Andrea Tartaro wrote:
>
> If you can send me your data, it would be very useful for me to see what
> it is doing. If you take out the zero-area triangles, the strip would
> break making tri strips even shorter. If you want more information on
> your strip length distributions, the StripifierStats class has a method
> toString() which will print out a distribution of strip lengths (eg. how
> many strips are length 1, 2, 3, 4, 5, 6, 7, 8, 9, 10-19, 20-49, 50-99,
> 100-999 and 1000 or more).
Here is an output for kong.3ds (I got the file from openfx, but I
suppose it is standard 3ds example).
num orig tris: 12472
num orig vertices: 37416
number of strips: 6675
number of vertices: 27488
total tris: 14138
min strip length: 1
max strip length: 101
avg strip length: 2.1180524344569287
avg num verts/tri: 1.9442636865186023
total time: 851
strip length distribution:
1=6230 2=17 3=3 4=30 5=11 6=43 7=15 8=20 9=5 10-19=133
20-49=154
50-99=13 100-999=1 1000 or more=0
Original number of tris is ok. But after stripification stats panel
shows
Triangles: 20813
Average strip length: 4.1180525
Which seems a bit different from stripifier stats output. I got the
output by adding print in scenegraph/traverser/GeometryArrayStripifier.
If we will look at printed output, stripifiers seems to do quite nice
job in finding long strips, but at the same time leaves horrible amount
of free floating triangles. But it might be a problem with model - maybe
triangles are not exactly next to another ?
With other file (this funny man in a hat)
num orig tris: 14313
num orig vertices: 42939
number of strips: 7665
number of vertices: 31866
total tris: 16536
min strip length: 1
max strip length: 125
avg strip length: 2.1573385518591
avg num verts/tri: 1.9270682148040639
total time: 1091
strip length distribution:
1=6865 2=121 3=30 4=69 5=35 6=63 7=50 8=49 9=37 10-19=215
20-49=116
50-99=10 100-999=5 1000 or more=0
This time number of tris is a bit too low (it was 15092) - probably some
are in different tree. After stripification in panel there is following
data
Triangles: 24201
Average length: 4.1573386
If you want I can send you 3ds file by private mail (please contact me
directly if you are interested to not clutter the maillist).
Artur
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
