stick to saving files in XISF (PixInsight's default format). This format supports (lossless) compression which will help a bit. But stacked files will be large, especially if using 64-bit outputs - this is normal. I regularly rack up ~100GB processing folders, keeping all my intermediate data (a single frame is 80 megs or so, stacked image is around 120MB). Stacked RGB image is about 280MB. AP with modern sensors generates a ton of data!

Having said that, looks like you're stacking externally with something else and then bringing FITS in - may be worth learning PI's stacking tools, they're really top-notch and straightforward once you get the hang of it (BatchPreProcessor will do a very good job of making getting to images ready to stack for you). That'd probably also fix your other issue.

[groberts]

Thanks for the comments, at least it is 'normal' then - at this rate I'll need a computer the size of a house!

These were stacked with PI but will take a look at BPP and see if that improves matters - when stretched the stacks themselves look OK.

Graham  

[discardedastro]

6 hours ago, groberts said:

Thanks for the comments, at least it is 'normal' then - at this rate I'll need a computer the size of a house!

These were stacked with PI but will take a look at BPP and see if that improves matters - when stretched the stacks themselves look OK.

Graham  

If you're stacking with PI then definitely stick to using XISF as a file format.

And yes, fast storage is the best improvement you can make to most PCs to improve performance of PI - I have a 2TB Samsung 970 EVO Plus NVMe SSD for my active datasets and I can completely saturate even that (though there is a Ryzen 3950X to match - with half the cores, that disk is probably not going to bottleneck). But if you just hang onto your raw frames and your pre-stack intermediate data you'll be dealing with dozens of gigabytes per night of imaging quite readily, so a small fast-as-you-can-get SSD paired with a larger storage volume for older data is the way to go. I use a big 'ol pile of disks in the garage for storage of all my raw data.