Hello, I'm working on viewing RAW files within the FITS Viewer. They won't be as useful as FITS files, but it might be convenient for them to get displayed there (and also in Focus & Guide modules). So I need one of each type (CR2, NEF, ARW). It should be a star field or DSO that can be solved by astrometry.net. Please include the pixel dimensions and image angular scale (wxh)

No one has CR2/ARW files?!!!

I put one CR2 from 5D Mark III of M31 at s2.org/~jpaana/IMGL9192.CR2

400mm focal length, 5760x3840 pixels, 6.25um pixel size, solve-field --downsample 4 gives:

log-odds ratio 171.453 (2.89245e+74), 48 match, 0 conflict, 199 distractors, 81 index.
RA,Dec = (10.5038,41.678), pixel scale 3.3158 arcsec/pix.
Hit/miss: Hit/miss: -+++--+

---

+---++-+++

---

++++--++-++--+----+---+-++-+

---

++---++----+

---

+----+---+--+---++-
Field 1: solved with index index-4211.fits.
Field 1 solved: writing to file ./IMGL9192.solved to indicate this.
Field: IMGL9192.CR2
Field center: (RA,Dec) = (10.505016, 41.683592) deg.
Field center: (RA H:M:S, Dec D:M:S) = (00:42:01.204, +41:41:00.930).
Field size: 5.33195 x 3.56503 degrees
Field rotation angle: up is 171.291 degrees E of N
Creating new FITS file "./IMGL9192.new"...
Creating index object overlay plot...
Creating annotation plot...
Your field contains:
The star νAnd

Thanks! What's usually the time difference between downloading CR2 image vs. FITS?

Haven't timed, but it does make a substantial difference on my Raspberry Pi 3 as SD card IO is not the fastest. Also downloading over wifi to my PC is slower as the FITS is usually about 2x the CR2 size. Too bad the camera doesn't support subframing, that would make life a lot easier and is why I've moved pretty much over to using Atik 383L+ instead when not doing wide field.

Yeah, I'm considering options for approaching this issue. CR2/DNG/ARW cannot be subframed. It might be possible to open the file, convert to jpg, subframe, then upload, but that would take too much time as well. GPhoto driver already process all formats to FITS, and even if I wanted to display CR2 within FITSViewer, I'd have to use libraw to convert to JPG, then load it in FITSViewer and lose all the FITS header info that might be useful.

I'm thinking this: Use FITS for focus/guide/alignment and RAW for capture. What do you guys think?

I guess it would be possible to modify dcraw or libraw to support decoding only a part of the full raw image (maybe at MCU granularity or something), which might speed things up downstream. Also support for the smaller raw formats could be useful for some cases and kind of emulate binning modes (for example Canon's sRAW is 1/4 size of the full image, mRAW is a more complex downsampling) and should be enough for alignment and probably also for focusing and guiding. Even switching to JPEG might be acceptable for those.

This will be perfect for me if the selection FITS/ARW could be made automatically be ekos

You need ARW file again for your test?

With libraw, you can decode a RAW file and then copy its data (or part of it), but then you'd have to save the subframe back as the original format (CR2/NEF..etc) which I don't think it's possible with libraw and even adds more overhead. Currently, after libraw decodes the RAW images, it is subframed if necessary and saved in FITS. libraw frames are always 16bit which might explain why its size is overall larger as some cameras are 10, 12, or 14 bits so it went for 16 to cover all.

I dug a bit deeper and it seems lossless JPEG doesn't offer a handy boundary to do partial image decode like lossy JPEG does as MCU so that point is moot. Canon's CR2 is lossless JPEG encoded internally so the data is smaller compared to raw 14 bit (or 12 with some sensors) data, even though it does have JPEG previews and other metadata embedded.

Yes it's good to have it here in case I need it for testing in the future.

ok.

here is a ARW file of M45.

4256x2848
8.4 um pixel size
420mm focal length