Questions about Normal Maps and.........

[cartrite]

Hi,

I'm using the VT surface and normal maps from John van Vilet downloaded from the motherload, and sometimes all the craters look like mounds. Is this normal? Is it my card? Using a GeForce 5500 driver 78.01 256 mb ram. This occurs with all rendering modes.

[img][img]http://img226.imageshack.us/img226/4056/m31frommarsorbit48er.th.jpg[/img]

Also, the image above also shows m31, something I've been working on. Is there a way to surround it in a gas cloud to make it look more realistic?

Also, the image below shows a VT close up of mars (holden crater region)
that I've ben working on. The tex was derived from an image from JPL's image gallery on Mars. I need a texture of that general area (not that close level 6 or 7 ?) to supply the close up with some background. Anyone out there know where I could get one? I've looked all thru the image gallery at JPL. This VT jumps from level 4 to level
to level 8, so I'm looking to add some background in levels 5, 6, or 7.

[img][img]http://img91.imageshack.us/img91/2830/mars48ww.th.jpg[/img]

Later,

cartrite[/img]

[Don. Edwards]

This is an optical effect that happens most of the time when the planet is lit from the wrong side. So if you have the sun of the left, try rotating the planet around so that the sun is to the right and see if visuals look right. the reason this hapens is that when the normalmap was mage it had either some shadow detail in it or the normalmaps hiehgt wasn't done right. There is also the posibility it got inverted in the proceesing.
I think the normalmap you are speaking of is for Mars? if so I will have a new Mars texture and normalmap ready for downloads vary soon. They won't be virtual textures but by the specs of your video card you have more than enough memory to use them. Hope this helps.

Don. Edwards

[buggs_moran]

Hi,

Also, the image above also shows m31, something I've been working on. Is there a way to surround it in a gas cloud to make it look more realistic?

You can try the technique of John Whatmough (quoted below), who tragically isn't with us anymore. However, I do not know the results of an object at that distance. Selden showed me that it was possible to have a barycenter further than 16000 LY but your talking 150 times that distance. Celestia, as far as I know, cannot do volumetric objects other than atmospheres...

http://www.celestiaproject.net/forum/viewtopic ... eroid+belt

Basically, the belt is a combination of three "planets", each with a zero orbit that puts them inside the star. The first has a giant ring that extends 100,000,000 km. The second is highly oblate and has a atmosphere height the same value as the ring width. The third is the same as the second, only less oblate and smaller. The actual radius of each "planet" is ~600000 km, so the "planets" themselves are hidden inside the star. And, of course, the radius of each is slightly different than the others to prevent them from "blinking".

Only flaw so far is that you have to turn off ring shadows. Otherwise the "planets" cast a long thin shadow on the ring. You can make the ring emissive, but then you get an odd brightness gradient across the ring. I've toyed with using a null mesh to remove the planet shadow, but this tends to crash the program.

Also, zooming in on the star penetrates the atmospheres, so you'll get the same kind of blinking and odd effects as when you get too close to a planet with an atmosphere, only more so because there are two atmospheres overlapping.

-John

[cartrite]

Sorry to hear that John's no longer with us. That picture he did was great. He sure was talented. That idea he came up with is something. I'll have to look into it. Thanks for the tip.

Don, thanks for the tip also. You know, Ive notied the bumps on the craters for a while now. Sometimes they looked bumpy and sometimes they looked like normal craters. I never put it together that the light direction was causing it. Your responce brought up another question. In the opengl info box it says that my max texture is 4k. Is it possible for some of you people out there to use 16k textures? I have the older pci bus that runs at 133 mhz. Is that why I only get 4k? And your textures on the motherload site, I thought they were your new ones. I managed to VT the 16k version that is on the motherload now. Both the Nomal and Surface. Together they zip up in 85 mbs. I'm testing them now to see if I did them right.

Thanks Again

cartrite

[cartrite]

Mabey I'm going crazy,

I sat in orbit over mars. I took 3 screenshots. The first was of Don's 4kversion, the second was the 16k version, 3rd John van Vilet's VT.
Same spot, just switched textures. First two showed bumps, third showed craters. (But, and this a BIG BUT), when the screen shots were viewed with Windows Picture and fax viewer, The Bumps Appeared and Dissapeared at random. Driver Problem? Well I'm leaving this issue for the twilight zone.

cartrite

[maxim]

It's simply a perception problem. No need to blame your video drivers.

It can not only be seen on celestia textures, but also on actual pictures released by NASA or ESA. One reason is i.e. that we are used to have lighting coming from above. If light comes from the lower side of the screen your perception tends to switch. If your perception is stable, or starts jumping is also related to your /mood/. Sometimes it's heavy and irritating, sometimes not. If you can identify and keep beeing aware of the direction of the light source, things should calm down.

(There are additional reasons too)

maxim

[cartrite]

I'm sure ther are a million reasons this happens. What amazied me was this:
1. I took screenshots to show others what was happening.
2. Depending on what veiwer I used, the effect was there or it wasn't.
a. The gimp shows only a normal shot with craters.
b. All the other viewers show the effect on one occation, no effect on the next.
So then I wondered if other people using a different system can even see what I see.
Anyhow my basic question is answerd. That is, Was this a common problem or a local hardware software problem?

cartrite

[danielj]

Now YOU are talking about the benefits of more video card memory.But when I asked about in a latter post,they say you have to assume a compromise between speed and memory.Now what?I have already bought my video card and it has only 128 MB!!!Why didn??t you explicit that 256 MB video card was highly desirable?If I knew that,I would had bought a Geforce 6600 256 MB.Now I have a BIG problem in my hands.I spend almost all my money and don??t have enough to buy a 256 MB video card.


I think the normalmap you are speaking of is for Mars? if so I will have a new Mars texture and normalmap ready for downloads vary soon. They won't be virtual textures but by the specs of your video card you have more than enough memory to use them. Hope this helps.

Don. Edwards[/quote]

[t00fri]

Now YOU are talking about the benefits of more video card memory.But when I asked about in a latter post,they say you have to assume a compromise between speed and memory.Now what?I have already bought my video card and it has only 128 MB!!!Why didn??t you explicit that 256 MB video card was highly desirable?If I knew that,I would had bought a Geforce 6600 256 MB.Now I have a BIG problem in my hands.I spend almost all my money and don??t have enough to buy a 256 MB video card.

Daniel,

you now go on with reproaches that people in this forum did not give you the right advice.

Noone can replace application of your OWN common sense!

I will never understand, why you did not make yourself sufficiently knowledgeable by reading various serious reports carefully and draw your OWN conclusions from them before buying.

A few back-of-the-envelope calculations would easily have shown you that more memory is always the best choice with regard to Celestia. I would always have bought a 256MB card, no doubt.

But for making a balanced decision, ONLY YOU have a clear view about your available TOTAL budget and the costs of hardware in your country. When you spend relatively more money for your graphics card then presumably you will have to save on the CPU, the RAM, the speed and size of the harddisk or the quality of your monitor. ONLY YOU have a clear view about other possible applications you want to do with your new computer and hence about the optimal RELATIVE cost weight for the individual hardware components.

Only the (sometimes tedious) process of ACCUMULATING KNOWLEDGE and of practising to apply your COMMON SENSE to a pending decision will prevent WRONG DECISIONS in the long run.

Whether you like that or not: it's simply a FACT! In addition, following my advice will have many benefits for you also outside the narrow perspective of Celestia.

Just learn relying more on YOURSELF, there is great hidden potential, I am sure!

Bye Fridger

[Boux]

cartrite,
Don't blame yourself, your video card, size of memory, relevance of normal map versus surface map, whomever or whatever.
Maps are 2d only and both eyes send the same information to the brain.
Our brain has no way to know whether what we are seeing is a crater or a mountain (a "bump").
The silly thing is that sometimes our brain temporarily decides that it is a crater then the next couple seconds it switches to the perception of a bump.
It's a perception artifact.
When the light direction is obvious, then it can help because an additional information is sent to the brain which has a better chance to choose the "right" solution.
The same when there are remnants of a meteorit in the middle of a crater. The brain interprets it as a likely higher ground (I believe this is why this aberration happens less often when we look at the Moon.
Our cultural background and the standard clich?© of an impact crater tells us that many craters on the Moon have such remnants in the middle of many craters).
We are looking at 2d textures, not 3d. Any additional information is synthesized by the brain based upon a mix of cultural elements, souvenirs, frequence of similar situations, other images already seen and stored in memory and probably current mood.
We are in the realm of perception and psychology here

[maxim]

cartrite,

...

The silly thing is that sometimes our brain temporarily decides that it is a crater then the next couple seconds it switches to the perception of a bump.

...

We are looking at 2d textures, not 3d. Any additional information is synthesized by the brain ...

Thats some of the other reasons I talked about. You're looking at the surface from directly above. If the brain would have additional informations like an object partially (even a very tiny part) hidden by another, or some movement of the ground or the light source, it would be /much/ easier for it to decode the proper view.

maxim

[Spaceman Spiff]

I have trouble seeing this crater pit/mound inversion from that first picture, presumably because it shows the night side of Mars under ambient light. With effort, it does look like the craters are pits - I recognise Galle (the 'smiley face crater') as a crater - but this could be due to shadowing retained in the original texture map. Bump mapping doesn't apply on ambient lit night sides does it? So how can we see the effect?

The above comments about the ambiguity illusion concerning shadows is classically demonstrated on TV with a mask of a face, where it looks like the face has its nose pointing out. Then the face is rotated, and suddenly all the features move to the wrong side, and one realises that the face was a hollow mask with the nose pointing away. Um, is that clear?

On 2D pictures of unfamiliar surfaces, this is always a problem, especially with real high resolution spacecraft photos of the Moon and Mars.

Finally, wasn't there a bug in an older version of OpenGL (1.2?) which made bump map shadows away from the equator rotate until they faced the wrong way near the poles? Could that be the case?

Spiff.

[danielj]

So the only difference I will see,in Celestia,between my old FX 5700 and my new 6600 GT will be the frame rate?No graphic improvement,at all?

[cartrite]

I can understand that perceptions can change in an active environment.
But to me, taking a snapshot freezes a scene in time and space. It is no longer active. No changing light direction. No movement. Everything is frozen. So something is on a picture or it is not.
But in the world of ones and zero's, things can change for any number of reasons. I guess. Anyhow, I started doing a texture of "The Face on Mars".
I'm trying to make a model for it now. So it appears to be a mountain if viewed from the side. That means I'm trying to learn how to use Blender. Is there a way to model a VT or will I have to use a single texture. Kinda like StoneHenge. Here is a shot of the VT.

[img][img]http://img326.imageshack.us/img326/7664/face0op.th.jpg[/img]


cartrite[/img]

[jestr]

Hi Cartrite,at present it isnt possible to use a VT on a model,but if you have the patience it is just possible to cut the model up into the relevant pieces (to go with the VT tiles),but hard to get them all lined up right in Celestia and up close you can see the joins.Jestr

[cartrite]

Yeah thats what I thought, If I ever get the model done, I may try them both at the same time. The original was greyscale 2400x2400. I scaled one up for level 10 and scaled one down for level 9. I'll use the 2k. The main purpose I'm trying a model is it to see hills on the horizon, while Im siiting in very low sync orbit, or on the surface watching the stars go by. Just speed up time and let em rip. No light pollution here.

[img][img]http://img392.imageshack.us/img392/6179/sunset13ky.th.jpg[/img][/img]
[img][img]http://img392.imageshack.us/img392/306/smc39sw.th.jpg[/img][/img]

By the way these were taken with FT-1. The sun by Runar Thorvaldsen and SMC by Fridger and Toti. Beautiful jobs.

cartrite