WMAP... Selden are you holding out on us?

[ogg]

All I can say is I feel teased. Selden posts (several times) a " " screenshot displaying his wonderful 2dF and Redshift addons plus 'others'. One of them being the WMAP microwave background radiation.
Perhaps I've completely overlooked the link to it, the forum search decided not to show me anything and/or I've gone mad, but if not:

PLEASE
PLEASE
PLEASE
PLEASE
PLEASE
PLEASE

PS deepest praises in your direction and Chris's.

[selden]

Patience is its own reward

I've been reluctant to release it because it has a big ugly seam where the edges come together.

I used Iris v4.03 to convert from the publicly available Aitoff projection to simple cylindrical. There are several fencepost errors in the map conversion code in that version of Iris. I reported them to the author but never heard back. v4.12 is available now, but I haven't had a chance to see if they've been fixed. Hopefully I'll be able to do that this evening (12 hours from now).

[ogg]

btw: Not sure if you'd have seen this already or not, but there's a 2k rectangular projection available from the Wmap website itself, not directly but as the texture used for their VRML model (model itself only ~300k, texture is 2k by 1k jpg, about 2mb). I haven't had a good look at it for seams though.

-edit- it's now available herefor anyone interested.


Keep up the good work and I'll try rewarding myself as best I can.

[selden]

OK, here it is:

http://www.lns.cornell.edu/~seb/celestia/wmap.zip (1.2MB, 9Mar04)

[updated 12Mar04 with models that have "protective shells" to try to avoid clipping problems seen on some systems.]

I won't include another copy of that same picture here

It contains 2 2K jpeg spherical maps, one of the all-sky background with a radius of 13.8GLY, and one of Q-Band radio waves with a radius 65K LY, the outer boundary of the molecular hyrogen gas clouds associated with the plane of the Milky Way.

Unlike my recent CMOD efforts, this Addon uses 3DS spherical models to hold the all-sky maps, so it can be used with Celestia v1.3.1 [not 1.3.0] or later.

Both models have only inward-pointing normal vectors, so the maps are only visible on the insides of their spheres.

[ogg]

Thanks Selden! Downloading now

[don]

It contains 2 2K jpeg spherical maps, one of the all-sky background with a radius of 13.8GLY, and one of Q-Band radio waves ...

Howdy Selden,

The Q-Band add-on displays okay for the most part, except for the noted seam. But I also get small black "patches" (geometric shapes) when I move the mouse ever so slightly (holding down the right button).

With the other one, if I move the mouse ever so slightly (holding down the right button) I get small "patches" (geometric shapes) of the image on-screen. Otherwise, no image at all.

Anything special I should be doing?

-Don G.

[selden]

Don,

I'm sorry to say that I don't see either of those problems. My guess is that you're seeing interactions between Celestia's new model code and ATI's graphics drivers. I suspect there are bugs in both

You might try looking at them using Celestia v1.3.1 final to see if it makes any difference.

It doesn't work with v1.3.0 final. The model specifies a wildcard image filetype, so it can be used with png, jpg or dds files. I included 2K jpeg files. Wildcards aren't recognized by 1.3.0.

[don]

Hi Selden,

The same thing happens in 1.3.1 final, so I suppose it's those nasty ATI drivers gettin' in the way again.

[selden]

Another possibility is the clipping region gremlin. I'll try creating some models with larger invisibie shells around them to see if that helps. My recent set of models haven't had anything extra like tha. I won't be doing that until later this evening, though.

[don]

Thank you Selden. You can just send it to me via e-mail if you want, and I'll test it with my ATI card.

[selden]

I have updated wmap.zip to use two new models that have "protecting shells" around them. They look unchanged on my system.

Please let me know if they make any difference for you.

http://www.lns.cornell.edu/~seb/celestia/wmap.zip

[don]

Hi Selden,

Sorry, it does the same thing. I even tried disabling all hardware acceleration via Windows. No luck.

[selden]

Don,

That's very strange...

The only substantive difference between "WMAP Q-Band" and "WMAP Background" is the radius, which is specified in the DSC file (wmap-trans-prot.dsc in this version). You might try reducing the diameter of the background object to see what that does.

Do my other all-sky maps work for you?
See http://www.lns.cornell.edu/~seb/celestia/billboard.html#7.2

The primary difference between this model and previous all-sky maps is that I made the material definition in this model translucent: the opacity value is set to 0.3. The others depend on an Alpha channel in the texture image.

[don]

Hi Selden,

Since you pointed me to the Radius, I did some experiments...

1. Started with the same radius as the Q-Band (Radius 13.0e5). This works.

2. Switched to the Background radius (2.76e10) but decreased the exponent to 6 (2.76e6). This works.

3. Increased the exponent to 7. Started displaying problems.

I hope this tells you something?

I've not tried your other spheres lately, but I seem to remember them working last year. I will try these again later.

[don]

Do my other all-sky maps work for you?

* mw-msx.zip works nicely.

* h-alpha_v1.zip says it's supposed to be 6.5MB, but the download is only 3.17 MB. It displays nothing, with either model selected. And I still see galaxy blobs.

[Toti]

Selden:
With the original .dsc, I had the same problems that Don related, but on my NVidia RIVA TNT2: when I changed the view direction with the mouse, I saw dark poligonal patches that showed up very briefly.
So I modified the radius entry and it now works OK. By the way, the upper usable value is some number between 8e6 and 8.3e6.

Thank you for the add-on!
Bye

[selden]

Don,

Sorry: the size of h-alpha_v1.zip was a typo. It's really about 3.3MB. Of course, "it works fine for me!" -- I just downloaded and installed it to make sure. Most of the sky should be dark with the default texture. If you look toward Orion, though, there should be lots of nebulosity visible. (Select Betelgeuse)

The h-alpha DSC specifies a radius of only 10000, so I guess that's not the problem.

Toti,

Apparently the depth of the z-buffer is much more of a limitation than I'd realized. My Ti4200 supposedly has a 32bit bit z-buffer. I can see deep-space all-sky maps that have a radius as large as about 3e19. Obviously other cards don't let you go so far

[don]

If you look toward Orion, though, there should be lots of nebulosity visible. (Select Betelgeuse)

No go. Sorry Selden.

[maxim]

While others are working fine, h-alpha also doesn't display anything for me.

Spec: GF4 MX 440, 64MB, 53.03, V1.3.1

maxim

[selden]

Are there any error messages in Celestia's console log? ( type a tilde to see it, tilde space on European keyboards)

Since both versions work perfectly for me, it's hard for me to guess what might be going wrong.

One problem for some cards is that V1 contains 4K surface textures. The jpeg expands to about 32MB. Cards with 32MB or less memory would have to support dds. I was going to create a smaller version but didn't get around to it. I'll do that shortly.

To see if the size of the images is your problem, you could try using this SSC to use them as AltSurfaces. Put it in the same directory with h-alpha.dsc.

Code: Select all

AltSurface "testdds" "Sol/Earth" { Texture "h-alpha.dds"}
AltSurface "test4k" "Sol/Earth" { Texture "h-alpha-a.jpg"}