Page 1 of 1

Zooming on an exoplanet not possible?

Posted: 19.10.2007, 01:06
by ElChristou
Guys, noob question again... (lately I think my neurone must have some kind of problem!) I wanted to watch the travel of an exoplanet in front of it's star, but apparently I'm unable to zoom enough for this. What's up? We can zoom on the faintest galaxy and not on a star no so far from us? ... or it's just me...

Posted: 19.10.2007, 01:36
by Cham
I think there's a bug with the zoom function, at very high zoom values. This bug may be Mac related (I'm not sure of this, though).

Zooming to very high values (small angles) makes Celestia to slow down to a crawl.

Dirkpitt, can you confirm ?

Posted: 19.10.2007, 01:58
by Johaen
Zooming to very high values (small angles) makes Celestia to slow down to a crawl.

I cannot confirm this in Windows. Zooming in to max maintains 60fps.

I wanted to watch the travel of an exoplanet in front of it's star, but apparently I'm unable to zoom enough for this. What's up? We can zoom on the faintest galaxy and not on a star no so far from us? ... or it's just me...


It seems stars really are that small in comparison to a galaxy. Zoom into the Sun to max, and then "fly away" from it. Once you get about 3 light years away, it really is nothing more than a point. Atleast that's my observation.

Posted: 19.10.2007, 10:48
by ElChristou
I don't have any slow down, just a point as star and problems with labels...

Johaen, you can zoom on some DSO very very far away and invisible at normal FOV when a star is already visible...

Re: Zooming on an exoplanet not possible?

Posted: 20.10.2007, 08:21
by chris
ElChristou wrote:Guys, noob question again... (lately I think my neurone must have some kind of problem!) I wanted to watch the travel of an exoplanet in front of it's star, but apparently I'm unable to zoom enough for this. What's up? We can zoom on the faintest galaxy and not on a star no so far from us? ... or it's just me...


The amount of zooming is limited by the use of single precision floating point arithmetic by graphics hardware. Celestia prevents the FOV from being set too narrow in order to avoid some unpleasant jerkiness that is the result of precision limitations. You can experiment with the limit by adjusting the value MinimumFOV in celestiacore.cpp.

Note that Celestia relies on this the limited to FOV to optimize star rendering: it's assumed that stars further than a certain distance from the camera can always be rendered as points. So even if the precision errors weren't too distracting, you wouldn't be able to observer extrasolar transits from the Earth without making additional modifications to Celestia, like making the distance limit a function of the FOV (which should probably be done anyway, but it's not high priority.)

--Chris

Posted: 20.10.2007, 08:22
by chris
Cham wrote:I think there's a bug with the zoom function, at very high zoom values. This bug may be Mac related (I'm not sure of this, though).

Zooming to very high values (small angles) makes Celestia to slow down to a crawl.

Dirkpitt, can you confirm ?


I don't observe any slowdown. Maybe it's related to the settings you're using? If you can narrow down what is causing the problem, I definitely want to know about it.

--Chris

Posted: 20.10.2007, 10:01
by Vincent
Cham wrote:I think there's a bug with the zoom function, at very high zoom values. This bug may be Mac related (I'm not sure of this, though).

Zooming to very high values (small angles) makes Celestia to slow down to a crawl.


I can easily reproduce this bug on Windows with Celestia 1.5CVS.
It essentially happens in full screen mode, when zooming with Shift + Mouse left drag.
When I reach a zoom value of ~ 28503x and I still try to zoom further, Celestia displays the following string for FoV, and then freezes:
Image

Re: Zooming on an exoplanet not possible?

Posted: 20.10.2007, 10:36
by ElChristou
chris wrote:The amount of zooming is limited by the use of single precision floating point arithmetic by graphics hardware....
So even if the precision errors weren't too distracting, you wouldn't be able to observer extrasolar transits from the Earth without making additional modifications to Celestia, like making the distance limit a function of the FOV (which should probably be done anyway, but it's not high priority.)


Tx Chris, if you got it somewhere in your list it's fine. Reproducing an extrasolar transit would be nice to illustrate how those planets are detected...

Posted: 20.10.2007, 15:46
by Fightspit
I already reported this bug some month ago here:
http://www.celestiaproject.net/forum/viewtopic ... ght=#88313

Posted: 20.10.2007, 15:54
by eburacum45
I'd go to the star, then back away towards Sol; at a distance of several tens of AU's, you can get a reasonable transit. But are the orbits accurate enough to show these transits correctly?

Posted: 20.10.2007, 16:13
by Cham
Chris,

the bug is occuring while I'm in full screen mode (I'm always in that mode anyway, so I can't tell about the windowed mode).

When I zoom in to extreme values, and the nebulae are switched ON, I can then see some back models to the front ! I already reported this issue before (see also the thread indicated above by Fightspit).

I also indicated the bug to Selden, some time ago, while testing his Hubble FOV addon. While located at Hubble's POV, and zooming to extreme values, I then see the hubble model at the center, in front of me, while it is actually in my back ! And the slowdown is occuring at the same time. This bug is trully weird. The Hubble model is drawn at the center as a tiny model at the center. I also reproduced that weird effect with a grid (which is actually in my back and shouldn't be visible) : zooming in brings a deformed version of the grid to the front. It is like if ALL the nebulae are trying to be drawn at the same time (slowing down Celestia), as a tiny version at the center of the screen.

Posted: 21.10.2007, 09:48
by Vincent
chris wrote:I don't observe any slowdown. Maybe it's related to the settings you're using? If you can narrow down what is causing the problem, I definitely want to know about it.

I think I figured out what was going wrong in celestiacore.cpp.
I'll post a fix in the dev list. In the meanwhile, using this version
of celestiacore.cpp should solve the problem:
http://vincent.gian.club.fr/celestia/celestiacore.cpp

Sorry Christophe (ElChristou), let's get back to your discussion now... :wink:

Posted: 22.10.2007, 21:06
by Vincent
Fightspit wrote:I already reported this bug some month ago here:
http://www.celestiaproject.net/forum/viewtopic ... ght=#88313

I've commited a fix to CVS.
Fightspit, could you please confirm that the bug is gone ?

Posted: 22.10.2007, 22:36
by Cham
Vincent wrote:
Fightspit wrote:I already reported this bug some month ago here:
http://www.celestiaproject.net/forum/viewtopic ... ght=#88313
I've commited a fix to CVS.
Fightspit, could you please confirm that the bug is gone ?


Vincent, please, what is this fix supposed to do ? On my system, I don't see any difference at all.

Posted: 23.10.2007, 16:21
by Vincent
Cham wrote:Vincent, please, what is this fix supposed to do ? On my system, I don't see any difference at all.
Martin,

This fix has removed the bug that I reported a few posts above:
Vincent wrote:It essentially happens in full screen mode, when zooming with Shift + Mouse left drag. When I reach a zoom value of ~ 28503x and I still try to zoom further, Celestia displays the following string for FoV, and then freezes:
Image

This bug doesn't seem to be related to the one you're encoutering.
'My' bug was related to a floating point precision issue and thus,
it was OS (and compiler) dependant.