Hi Curt,
I suppose I'm less sensitive to the abrupt change problem as most of the
scenery both face material and triangle edges show some form of abrupt
change due the the mountainous terrain that covers most of the country.
Personally I like the blending solution for edges although this still
doesn't solve the stretching problem. This could possibly be done for both
face material changes and slope changes (smoothed edges)
There seem to be two issues;
a) the stretching on slopes and
b) the seamless edges.
At the moment we have pretty much seamless edges on adjacent tiles of the
same face material. This isn't the case for edges on adjacent tiles of
different face material.
This is a known issue. I suggested a way to try and get a handle on it,
unfortunately this would be at the cost of seamless edges. But this
investigation process doesn't have to effect the production version. Could
this be done locally by someone who can implement different slope alogrithms
and report back findings.
I think the objective data would assist, rather than what appears to be most
subjective assessment albeit very well educated.
I might be way off target here and I hope that I haven't caused offence to
anyone. I realise I'm making rather bold statements for a newbie.
Cheers
Dene
From: "Curtis L. Olson" <[EMAIL PROTECTED]>
dene maxwell wrote:
Hi Paul,
From: Paul Surgeon <[EMAIL PROTECTED]>
On Sunday 15 January 2006 10:25, dene maxwell wrote:
> Hi Paul, to my way of thinking the resolution is not important.
Pythagarus
> is more important, the distance as seen in a birds-eye view as seen by
FGSD
> is not the distance of the terrain. Hence if you cut a material to a
> birds-eye distance of say 10 what ever units it will be stretched to
say
> 14.14 units when the slope is 45 degrees. The greater the slope the
larger
> the slope distortion, in the nth degree a vertical slope will have
zero
> length stretched to infinite length.
>
> Dene
What I mean is that from a "top down" view you need to increase the
resolution
of textures on the slopes so that when you look at the slope from a
perpendicular angle you still get the same resolution as on flat ground.
Example :
If a flat piece of ground has a texture resolution of 1m/pixel then a
slope of
45 degrees should also have a texture resolution of 1m/pixel if you look
at
it from a perpendicular perspective.
agreed
A 45 degree slope viewed from the *top* should have a texture resolution
of
0.707 m/pixel. So from a "top down" view the texture resolution will be
higher. This will "unstretch" those slope textures.
Paul
this certainly seems to follow the geometric logic. One problem...it
doesn't seem to work.
Without being familiar with the code concerned, my impression is that this
logic is not being applied in a way that produces the desired result or
something very important is being over looked in the logic. Perhaps
applying this sinusoidal algorithm is the problem, why not try giving all
perspectives a 1m/pixel view and see if the result is more pleasing? This
might aleast give some idea where the observed stretching is
occurring....yeah?
Here's the downside though ... if you change the texture resolution or how
you project the texture onto the surface based on slope, then you lose the
seamless tiling across triangle edges. We eventually need to come up with
a better way to blend/dither non-tiled edges together, but for the moment
we can't do that.
Regards,
Curt.
_________________________________________________________________
Discover fun and games at @ http://xtramsn.co.nz/kids
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Flightgear-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/flightgear-devel
