The limit for a single sculpty?

I was wondering what the limit on 'seperate' objects from one sculpty is?

I saw one that appeared to be I believe 20 stones yesterday.. in Avilion, as part of a pathway.

I'm using the cylinder that DanielFox uses in one of his tutorials atm to attempt making a mutlitude of 'seperate' objects, though I'm also testing the limits on movement of sections (crossing the thin thread faces and such, positioning the 'seperate' pieces in variuos places).

But waht is the most possible, or at least most done that anyone knows of?

I'll be attempting various things with sculpties along these lines over the next few days, and was wondering if anyone knew.

trial 1 fails =P

11 'seperate' pieces, with some crossing, random positioning

trial 2 success. 9 'seperate' pieces, no crossing, no scaling to a bounding box within wings (The first was, I'm not sure if that caused my problem, I'll be checking a few things later)

I didn't even bother to move the ends of each 'individual' piece to make them just cylinders.. just pinched and exported.

Created this 1-prim bookshelf/cupboard today.
It's 1 sculpty with 8 'seperate' planks that I moved / rotated and scaled around.
Didnt find the limits yet though

i have nothing to add to the thread just to say

you were in avilion and didn't come say hello? dissapointed

-Crunch

i was standing next to you while you were talking to a Lord and Lady rofl..was a dark fae

(I think Aminom Marvin was the first one that showed me/discovered those invi sculpties I call them?!)

lol next time slap me across the back of the head so i notice you

alright...so how can i do it with 3dsmax?

where is the tutorial? Would be really helpful, i was told how it works, but not how to do it.

The nominal answer is 8 separate (non-touching) parts from a sculpted mesh.

A sculpty is comprised of a 32x32 mesh. Each separate piece requires 2 nodes plus two rows of the sculpty mesh. (The nodes force the connection between pieces to be a 1 dimensional line, which doesn't render thus making the parts visibly disconnected).

Thus 4 of 32 mesh rows per piece, 32/4 = 8 separate pieces.

To be more complete:

There are some ways to cheat a bit. Use a spherical topology so you can skip the nodes on the ends, make one or two "separate" pieces share a node (thus still touching at a point or edge), or use only one non-node mesh row per part (if you are only making "ropes" or rungs).

To do what? Make a single object look like multiple?

Here's a quick run-down of the basic method:

Scale two rows of vertices down so far that they are not visible in a shaded view. The sections on either side of those two rows now look seperate as a sculpty when imported losslessly.

Really, that's pretty much it. You just scale sections small enough to make a very very very thin (Basically invisible) line from one section to the next.

My new record for the moment, 80 separate (non-touching) kind of triangles-shapes in 1 sculpty.

(Used it for folliage next, for some alien-trees, testing on the betagrid)

basic; Use 1point-poles inbetween, to get seperate kind of objects in 1 sculptie. The ?-mark above my head is 1 sculpty, between the dot and top questionmark. I scaled those rings so small/thin it became a 1point-line.

advanced; connect your shapes together as a 1point-line, makes you able to bypass using poles. (the bottom of this tri-shape here is a 1point-line, y/z the same. X has distance)

How you do it in 3D max. well for me its more a bit of sculptypaint/and a lot of photoshop work(checking the points), once i got that shape/sculpty, i can us it next to rotate/scale the points that are free, and not on this pole-line.

Very nice, Cel, seems abstractly organic.

Is that essentially a 3d asterisk? As in all of those are connected at on central point? Looks like it but not quite sure.

That relies on some assumptions about how it is made. Here's 25 non-touching parts in one sculpty. (It would have been 26, but I glitched one of the joins.) I'm pretty sure this is not the most possible.

Well I tried 88 non-touching parts, but somehow I could not use the 33th row to work for a plane-type sculpty, so made it 80 non-touching triangles.


Older: 48rocks in 1 sculpty.


>> Is that essentially a 3d asterisk?
>> As in all of those are connected at on central point? Looks like it but not quite sure.

I used the 80-triangle sculpty I created, rotated the points and scaled them. to make a kind of flower/3D folliage.

I thought that was how it was done, and i did it that way i thought...i basically took 2 rows of vertices, and centered them in 3d space. 0, 0, z. Leaving the Z value where it was. So i saw a line on the screen, but when i brought it into SL, there is still a thin line connecting both shapes.

so setting all vertices down to 0 on the x and y, for both rows, did not do it...as far as i could tell

Make sure you upload losslessly, and allow it to rez fully. Non-lossless doesn't have the detail accuracy needed for this technique, and lower LOD's than the best will show the line.

For the lossless, I suggest SLImageUpload instead of the SL uploader, as it works better at this point.

SLimage upload was the key, thank you

I gather the 80 triangular pieces are each 2-dimensional, right? Is 48 your record while keeping each piece 3-dimensional?

summary of stuff you may or may not know already:
Sculpt type plane is a square grid of 32x33 vertexes. (other sculpt types are 32x32 vertexes only, they ignore the last lowest row of the image)
Vertexes are set by an image, UV is set by xy position on the image, 3d position set by rgb of the image pixel there.
Vertexes are connected horizontally, vertically and diagonally from bottom left to top right.

All COLINEAR polygons are skipped by the renderer (w-poly friendly), they are not klickable and only shown when selected in edit mode.
To achieve colinear polygons you require accurate sculptmaps that can place 2 vertexes on the same xyz coordinates (good generator without rounding errors) and lossless uploading and a not-messed-up viewer.
This allows for nice separations on close up high detail, but makes a mess on the next lower level of detail:

To optimize for the "next lower level of Detail (LOD)" you must know what vertexes are ignored at first (due to downsampling of a 3d shape):
If you start counting your sculpt maps vertex rows and columns at "0", your top left pixel is <o,o> and call its place "x"=<0,0>; row and collum 0, 31 AND 32 are NEVER ignored due to downsampling.
at first lod decrease it ignores all ODD vertexes (except 31, wich is never ignored), thats all that are (x&1): (logical & on integer bitfield).
at second LOD it also ignores all that are (x&2), ignoring (x&3) in total.
at third LOD it also ignores all that are (x&4), ignoring (x&7) in total.
at fourth LOD it also ignores all that are (x& ignoring (x&15) in total.
at fifth LOD it also ignores all that are (x&16), ignoring (x&31) in total. (but still not 31)
LOD downsampling stops here, 32 is also never ignored.

what you make of that is a more complex task.

This isn't quite right. A plane uses 33x33 distinct vertices. You can count them in the attached wireframe view.

The general rule is that every sculpty has 32x32 (x2, counting triangulation) faces. To get this, 32 distinct rows (or columns) of vertices will be used if the columns (or rows) wrap, and 33 otherwise.

>I gather the 80 triangular pieces are each 2-dimensional, right? Is 48 your record while keeping each piece 3-dimensional?

Well the 80triangular pieces are 3-dimensional, not 100% flat. Just imagine what 3D object you can build with four 3D-triangles. I send you a copy of it when i log on.

id rather just count the same-colored squares on my 64x64 pixel image up to 32x33 (witout converging borders) while you tell me where the data for your 33th column is stored.

The 33rd column of vertices should be stored in the 64th column of pixels. This is assuming the programmer of your converter is aware of this special case.

Cool. At first, that just didn't seem plausible. But now I see how to do 88 with 32x33 vertices and 80 with 32x32.

There are still some "wasted" vertices. I'll bet we still haven't found the max.