FITS Format

[MRPUGH]

Chris

Why remove the ability for users to select an output format?  That is all I want because I do not use PI for image calibration as there are, in my opinion, better tools out there for that particular process. So having calibrated my images I bring them into PI for alignment and move them back into my chosen software for noise rejection and image combining (again into a package that is much better in my opinion for this), because I choose not to use PI.  So the point here is simple - the PI developers are trying to tie people to their software that uses an image format unrecognisable to any other software package.  OK, so maybe in the future the format will be read by other packages but given this is the first formal release of the format it is a bit naive and immature to assume that everyone wants to stay inside PI....just not so in fact I can tell you that a LOT of people shy away from it because the GUI is intimidating, the process pages are user unfriendly and the pop-up help files utterly nonsensical.

I did find the Batch conversion process but for goodness sake - just write back in the ability for users to select an output format!

Martin

[Niall Saunders]

Hi Martin,

Given that you seem to have an insatiable urge to want to process intermediate XISF files generated by PixInsight, and that you need to be able to do this outwith what you apparently consider to be unnacceotable constraints of PixInsight, why do you not simply convert those XISF files that you are interested in back to FITS?

We have tried to explain 'why' certain Processes within PixInsight require the use of the XISF format - if the use of the XISF format leaves you so distraught, then simply do not use those processes !

You state that you bring your images into PixInsight "for alignment" - it seems to be the only reason that you are using PI. Yet you then state that "the PI developers are trying to tie people to their software", stating that PI "uses an image format unrecognisable to any other software package" as your only evidence.

I have already pointed out that PixInsight required a format that was more capable than FITS, but - at the same time - conckuded that the new format uses an image format unrecognisable to any other software package had to be made transparent right from the very outset (and so the XISF specification was released as early in the development phase as it was possible to do so).

I also suggested that, if other software packages cannot accept any format as an alternative to the ancient FITS format - that is their problem, not something that PixInsight needs to concern themselves with.

I assume that, at the same time that you have come on-line here to rip the XISF format and the Totalitarian State that you beleive is the world of PI, you have also visited the thriving on-line user group(s) for the other software that you work with, to ask them why they have not made use of the freely available XISF format and documentation.

If so - what was their response?

[What, you couldn't find a thriving on-line discussion forum, or you did, but nobody replied? Hmmm, I am at a loss as to suggest how you move forward]

Finally - comments such as "Get over yourself" from you are hardly likely to endear you to any user group. You asked a perfectly valid question, and have received pretty comprehensive replies - none of which involved that kind of attitude. Let's all try to keep this discussion 'civil'

[sreilly]

I strongly suspect that Martin is referring to using layers in PS and that's still very possible if that's the route you desire by simply taking the stacked or aligned images and converting them to .fits. It's simple and easy, much like in the days before Eddie wrote the fits plug-in for PS. It's not a big deal nor does it require attitude towards those attempting to help. While Martin is a pro with Photoshop he wasn't always and if he put half as much effort into PI he may well find himself doing far more in PI and liking the results. And forget about the documentation comment as there was none for PS as far as astronomical image processing went in Photoshop before others started to document how they processed their images. In fact I doubt you'll find any in their manuals to this days for our processing needs for such data. And they don't directly support .fits yet, at least not up to the versions prior to their Cloud version as far as I know. Different strokes for different folks as the saying goes.

Yes, I did a 360 after I realized that they was a reason and need for the new extension. If the format allows me to use the newer processes then why wouldn't you? As noted before, you can go into Global Preferences and set the "Save As" format to .fits.

[Juan Conejero]

Hi Martin,

just write back in the ability for users to select an output format!

I'm afraid that won't happen, neither for StarAlignment (which apparently is the only part of PixInsight you are using... overkill, maybe?), nor for the rest of tools and scripts that can generate disk image files. The problem is not that we don't want you to use a particular format—of course, you are free to use whatever format you want to store your data, how could it be otherwise?—, but that we cannot rely on FITS to support (in any reasonable way) new features and functionalities that we are implementing and will be implementing in the future. If you want to use PixInsight (you obviously don't, but let me say this rhetorically), better forget FITS as soon as possible because it has been deprecated in our platform, and all internal dependencies on FITS metadata will be removed during the 1.8.5 cycle. FITS will only be used to feed raw data at the very beginning of the whole process.

So the point here is simple - the PI developers are trying to tie people to their software that uses an image format unrecognisable to any other software package.

That's not true. You can always use FILE > Save As and select the FITS format to save an image. Or use the BatchFormatConversion script and set output extension = .fit. Or use the ImageContainer tool in a similar way. There are even third-party applications that can read XISF files (on macOS, you can use Code Obsession's Observatory and Acorn; see this forum post). Of course, you may lose some extended features by doing this conversion, since FITS cannot store image metadata and properties that our tools serialize in XISF format. This will happen more and more as we continue our development, but that's your choice.

As for other software packages, as Niall has said, that's not our problem. We can read and write FITS, TIFF, JPEG, JPEG2000, WebP, PNG, BMP, XPM, XBM, etc., and read-only even more formats. So format wise, I think we have no relevant problems.

[pfile]

martin, there is a hint for you in this thread, if you are handy with an editor and happen to have some old process icons laying around.

https://pixinsight.com/forum/index.php?topic=11340.0

rob

[mvln]

Please give us back the option to save calibrated / registered files as FITS.
Other tools do not support the new format.

[Niall Saunders]

You would have to very carefully explain, in detail, why you can no longer proceed with your processing if you are unable  to save your specific files in FITS format. You would also need to specify, in detail, exactly which images you need to be able to save in FITS format, and when in your processing methodology you need to be able to save these in FITS format.

That way, we might be able to help you understand 'why' the use of FITS is just no longer required, and - perhaps more importantly - why the contiinued use of the FITS format is just simply no longer tenable.

The XISF format was not created 'just because PixInsight wanted to use a propietary format', it was created out of absolute necessity. Without the use of the XISF format, and constrained by the now archaic FITS format, many of the new processes in PixInsight would simply not be possible - period.

An analogy here might be, "Do you want to save these files of yours in JPG format?". Almost certainly not! Why? Because you would not accept the loss of information that would occur if 'forced' to use such a format - despite it being perfectly 'portable'. Well, you will soon see that, as PixInsight continues to develop, the FITS format will become les and less usable. As Juan states elsewhere on this thread, the FITS format will become 'deprecated' in PixInsight (remembering that 'deprecated' does not mean 'dropeed' or 'no longer supported' - instead its meaning becomes closer to 'no longer needed') and users just won't think about FITS once they have loaded their data.

Personally, I have not saved a single FITS image since the initial release of XISF.


However, it is only certain key stages of certain key processes that mandate the use of the XISF format. The generation of FITS files at these interim stages would serve no real purpose whatsoever - after all, as already explained, PixInsight would not be able to proceed without the critical extra data stored in the XISF headers, and any third-party programs (that cannot read XISF data ***) would have absolutely no need to be able to read those interim XISF files in the first place.

*** This lack of third-party support is not a problem that PixInsight has created. If third-party software developers choose to fail to implement a royalty-free, open-source, fully-documented, well-established image format such as XISF, then that is their problem. This was always going to be the case, and is why the full XISF implementation and documentation, as well as all source code, was released free-of-charge, long before PixInsight reached this current stage in development, where the use of XISF has now had to become mandatory. These third-party vendors were free to join in on the development discussions, and their input would have served to improve XISF as we now know it. However, when third-party developers have an extremely 'lethargic' attitude to developing their product to meet the growing demands of their existing customers, or if they charge scandalous amounts of money for their customers to enjoy the privilege of having bugs removed from the existing code release, then it has to be the customer that makes the necessary decisions as to how best to invest their money

But, that is a personal (non-affiliated) rant about 'external software' - if this is not the problem that you are facing then, as described above, tell us in plenty of detail about the issues that you are having right now, and we can try and help you.

[HomerPepsi]

I'd just like to add a use case for fits for outputting imagecalibration files.  For analysis in MaxSelector, CCDInspector, etc...