Celestia Forums

The first test version with sRGB-capable output is downloadable from https://github.com/CelestiaProject/Celestia/actions/runs/6538920137 . It requires two parameters set in celestia.cfg:

Please test. For those willing to build from sources please use `srgb-test-v1` branch.

There is no file to download at the link you specified ... all I see are jobs, check marks, but no downloadable file (except workflow file which I don't need), and the latest version in Releases is 1.6.3, no EllipticalOrbit-HyperbolicOrbit version ...

(Sorry, I don't use Github much so I'm not very experienced with that ...)

You need a github account to be able to download (sorry).

Discord user NGTS addict uploaded it to Gdrive: https://drive.google.com/drive/folders/1KJKphu85i-4wQfHqr2tcI88eZ8vZ6Z9U?usp=sharing

Woah, amazing! The Earth (and maybe other planets too, I didn't test that) look way better than in older Celestia versions.



One annoying thing: It's very hard to unselect things since almost everytime I click in free space (black), it marks a star cluser ...
So I had to disable data/openclusters.dsc by renaming the .dsc to a non-Celestia extension and to remove it from celestia.cfg.

EarthMoon wrote:One annoying thing: It's very hard to unselect things since almost everytime I click in free space (black), it marks a star cluser ...

That's my fault, sorry! Fortunately for you its already obsoleted by the UCC, so i need to update it again.

I found an issue: There are noises on the day-night edges of planets when you zoom at the planets from big distance while viewing from the day side and only a very small black (night) part is visible. The planets have not beed modified in any way.

Neptune and Uranus (as example) as seen from Voyager 2 with high zoom-in:




Times and dates are visible in the upper-right corner. The Voyager addon can be found here.

Do you have 1.7 without sRGB changes? What do you have there?

EarthMoon wrote:There are noises on the day-night edges of planets when you zoom at the planets from big distance

yeah I've known about this for a while since the per-pixel rendering was introduced. I imagine it's probably got something to do with the amount of samples, so the only way to fix it would be to add more samples, but then that would take a big toll on performance... unless there's another way to fix it that doesn't involve adding more samples

Or add a setting for samples, so users can choose between performance and quality.
E.g. a Render > Samples in the menu bar which opens a text field where users can write a number (higher number = more samples)

The Github only has versions up to 1.6.3 in the releases tab so I can't download 1.7.0 ...

And building from source is a hard pain (at least with the tools I have) and would waste too much time.

And now I found a new bug: The orbit/rotation data for the moon get invalid at the year change 3999-4000. If you are in sync orbit on the moon and speed up time fast, you will see the Earth visually follow some lissajous paths on the moon's sky. At the year change 3999-4000, the Earth abruptly moves off this path over months and years. And the moon's trajectory follows a slight jump at that time (see below).


Open the below cel:// link, set the time scale to 100000 and unpause. You should then see a slight jump in the moon's orbit on its backside.
cel://SyncOrbit/Sol:Earth:Moon/4000-01-01T02:20:16.03711Z?x=9MQ6ZP2v4P///////////w&y=MJJKnS2/Mg&z=HXbQWVyp6////////////w&ow=0.44150567&ox=0.2469642&oy=0.7748661&oz=0.37903035&fov=45.883354&ts=100000&ltd=0&p=1&rf=71161779&nrf=255&lm=6&tsrc=0&ver=3

1. Use proper topics.
2. See VSOP87 limits.