• Create Account

# Lauris Kaplinski

Member Since 02 Aug 2006
Offline Last Active Feb 18 2013 02:33 AM

### #4928286Collada vertex format and creating Terrain

Posted by on 04 April 2012 - 01:57 PM

You cannot expect anything about the order of vertices/polygons, unless your modeling program explicitly says, that it keeps them in certain order in exported file.

Instead I'd suggest simply walking through all vertices, rounding their X & Y (maybe scaled by sme factor, if your grid step is not exactly 1) coordinates to nearest integer (thus getting array coordinates) and saving Z as the elevation at that point.

### #4927842Bone animation and skinning

Posted by on 03 April 2012 - 05:36 AM

Bind pose is the skeletal pose, that corresponds to untransformed skin. I.e. for normal biped characters it is T-pose. I assume it is the global transformation - if not, you have to calculate the global bind matrix yourself.
The vertex transformation is done the following way:

Let matrix B2S be the rest (bind) global transformation of bone.
Let matrix P2S be the rest (bind) global transformation of the parent of given bone
Let matrix AB2P be the animated local transformation of bone
Let V be untransformed vertex and V' animated vertex

Then

V' = (P2S * AB2P * B2S(-1))

The trick is to precalculate everything you can - B2S(-1) can be calculated once for bone, P2S * AB2P can be calculated once for animation frame

Now, if you have more than one bone transforming given vertex you just have to calculate weighted average of all transformations (i.e. weighted average of transformed vectors).
How to best do that in vertex shader depends on shader model used. If you do not want to use integer attributes and dynamic loops, you will use matrix palettes. Basically each vertex is always influenced by all bones. If in reality given bone does not transform given vertex, it's weight will be 0 for that bone.
Simple matrix palette implementation can kill performance because you have very long list of matrix operations for each vertex, most with weight 0. Depending on model and skeleton complexity it may be beneficial (or necessary) to break model into fragments so that each fragment is influenced by only subset of all bones.

### #49274322D vs 3D and some other newbie questions

Posted by on 02 April 2012 - 04:53 AM

- I'm confused by isometric games. Are they 2D, 3D, or can they be both?

Isometric projection is one kind of 3D projection to 2D surface, the other being perspective projection. So in theory isometric games are 3D.
Now isometric projection does not have perspective - i.e. all objects appear the same size, regardless of distance. Also, if you do not plan to allow changing of camera angle, all objects ALWAYS appear the same. Thus it is possible to render isometric world with 2D tileset and sprites - and it is often done that way.
Whether it is actually easier to do it in 2D or 3D depends on many factors.

### #4879400I have a problem with theory of relativity

Posted by on 01 November 2011 - 03:42 PM

This is the scenario:

A spaceship is launched from Earth at a relativistic speed towards a distant planet. One observer is placed on the spaceship and one is placed on the planet.

At the time of launch, two stopwatch clocks are started simultaneously -- one on the spaceship and one on the distant planet.

This alone is meaningless statement in STR.

Simultaneity depends on reference frame for any spatially separated events (in this case spaceship and planet). I.e you have to specify in which reference frame - "Earth/Planet" or "moving spaceship" these stopwatches were started simultaneously.

This issue is muddled by the acceleration spaceship has to undertake in above example. But you can make it cleaner so, that spaceship flies with constant speed and synchronizes its clock at the moment, it is passing by Earth in its way towards the planet.

When the planet-based observer sees 20 years on his stopwatch, the spaceship flies by the planet. Here are my questions:
1. What will the planet-based observer see on his own stopwatch? answer: 20 years

Correct by definition.

2. What will the planet-based observer see on the ship stopwatch?

If clocks were started simultaneously in Earth/planet reference frame, then something << 20 years
If clocks were synchronised in moving spaceship reference frame, then something >> 20 years

3. What will the ship-based observer see on the ship stopwatch?

Exactly the same value that planet-based observer is seeing (if they are close enough to each-other you can ignore all relativistic effects)

4. What will the ship-based observer see on the planet-based stopwatch?

Also the same value that planet-based observer is seeing (i.e. 20 years)

### #4877151View matrix to reflected view matrix

Posted by on 26 October 2011 - 06:12 AM

To get mirrored view matrix (for reflection) you should multiply your view matrix (generated with lookAt) with mirror matrix of given plane.

M'camera = Mreflection * Mcamera

You can take a look at Parts 1.1 and 1.2 of my reflection tutorial, the construction of reflection matrix is shown there (or check any other 3D/matrix/vector geometry reference).

If your water level in scene is Y, the water plane is (0, 1, 0, -Y) - assuming Y is upwards direction.
You may also want to add extra mirror transformation (around X or Z axis in your case), to keep the winding (culling) order of polygons intact.
You can also use the same mirror matrix to transform the view and target points (and possibly up vector) for lookAt function.

### #4867120Possible neutrinos travel faster than light

Posted by on 29 September 2011 - 04:36 AM

Neutrinos moving faster than light in vacuum does not automatically refute the theory of relativity.

• It may be, that relativity still holds, but light itself moves slower than C
• It may be, that neutrinos travel faster than C, but this process does not carry information
• Even if neutrinos can carry information faster than C it may mean, that relativity still holds but causality, as we know it, does not exist in physical world
In any case, relativity, faster-than-light communication and causality do not seem to fit together. I personally hope that the last one will be thrown out and the universe turns out to be perfectly deterministic again

3. Don't worry, the last one is definitely not the case. Violations of causality are prone to logical contradictions. Logic is the only thing we can always assume in any coherent discussion- with out which, the principle of explosion makes the conversation itself meaningless. In so far as we're thinking or talking about anything, logic holds- and so must causality.

Causality in philosophy already is logically incoherent, so not big loss here.

Also strictly deterministic theories, like pure Newtonian physics are time-symmetric and thus have no place for true causality. If ball A hits ball B and "causes" it to move, you can simply invert time and say, that ball B "caused" the movement of ball A backwards in time. We can only distinguish between them because of intuitive assignment of causality with forward-in-time influences, but this has no any physical meaning.

If Qm anf GTR taught us anything, the most important thing is, that we cannot carry the notions of our day-to-day logic naively into physical realms. Of course the theories have to be logically coherent - but "normal" causality does not belong to that part of logic.

2. Neutrinos carry information; that's how they were detected. Neutrino vs. no neutrino is information.

Probably. I simply do not know enough to exclude any entanglement-like setups, where we can post-factum establish, that neutrinos in CERN and Italy "appeared" within certain time interval but not have enough control in CERN to "set" the starting point of given time interval with enough precision.

1. This is... extremely unlikely. Like I mentioned in my post above, light may be being slowed down by something, but the only thing that would have been slowing down light in this experiment relative to cosmic observations would be gravity, and if we allow for gravity to slow down one thing, and not another, we have to disassociate space and time, which refutes special relativity by creating a special reference frame (space) independent of the passage of time experienced by different matter/energy in different places.

Why is gravity "the only thing" that could have slowed down light?

How about our theory of light is completely wrong and it never moves with speed C but always slightly slower depending on other factors than gravity? Thus both special and general relativity can still hold (as C being constant in all reference frames but not the speed of light).

### #4866433Possible neutrinos travel faster than light

Posted by on 27 September 2011 - 07:42 AM

FTL travel indeed would make traveling backwards in time possible, but not because of some naive interpolation of equations.

According to the theory of relativity, if two space-time points A & B are separated by space-like distance (i.e. light cannot travel from A to B before B takes place and vice versa), the time-order of those events is not definable. I.e. there are always some reference frames, where A happens before B and some where B happens before A.

Now, if somehow something travels from A to B FTL and reaches the spatial location of B "at or before" B takes place, then in those reference frames where B takes place BEFORE A, it has traveled backwards in time.

What do you mean, 'has travelled backwards in time'? Indeed, faster than light travel would play funny tricks with our perceptions of causality, but not with causality itself. Yes, if you go faster than light, people get to see your present state before ever getting to see your past states. You can do essentially the same thing with a good old mirror.

According to the theory of relativity there is NO preferred reference frame - thus the descriptions of reality from all inertial reference frames are equally true.
As I gave in above example - for space-like events in space-time there are no fixed (true) order of precedence. If events A and B are separated by space-like distance, then always for some inertial reference frame (let's call it frameX) event A happens before event B, for some other frame (let's call it frameY) event B happens before event A (and, of course for some frames they are synchronous). And all these descriptions are equally true descriptions of the Universe.

Now, if signal reaches from A to the spatial location of B before ot at the time B takes place, then viewing the situation from reference frame Y:

• Signal reached from A to (the spatial location of) B before B took place
• B happened BEFORE A
Thus signal travelled backwards in time.

### #4866373Possible neutrinos travel faster than light

Posted by on 27 September 2011 - 03:58 AM

snip

What are you talking about the theory of relatively does talk about time and therefore time travel. According to the theory of relatively the faster something travels the slower time passes for that object. Also the higher the gravity around an object the slower time passes for it. Therefore if you travel faster than the speed of light, according to the equation, you would go backwards in time. The notion of time travel was originally proposed by scientist studying relatively, not science fiction writers.

O RLLY?

Indeed, more velocity means slower passing of time; but clearly the geometry of the equation doesnt allow for the extrapolation you posit here, since time dilation as a function of v isnt even differentiable at v==c; it has an infinite slope. The correct extrapolation of ever more velocity isnt into negative time; the correct extrapolation is that the function doesnt extend into the v>c domain at all cause it completely curves away from that domain and doesnt point toward it at all.

FTL travel indeed would make traveling backwards in time possible, but not because of some naive interpolation of equations.

According to the theory of relativity, if two space-time points A & B are separated by space-like distance (i.e. light cannot travel from A to B before B takes place and vice versa), the time-order of those events is not definable. I.e. there are always some reference frames, where A happens before B and some where B happens before A.

Now, if somehow something travels from A to B FTL and reaches the spatial location of B "at or before" B takes place, then in those reference frames where B takes place BEFORE A, it has traveled backwards in time.

### #4866371Possible neutrinos travel faster than light

Posted by on 27 September 2011 - 03:51 AM

IF this is proven correct and IF we can make use of it the only real use I can see is FTL communication. Now, with respect to light speed this is a form of time travel, in that information could make it from one location to another before light (to use the sci-fi example you'll be able to see the enemy fire his laser slightly before it starts trying to remove your hull) but it only remains 'time travel' in that sense.

I think it is quite well established, that there is no actual difference between "communication" and "effect". Thus if any information can be transmitted FTL, true time travel is also (theoretically) possible.

This also would make type 2 perpetum mobile possible, as you could use the information abut future state to influence current state, thus always choosing the state, that will result in minimal future enthropy.

### #4865999Skinning Issues

Posted by on 26 September 2011 - 04:34 AM

You description seems correct, so maybe you have done a typo in matrix multiplication.
Btw, what exactly do you mean by bone bind pose matrix? Is it relative to parent or relative to skin?
I usually use the following notation:
• B2P - bone "rest" orientation, relative to parent
• B2S - bone "rest" orientation, relative to object/skin - it is the product of the chain of matrices from root to bone (B2S = B2Sparent * B2P)
• S2P the inverse of B2S
• AB2P - animated orientation of bone - relative to parent
• AB2S - animated orientation of bone - relative to object/skin (AB2S = AB2Sparent * AB2P)
• The matrix you will need (the one transforming skin vertices) is:
• AS2S = AB2S * S2B

### #4865876Shadow-map for multiple omnidirectional lights

Posted by on 25 September 2011 - 04:04 PM

Maybe I am missing something, but IMHO you are writing the distance from light to shadow buffer.

I.e. your fragment is in light, if shadow map value is GREATER than fragment distance from light.

### #4865852Atmospheric Scattering - Out-scattering integral solution?

Posted by on 25 September 2011 - 02:47 PM

Maybe I misunderstood something, but IMHO:

h =/= ((A + (C*x/N)) - R)

i.e. altitude is not a linear function. You can start from upper atmosphere (A), go through point at sea level and end in upper atmosphere again (B).

### #4865235Shadow-map for multiple omnidirectional lights

Posted by on 23 September 2011 - 12:36 PM

Thanks for your answer. I've read about the conventional way with a geometry shader, the 6 passes method and also paraboloid-mapping. But isn't there a way to skip the projection?
I thought, if I transform the vertices in light-space, i could use their direction as a cubemap coordinate and their distance from the light-space null-vector as depth-value.
Is there really no way to write from fragment-shader to a cubemap with directional texture-coordinates?

I've tried to set gl_Position in the vertex-shader of the shadow-map-calculation to the normalized vertex-position in lightspace and to use the vertex's distance to the light-space null-vector as fragment-color. But if I use this shader with an FBO which has cubemaps as depth- and color-attachments only one face of the cube is used, the other faces are static but seem to be undefined.

Okay, I probably understood what you meant.

Basically you want to replace standard projection (frustum) with custom calculation of gl_Position.

I think the biggest problem is, that your triangles will not be rasterized over cubemap. Rasterizer projects your triangle to XY viewport plane and then runs fragment shader for all pixels - if, for example, the triangle is parallel to XZ plane in clip coordinates (i.e. after perpective division) it will not be rasterized at all because it covers 0 viewport pixels.

Also do not forget, that single triangle can cover many sides of cubemap. The maximum is 5, although 2 and 3 are probably more common. Rasterization is bound to single layer/viewport I think. And I am pretty sure rasterizer will not select cube map side automatically - you have to do it in geometry shader.

### #4865169Shadow-map for multiple omnidirectional lights

Posted by on 23 September 2011 - 09:09 AM

You need six projection matrices - one for each side of cube. And thus also 6 lights-pace matrices - each orienting negative Z to the direction of corresponding side.

You can still render shadow map in one pass using instancing or geometry shaders, but there is no way to avoid separate projection matrices. Except, of course, dual paraboloid shadow map, but this is quite different issue.

Which method (rendering in 6 passes or using geometry shaders) is actually faster depends AFAIK on GPU and driver versions. If you have frustum culling enabled the amount of geometry rendered is not too different. Doing 6 sides in one pass may actually be bad for memory locality, as all your triangles are each time guaranteed to touch all cubemap sides - but I do not know, how much this may be issue on modern GPU-s..

### #4862857normal and wind order

Posted by on 17 September 2011 - 11:52 AM

Normally if you already have triangle normal you do not need to do any other back-face/front face testing - if normal is towards you, triangle is front-faceing and vice versa.

Now, if you want to know the winding order of triangle ABC, you have to calculate the cross product AB X AC. If this is oriented towards you (can be checked with dot product), triangle if CCW.

PARTNERS