 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Jim Kress wrote:
>
> It depends on how you want to create the df3 file. there have been several
> posts in the povray newsgroups about utilities that create a df3 file from
> various types of images.
>
> I've written some code to take a 3D array of x,y,z data and from that
> directly create a df3 file. If anyone is interested, I can clean it up and
> post a copy in the binaries.utilities group.
>
> --
> Jim
That would of course be appreciated but I think that the one thing
on many peoples minds at this point is what exactly should a .df3 file
contain ? What does it do and how can a person control it to make the
media conform to a shape at will ?
If you can offer even a simple overview of what can be gained and where
to start you would be doing a lot of people here a big favour. My self
included. I have some vague notions and have explored the field a bit but
my own results to date have not yet been satisfactory and I owe that mostly
to lack of sufficient input from a knowledgeable source. I found the docs
sorely lacking on this subject for some reason.
--
Ken Tyler
mailto://tylereng@pacbell.net
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Fri, 16 Apr 1999 11:57:01 -0700, Ken <tyl### [at] pacbell net> wrote:
> That would of course be appreciated but I think that the one thing
>on many peoples minds at this point is what exactly should a .df3 file
>contain ? What does it do and how can a person control it to make the
>media conform to a shape at will ?
Have you ever found yourself wishing you could create a 3D bitmap?
Perhaps something simple, like alternating octahedra and tetrahedra
in a nice lattice. Or something complex, perhaps a swirly pattern
that can't be modeled with POV's built-in procedural textures. Say
you have a nice algorithm to generate it but the only language you
know is Visual Basic, so patching POV is out of the question.
Well, a .df3 file is just that: a 3D bitmap. When you use one, it's
just like an imagemap (well... more like Nathan's image_pattern patch,
in that it works with color_map) except that it varies in all three
dimensions instead of just two.
You're the artist here; you tell me how useful such a thing is. :)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Ron Parker wrote:
>
> On Fri, 16 Apr 1999 11:57:01 -0700, Ken <tyl### [at] pacbellnet> wrote:
> > That would of course be appreciated but I think that the one thing
> >on many peoples minds at this point is what exactly should a .df3 file
> >contain ? What does it do and how can a person control it to make the
> >media conform to a shape at will ?
>
> Have you ever found yourself wishing you could create a 3D bitmap?
> Perhaps something simple, like alternating octahedra and tetrahedra
> in a nice lattice. Or something complex, perhaps a swirly pattern
> that can't be modeled with POV's built-in procedural textures. Say
> you have a nice algorithm to generate it but the only language you
> know is Visual Basic, so patching POV is out of the question.
>
> Well, a .df3 file is just that: a 3D bitmap. When you use one, it's
> just like an imagemap (well... more like Nathan's image_pattern patch,
> in that it works with color_map) except that it varies in all three
> dimensions instead of just two.
>
> You're the artist here; you tell me how useful such a thing is. :)
I have as yet insuffecient data to perform a meaningful task.
I am not arguing it's potential for the artist I am crying for more
imput on how to use it. I even to some degree understand the 3d image
file concept but don't know where to start in it's implementation.
For example the program tga2df3. If I wanted to create a simple shape
let's pick something simple like a narrow cylinder with a sphere on top.
Your basic tootsi pop shaped object. What kind of tga images would I
feed tga2df3 to realize this shape when applied in a density statement ?
Give me fundamentals I will worry about concepts later. I need buttons
to push not fancy words to ponder. Give me a damn monkey wrench so I
can start to work with it.
Buzz ! Click !
--
Ken Tyler
mailto://tylereng@pacbell.net
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
You kind of answered your own question, the object you exampled with is
what you would model and slice into thin planar sections (along the z
axis most likely, using a box for CSGing let's say) as frame output.
Then use those in tga2df3.
I'm not one to explain them myself so I'll stop right there. Gee, I
thought you were able to do this kind of thing already though, you are
quite a capable person.
Ken wrote:
>
> Ron Parker wrote:
> >
> > On Fri, 16 Apr 1999 11:57:01 -0700, Ken <tyl### [at] pacbellnet> wrote:
> > > That would of course be appreciated but I think that the one thing
> > >on many peoples minds at this point is what exactly should a .df3 file
> > >contain ? What does it do and how can a person control it to make the
> > >media conform to a shape at will ?
> >
> > Have you ever found yourself wishing you could create a 3D bitmap?
> > Perhaps something simple, like alternating octahedra and tetrahedra
> > in a nice lattice. Or something complex, perhaps a swirly pattern
> > that can't be modeled with POV's built-in procedural textures. Say
> > you have a nice algorithm to generate it but the only language you
> > know is Visual Basic, so patching POV is out of the question.
> >
> > Well, a .df3 file is just that: a 3D bitmap. When you use one, it's
> > just like an imagemap (well... more like Nathan's image_pattern patch,
> > in that it works with color_map) except that it varies in all three
> > dimensions instead of just two.
> >
> > You're the artist here; you tell me how useful such a thing is. :)
>
> I have as yet insuffecient data to perform a meaningful task.
>
> I am not arguing it's potential for the artist I am crying for more
> imput on how to use it. I even to some degree understand the 3d image
> file concept but don't know where to start in it's implementation.
> For example the program tga2df3. If I wanted to create a simple shape
> let's pick something simple like a narrow cylinder with a sphere on top.
> Your basic tootsi pop shaped object. What kind of tga images would I
> feed tga2df3 to realize this shape when applied in a density statement ?
>
> Give me fundamentals I will worry about concepts later. I need buttons
> to push not fancy words to ponder. Give me a damn monkey wrench so I
> can start to work with it.
>
> Buzz ! Click !
>
> --
> Ken Tyler
>
> mailto://tylereng@pacbell.net
--
omniVERSE: beyond the universe
http://members.aol.com/inversez/homepage.htm
mailto:inv### [at] aolcom?Subject=PoV-News
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Fri, 16 Apr 1999 12:35:18 -0700, Ken <tyl### [at] pacbellnet> wrote:
>For example the program tga2df3. If I wanted to create a simple shape
>let's pick something simple like a narrow cylinder with a sphere on top.
>Your basic tootsi pop shaped object. What kind of tga images would I
>feed tga2df3 to realize this shape when applied in a density statement ?
You feed it cross-sections, I assume (not having used the program.)
For the Tootsie pop, you'd feed it a bunch of identical images with
small circles (for the stick) and then a series of circles getting
larger and larger, then reducing again to nothing (for the candy).
Picture what one of those egg-slicer gadgets you can get at the local
kitchen store would do to your lollipop, and you'll have a pretty good
idea what you need. Or, if you wanted to make a file for a human, try
this URL for a (somewhat graphic) depiction of cross-sections:
http://www.npac.syr.edu/projects/vishuman/VisibleHuman.html
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Yes i would like to create a DF3 using XYZ data rather than Bit maps!.
I've experimented with TGA2DF3 using an
animation of a sphere with a CSG Moving down the sphere to create a
sliced effect but still getting odd results.
If i could model the density it would make life much less frustrating.
Thank you Everyone who "tried" to answer such a difficult question,
Nathan Zwalsh
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Fri, 16 Apr 1999 13:21:07 -0500, Bob Hughes <inv### [at] aolcom>
wrote:
>I couldn't help but notice the "CSG-able" word said. I wanted a DF3 of a
>complex model once and came up with mixed results. Basically it didn't
>work in other words (although it almost did it seemed), and I repeatedly
>got a "patch object not allowed" during parse as I tried in vain to get
>the model sliced up right.
>So I guess this is impossible? No work-arounds? Well, I'm not giving up.
As I understand it your complex model was made up of patches, most
probably using moray or spatch. You are using windoze 95 (at least for
newsreading) so here's a windoze-only workaround: from spatch, export
your model to some polygon format (ex. vrml). Then, convert that to a
pov mesh. Then use uvpov to slice up the mesh since it treats meshes
as csg-able.
There is another problem with this method and I have not yet found an
universal solution. Since the plane (or box or whatever) used for
chopping gives the resulting surface its own texture, the result is
either a monochrome .df3 or a black .df3 :) (if no texture is given to
the plane). Suppose you've modelled a nice blob with separate element
textures, and you want to see the inside of it. No way, Jose. Sorry.
That's the reason I went a little off topic and gave up trying to
model a mushroom cloud density with blobs (the original poster's
question). Modelling is easy, but making the .df3 out of it is a pain.
There's a method that works for csg unions, though. Instead of slicing
the union with the plane, slice each part of it separately and texture
it afterwards. This may be quite a tedious work, but a macro might
help, like:
// suppose your cutting plane is caused sliceObj
#macro Object ( myObj )
intersection
{
SliceObj
myObj
#
Invoke it like this:
Object ( niceCSGableObj ) texture { coolObjectGuts } }
This will save... um, some time ... at least.
Man, this *is* an ugly macro, I know, but it should work.
Now, I summon all gurus: gather around and try to make up a way to
show the cross-section of a blob keeping its internal textures intact.
Here are some suggestions to try out:
Use a bounding (not clipping!) object that incorrectly bounds the
blob.
Use a weird orthographic camera setting (some combination of location,
look_at and direction which is beyond my skills) so that the screen
plane is *inside* the object. Will this work?
// object fits the unit sphere
camera
{
location -4*z
look_at 0
direction 4*z
orhographic
}
One last question: how about slicing objects with media and getting
correct results?
I've asked these questions before, but now 1) I am narrowing them down
2) there are more povers around and 3) pov skills increase with time
I can't test the stuff right now as there's already a line waiting at
the computer (I hope dad doesn't notice the slowdown... I put POV in
the system tray and then intentionally hung explorer, so the tray now
only holds a clock and a volume control :) ).
---------
Peter Popov
ICQ: 15002700
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 16 Apr 1999 14:50:40 -0500, par### [at] my-dejanewscom (Ron Parker)
wrote:
>If you have the Superpatch you could use trace() to do what you're wanting
>to do. In fact, you could write a macro that takes an object, a filename,
>and the desired resolution and automatically creates the df3 file for you.
>It wouldn't be the fastest thing on Earth, but it wouldn't be too terrible.
With all due respect to you, sir, and your SuperPatch, I think this
macro will be only suitable for a minimal share of all .df3
applications. If I have gotten your idea correctly, it will not be
able to either reproduce the internal structure of the desired object
nor the texture / interior. See my reply to Bob Hughes for more
details.
---------
Peter Popov
ICQ: 15002700
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
See inserted replies, please:
Peter Popov wrote:
>
> As I understand it your complex model was made up of patches, most
> probably using moray or spatch. You are using windoze 95 (at least for
> newsreading) so here's a windoze-only workaround: from spatch, export
> your model to some polygon format (ex. vrml). Then, convert that to a
> pov mesh. Then use uvpov to slice up the mesh since it treats meshes
> as csg-able.
Yep, there is one little sPatch piece in it, used 6 times. Thanks for
the work-around idea however it already is a bicubic_patch anyway, not
the same thing? (I know very little of these things). Though it was
messing up before getting to any of them. It looked like I had a decent
slice going at first with a plane, I think it was, then maybe messed up
at the bicubics. Keeping textures and all from what I could tell. I was
doing quick tests at low resolution. Then upon some changes, using a box
instead, I failed to get anywhere near the same result.
> There is another problem with this method and I have not yet found an
> universal solution. Since the plane (or box or whatever) used for
> chopping gives the resulting surface its own texture, the result is
> either a monochrome .df3 or a black .df3 :) (if no texture is given to
> the plane). Suppose you've modelled a nice blob with separate element
> textures, and you want to see the inside of it. No way, Jose. Sorry.
> That's the reason I went a little off topic and gave up trying to
> model a mushroom cloud density with blobs (the original poster's
> question). Modelling is easy, but making the .df3 out of it is a pain.
Yeah, I know what you mean, but I know I had colors of the original
model showing properly at first, as I said :)
> There's a method that works for csg unions, though. Instead of slicing
> the union with the plane, slice each part of it separately and texture
> it afterwards. This may be quite a tedious work, but a macro might
> help, like:
>
> // suppose your cutting plane is caused sliceObj
>
> #macro Object ( myObj )
> intersection
> {
> SliceObj
> myObj
> #
>
> Invoke it like this:
>
> Object ( niceCSGableObj ) texture { coolObjectGuts } }
>
> This will save... um, some time ... at least.
>
> Man, this *is* an ugly macro, I know, but it should work.
>
> Now, I summon all gurus: gather around and try to make up a way to
> show the cross-section of a blob keeping its internal textures intact.
> Here are some suggestions to try out:
>
> Use a bounding (not clipping!) object that incorrectly bounds the
> blob.
I did try clipping with a box inversed. And was using orthographic
camera. I don't quite follow your "inside" idea, however 'direction no
longer applies (nor 'angle') with this camera type, only 'up' and
'right' do.
> Use a weird orthographic camera setting (some combination of location,
> look_at and direction which is beyond my skills) so that the screen
> plane is *inside* the object. Will this work?
>
> // object fits the unit sphere
>
> camera
> {
> location -4*z
> look_at 0
> direction 4*z
> orhographic
> }
>
> One last question: how about slicing objects with media and getting
> correct results?
I do hope you mean slicing an object that has media, instead of use a
object with media to slice up another object.
> I've asked these questions before, but now 1) I am narrowing them down
> 2) there are more povers around and 3) pov skills increase with time
>
> I can't test the stuff right now as there's already a line waiting at
> the computer (I hope dad doesn't notice the slowdown... I put POV in
> the system tray and then intentionally hung explorer, so the tray now
> only holds a clock and a volume control :) ).
Bad boy!
> ---------
> Peter Popov
> ICQ: 15002700
Well, trial and error always prevails. It's that error part that bugs me
most.
--
omniVERSE: beyond the universe
http://members.aol.com/inversez/homepage.htm
mailto:inv### [at] aolcom?Subject=PoV-News
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nathan Zwalsh wrote:
>
> I need help in creating DF3 files to make a Mushroom Cloud caused by an
> Nuclear Explosion i've tried to experiment with the DF3's but no success
> on creating any thing close to what i wanted or expected. If anyone
> knows "ANYTHING" about the elusive DF3's please tell me. I'm still
> quite new to pov ( maybe one year under my belt ) so, please try to
> explain it to me as if i were a 2 year old.
>
> Thank you for your time reading this,
>
> Nathan Zwalsh
> nrz### [at] fatnetnet
Hi,
I am not altogether certain that a .df3 density file is the ultimate
solution to making a mushroom cloud. I am about to post a quick preview
of my own efforts towards rendering a reasonable facsimile of a mushroom
cloud. The image will be in the binaries.images group along with the source.
I have no more time to work on it but it might point you or someone else
in the right direction of research. Personaly I think I have made good
headway in the 1 hours time I had available to work on the problem.
Cheers,
--
Ken Tyler
mailto://tylereng@pacbell.net
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
