1.5.0 changes

[Cham]

O well, I may have found a new bug in CVS. It's about 3ds -> CMOD conversion, this time. The bug may be related to the CMODTool conversion utility, but this needs to be confirmed.

There's a 3ds model which I've corrected the specular reflection and colors, to adapt it to 1.5.0. Here's what it looks like in 1.5.0, using the 3ds version :



Then, I asked ElChristou to convert that model to a CMOD version, since my CMODTool is broken under OS X Tiger 10.4.7 (the CMODTool was okay with 10.4.6, but the OS upgrade broke it). The CMOD version looks like this, in 1.5.0 :




You can clearly see the difference. It's HUGE and terrible !

[dirkpitt]

I agree that alternate cursors ought to be implemented, for feature parity with other platforms if nothing else. There's no consensus yet on whether a Script menu ought to be a feature or not.

Then in the meantime, lets go for cursor

I'm having some difficulty implementing alternate cursors for the OS X version. Actually, I should say it's easily done but there is a 2-5 fps performance hit, or about 10%. With some tweaking, it doesn't seem impossible to do it without the framerate drop, but for the time being, I want to work on other things so this feature is being shelved.

[chris]

O well, I may have found a new bug in CVS. It's about 3ds -> CMOD conversion, this time. The bug may be related to the CMODTool conversion utility, but this needs to be confirmed.

There's a 3ds model which I've corrected the specular reflection and colors, to adapt it to 1.5.0. Here's what it looks like in 1.5.0, using the 3ds version :

You can clearly see the difference. It's HUGE and terrible !

One of the reasons that specular highlights look better in 1.5.0 is because I fixed a broken conversion of specular power when converting from 3DS to CMOD format. I just checked in a similar correction to 3dstocmod's conversion function. A version of 3dstocmod built from CVS should fix the problems. Let me know if you need me to provide one for you.

--Chris

[ElChristou]

I agree that alternate cursors ought to be implemented, for feature parity with other platforms if nothing else. There's no consensus yet on whether a Script menu ought to be a feature or not.

Then in the meantime, lets go for cursor

I'm having some difficulty implementing alternate cursors for the OS X version. Actually, I should say it's easily done but there is a 2-5 fps performance hit, or about 10%. With some tweaking, it doesn't seem impossible to do it without the framerate drop, but for the time being, I want to work on other things so this feature is being shelved.

For sure! A drop in the FPS for a cursor is not cool at all ...

[Cham]

One of the reasons that specular highlights look better in 1.5.0 is because I fixed a broken conversion of specular power when converting from 3DS to CMOD format. I just checked in a similar correction to 3dstocmod's conversion function. A version of 3dstocmod built from CVS should fix the problems. Let me know if you need me to provide one for you.

This is a confirmation of what I believed : it's the actual CMODTool available for OS X which needs an update. Dirkpitt ?

[dirkpitt]

This is a confirmation of what I believed : it's the actual CMODTool available for OS X which needs an update. Dirkpitt ?

I'm recompiling it right now...

[dirkpitt]

This is a confirmation of what I believed : it's the actual CMODTool available for OS X which needs an update. Dirkpitt ?

I'm recompiling it right now...

Done.

Download CModTool 1.0.3 for OS X (Universal)

I don't know if this build will fix the issues that were appearing on 10.4.7.

[ElChristou]

This is a confirmation of what I believed : it's the actual CMODTool available for OS X which needs an update. Dirkpitt ?

I'm recompiling it right now...

Done.

Download CModTool 1.0.3 for OS X (Universal)

I don't know if this build will fix the issues that were appearing on 10.4.7.

Many Tx Dirkpitt, go for a test!

[dirkpitt]

Sorry for the cross-post, but just wanted to let you know that I've made a new build of CModTool:

Go to CModTool 1.0.3 (build 4) announcement

This build should not only fix the issues that started appearing with OS X 10.4.7+,
but should also work with a lot more models than previous versions. Thanks Cham for testing!

[ElChristou]

Re TX!

[Cham]

I just recompiled Celestia 1.5.0 today, using a fresh installlation of the CVS code (8 november). I didn't made any other compilation since september. After playing with it for about an hour, I've noticed few things. Here's a short description :

1- The comet's tails are cutted by a "clipping plane". See the picture below :


2- The star glare effect is still disappearing at some distance. This is very annoying and I thought this was solved some time ago.

3- There's no star labels anymore. Using the "B" shortcut (or playing with the preferences options) don't show any star label. ??

4- Fridger's galactic labels code isn't there yet. I'm wondering why.

5- I'm still experiencing the dashed orbital paths problem while moving around an object. I never had this problem with Celestia 1.4.1.


6- The clouds shadow doesn't work. Of course, I'm using OpenGL2.0. I know Chris desactivated that feature some times ago, but is it still desactivated ? Why ?

7- I'm still getting some random tiny white dots on models (I never had this with 1.4.1). See the asteroid above, on its upper part. There are some white dots there, looking like stars, but they aren't. This effect is general and may occur on any model.

[selden]

I can't help with getting the problems fixed, but I can comment on 3, 4 and 6:

[I realize you know most of the keyboard shortcuts. I'm including them for people who don't. All of the necessary commands *should* be in the "Render/View Options" menu, but several are not ]

3. Unfortunately, Fridger's new star label code is incomplete. At the moment, a star label is visible only when all of 4 conditions are met:

a. Star labels are enabled (type b to toggle them)

b. the star is above a certain magnitude threshold (type [ and ] to adjust it)

Many of the brightest stars are labelled when "Magnitude limit" has a value of 6.5 or greater.

c. Any of the "OpenGL vertex program" render paths is enabled (type Ctrl-V to step through them)

d. the new High Dynamic Range code is enabled (type @ or shift-2 to toggle it)

(I call Fridger's code incomplete because requirements c & d should not be necessary.)

4. A galaxy's label is visible only when

a. galaxy labels are enabled (type e to toggle them)

b. the galaxy is well above visibility threshold; i.e. when "Magnitude limit" has a large enough value. (type [ and ] to adjust it)

On my system, for example, the label for M 31 is visible only when Magnitude limit has a value of 6.8 or greater. Since M 31 is the brightest, that means that no galaxy labels are visible from Earth at a smaller value than that.

6. Cloud shadows are visible when they are enabled. This requires either a celx command or for the "Show Cloud Shadows" box to be Checked in the Render/View options menu. I suspect it hasn't been added to the Mac's menu yet, although it is in the Windows menu. There is no keyboard shortcut to toggle cloud shadows.

[t00fri]

I can't help with getting the problems fixed, but I can comment on 3, 4 and 6:

[I realize you know most of the keyboard shortcuts. I'm including them for people who don't. All of the necessary commands *should* be in the "Render/View Options" menu, but several are not ]

3. Unfortunately, Fridger's new star label code is incomplete. At the moment, a star label is visible only when all of 4 conditions are met:

a. Star labels are enabled (type b to toggle them)

b. the star is above a certain magnitude threshold (type [ and ] to adjust it)

Many of the brightest stars are labelled when "Magnitude limit" has a value of 6.5 or greater.

c. Any of the "OpenGL vertex program" render paths is enabled (type Ctrl-V to step through them)

d. the new High Dynamic Range code is enabled (type @ or shift-2 to toggle it)

(I call Fridger's code incomplete because requirements c & d should not be necessary.)

4. A galaxy's label is visible only when

a. galaxy labels are enabled (type e to toggle them)

b. the galaxy is well above visibility threshold; i.e. when "Magnitude limit" has a large enough value. (type [ and ] to adjust it)

On my system, for example, the label for M 31 is visible only when Magnitude limit has a value of 6.8 or greater. Since M 31 is the brightest, that means that no galaxy labels are visible from Earth at a smaller value than that.

6. Cloud shadows are visible when they are enabled. This requires either a celx command or for the "Show Cloud Shadows" box to be Checked in the Render/View options menu. I suspect it hasn't been added to the Mac's menu yet, although it is in the Windows menu. There is no keyboard shortcut to toggle cloud shadows.

Selden,

I don't understand much about what you are writing about my code.

1) The dynamical star label code as YOU know it has been written by Chris, (who admittedly copied it essentially from my code for galaxy labels) All the "conditions" for it's related incompleteness are his not mine...

2) I wrote some extension to vary the labels' transparency as function of distance which Cham likes a lot. YOU don't know that code yet, I suppose, since I sent it only to a few people for testing.

3) The constraints for and restrictions on dynamical galaxy labels are different from those for star labels. This is related to the fact that the average apparent brightness of the NGC/IC catalog objects is way lower than the average brightness of the HIP stars.

4) My dynamical galaxy labels intrinsically rest on usage of my automag feature, like a sensible display of galaxies of such a large and dim brightness range essentially need the automag.

I'd appreciate if you would first inform yourself before elaborating about the incompleteness of "my" code....

[selden]

Fridger,

I guess I'm misremembering your posting.

Sorry.

Nontheless, the current star label code is incomplete, having the problems I listed.

[dirkpitt]

I haven't yet gotten around to adding a checkbox for Cloud Shadows in the OS X prefs panel. Chalk it up to plain laziness Adventurous users with Xcode installed may edit their net.shatters.Celestia.plist using Property List Editor. Search for the setting "showCloudShadows" and set it to 1.

[selden]

Fridger,

This seems to be the posting that I remembered:

http://celestiaproject.net/forum/viewtopic.php?p=72916&#72916

where you wrote

Hi again,

I added a crucial enhancement to the distance dependent label transparency for galaxy and star labels!

I assumed you had been making changes to the star code in the cvs repository, not just to your private copy. Obviously I misunderstood.

Sorry.

[dirkpitt]

I just recompiled Celestia 1.5.0 today, using a fresh installlation of the CVS code (8 november).

1- The comet's tails are cutted by a "clipping plane"

I can confirm this.

7- I'm still getting some random tiny white dots on models (I never had this with 1.4.1). See the asteroid above, on its upper part. There are some white dots there, looking like stars, but they aren't. This effect is general and may occur on any model.

I can also reproduce this, but only with FSAA (anti-aliasing) on. My setting is AntialiasingSamples 4.

[t00fri]

Fridger,

This seems to be the posting that I remembered:

http://celestiaproject.net/forum/viewtopic.php?p=72916&#72916

where you wrote

Hi again,

I added a crucial enhancement to the distance dependent label transparency for galaxy and star labels!

I assumed you had been making changes to the star code in the cvs repository, not just to your private copy. Obviously I misunderstood.

Sorry.

Unlike other devs, I have the habit to commit my code only to CVS when it is final from my point of view and after a bunch of people have played with it without complaints...

Bye Fridger

[chris]

1- The comet's tails are cutted by a "clipping plane". See the picture below :

Thanks for the bug report. I checked in a fix this morning; it corrected a bug that's been there since the beginning of time (or at least, since the comet tail code was written) and was probably only recently exposed by the new depth sorting code.

2- The star glare effect is still disappearing at some distance. This is very annoying and I thought this was solved some time ago.

The problem is solved in the new star code, which will be the default very soon. To enable it now, press the @ key. The old star code is still required for cards that don't support point sprites. The manual toggle between old and new star rendering will be going away, and Celestia will automatically select the mode based on your graphics hardware's capabilities.

Code: Select all

3- There's no star labels anymore.  Using the "B"  shortcut (or playing with the preferences options) don't show any star label.  ??


The dynamic star labeling code still needs to be made to work with the old star code. For now, just toggle the new star code by pressing @.

Code: Select all

5- I'm still experiencing the dashed orbital paths problem while moving around an object.  I never had this problem with Celestia 1.4.1.


I've had a very difficult time reproducing this. I'm assuming that it must be due to the different method that ATI hardware uses for clipping.

7- I'm still getting some random tiny white dots on models (I never had this with 1.4.1). See the asteroid above, on its upper part. There are some white dots there, looking like stars, but they aren't. This effect is general and may occur on any model.

I've had a hard time reproducing this. Dirkpitt said that it only occurred with AA enabled, but I don't see white dots whether AA is on or off.

--Chris

[Cham]

Chris,

About 5 : The dashed path bug may be related (?) to FSAA, but I'm not sure. I never had this problem with previous versions of Celestia (using the same computer, same OS version, same ATI card etc). It's only occuring with 1.5.0. On my system, it's very easy to reproduce, especially if I'm following an asteroid.

About 7 : I think the white dots are related to FSAA too. The dots are very subtle but systematic. This is also a new bug with 1.5.0, and never occured with any other versions of Celestia (prior to 1.5.0).

Only Dirkpitt confirm the #7 bug ? Are both of us the only Mac users with an ATI card and FSAA enabled here ?