As you can see by the photo I've attached there is bad clipping on my stalactites, yet in the preview there exists no such problem.
Anyone have any ideas how to correct it???
Also, is there ever going to be official documentation? I know there are tutorials and quickie lessons and the wiki - YET - no where are the limits of the input fields discussed. Fox example, Scale - does it range from negative ( < 0) infinity to positive infinity, or is there a set range. I seem to spend a lot of time having to save at each step because T2 keeps crashing when I try to experiment and end up entering a value somewhere that T2 did not like... It is very frustrating.
Thanks,
JR
Looks like you're using pretty extreme displacements here.
This can stress the renderer a lot in terms of huge rendertimes or incorrect results, like you show here.
I guess you used a second planet for creating the stalactites? You may try increasing the displacement tolerance from 1 to 1.5 or 2 and then see what happens.
You might even go to 3, if necessary, but wouldn't go any further, because rendertimes will become huge.
Martin
Where is the displacement "tolerance"? I've looked through all the settings and can not find "tolerance".
Actually it is a small plane with a power fractal applied to it and flipped upside down to create the stalactites. As far as the power fractal settings they are as follows:
Scale: -5000, Lead-in: 6.25, Noise octaves 19
Displacement applied Along normal, Amplitude 5719, offset 15, roughness .75, spike limit 5
I did find a way to get rid of the clipping, but it involved moving the camera, which is not what I wanted to do.
Thanks,
JR
The displacement tolerance setting is in the plane object-node, at the bottom.
Looking at the values for your displacements it doesn't surprise me that you have this problem.
Big scales, many octaves, high displacement factor and roughness...
So I think you should try upping the displacement tolerance to 2 for starters.
A better way of "fixing" this is to choose different settings for the fractal. Let's say you want 1m wide stalactites, then use 1m as feature scale and a couple of meter as lead-in scale and something like 0.1m as smallest. It will probably not give problems with rendering but also will be faster (lower number of octaves).
I see, I need to think even smaller. Thanks for the help!!!!
JR
Tangled-Universe meant to say "planet" object :)
No, actually TangledUniverse was right, I'm using a plane and I found the tolerance on the plane object. I kept looking at displacement settings on the power fractal, which is why I could not find it originally. Not even sure what it does, I'll have to look that one up...
Thanks for all your help!
JR
Try this clipfile if you like.
It's a 10m plane with a powerfractal and surfacelayer attached to it.
I couldn't get the plane rotated upside down, so I'll leave that to you ;)
Could you explain how you did that? I haven't found the logic of the vector values yet.
I'm glad I'm not the only one that does not fully understand the plane coordinates...
These are the values I used to flip it upside down:
Edge vector a: 1, 0, 0
Edge vector b: 0, 0, -1
I read something the other day, I think by Mr. Lamppost which I never knew, only one side of the plane is truly visible, so thus the need to flip it upside down for the ceiling of my cave.
I'm going to take a peak at your clip file - Thanks!!
JR
Negating the positive in either of the vector values of the plane will flip it.
Either of these values will do for inverting it; '-1,0,0/0,0,1' or '-1,0,0/0,0,-1'.
Rotating it off-centre is more tricky: A value of '1' or '-1' in the centre field(Y?) seems to rotate it by 45 degrees. You can then keep upping this value to '90' or '-90' to make it vertical. I get a bit confused over this area but it does work. I don't understand why 1 makes it tilt half as much as 90. Weird.
It must be some Euclidean geometry that we have never heard of...
@Dandel0: I only use a 1 to get it on its side, but forgot which one. No 90.
By the way, you might be able to do some interesting stuff with 2 planes, one flipped, both displaced.....
I use one of the edge vectors y coordinate, changing that to one (1) seems to flip on side. 2 planes would be interesting, right now one is a challenge. I applied a water shader to the plane then displaced it on the density channel with a power fractal. I'm getting some interesting stalactite formations - still trying to get the lighting right though...
JR
Give me a wee while and I'll post an example of this rotation thing. I cannot make a plane vertical by using a value of 1...
You are right, a value of 1 in either of the 'y' won't bring it to 90 deg - my mistake - had to rotate it around a bit to see the angle. Did find out values 1 thru 10 incrementally bring up the plane to roughly 85 degrees. Values 11 thru 19 don't seem to do much judging by angular position and shadows cast, finally I don't seem to see much difference between a value of 20 thru 90 in the 'y' field.
I would like to see the algorithm behind this. It is also annoying that I can't select a plane to manipulate it like other objects. The 'x' and 'z' vectors make sense if you envision a cube centered at 0,0,0 with each side a length of 1. Then you can figure out which way to position the plane in relation to faces on a cube.
Please correct me if anyone has a better explanation of the plane object.
To get around the limitations of the native plane object, I created a cube object in PoseRay and gave it just a minimal of height (for effects), so basically it is a plane - Imported it into T2 and makes life a lot easier. :)
JR
I don't understand why you don't get a plane on its side with a '1'. See this example. Just set the first row as follows (0-1-0), and it flips. It's a 4x8 meter wall with an image projected on it, by the way, hence the black sides.
---Dune
It looks like it flips to a perfect 90 degrees when viewed at that angle - same mistake I made. It is not at a true 90 degrees - when you revolve around it you will see that it is only around 70 degrees.
I see where I'm going wrong now, Ulco.
I'd been leaving the values at defaults and setting only the second 'vector a' at '1', when that's the case, you get the result of an exact 45 degrees when using these settings...
Vector A = 1,1,0
Vector B = 0,0,1
That's just changing one value from the defaults(and the reverse of this does the same but backwards, obviously).
Adding increments up to '90' to the middle 'vector a' parameter rotates this from 45 degrees through 90 degrees but, only when the first value is left at '1'. This is kind of confusing.
Zeroing these is the answer, so that only one parameter is applying rotation. Thank you. :)
The rotation values are quite tricksy on a plane object, you can stretch and skew them, aswell, along the X and Z's, this takes them off of square.
this is OT but...
maybe i'm seeing things,
Quote from: Dune on May 14, 2010, 02:42:58 AM
I don't understand why you don't get a plane on its side with a '1'. See this example. Just set the first row as follows (0-1-0), and it flips. It's a 4x8 meter wall with an image projected on it, by the way, hence the black sides.
---Dune
is that a Wireframe preview i see on those bounding boxes???
Quote from: zaai999 on May 14, 2010, 08:28:42 PM
is that a Wireframe preview i see on those bounding boxes???
No, it's just an optical illusion. It's really a cat jumping out of a bag! :o :D
cat's out of the bag on that one... :D :D :D
who puts cats in bags anyways? cats hate bags.. oh wait- Matt puts cats in bags ;D
Quote from: dandelO on May 14, 2010, 09:17:15 PM
Quote from: zaai999 on May 14, 2010, 08:28:42 PM
is that a Wireframe preview i see on those bounding boxes???
No, it's just an optical illusion. It's really a cat jumping out of a bag! :o :D
Yeah, it's wireframe. I guess I'm very priviliged as an alpha tester. But patience, thou time will come... But I will tell you this, it's a GREAT innovation of this already fantastic product! I couldn't have done without it in the things I'm working on right now.
well i'm glad its a reality now, this feature will be helpful in the future, i figured it would look something like that.