Adam Hauldren wrote:
> I believe the test I did was between different quality settings of 3DO,
> not comparative between different render programs.

   Hmm. I think I recall we compared them against Lightwave at various
anti-aliasing
settings as well, and we found that LW with no AA was faster, since we always
antialias.
As I recall, when you turned LW up to a comparable quality of AA, we were
faster.

> I suspect that he was referring to the generation of shadow maps for
> large numbers of 3DOs, which can add considerably to render time. In the
> case of cached shadow maps in scenes with static lighting, this only
> occurs once, but still represents a time-consuming process.

   Well, true, but I don't think any other program can do better. You can't
efficiently
raytrace shadows with the amount of geometry we typically render, and that many
shadowmaps
are going to take a bunch of time to render in any program.

> or two in various programs if that would help - I suspect you'd find
> that VNS was slower to render raw 3DO geometry but came into its own
> when dealing with large numbers of foliage instances and large terrain
> areas.

   I think my statement would be "When you compare apples to apples, WCS/VNS
should be no
slower than anything else."

   I just don't want a statement like "VNS renders 3D objects slow" to go
unchallenged
here, if we're not talking about a valid comparison.

> Regards
> Adam

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
                Highway Design: http://highway-design.net/

Chris 'Xenon' Hanson wrote:
> Adam Hauldren wrote:
>> I believe the test I did was between different quality settings of 3DO,
>> not comparative between different render programs.
>
> Hmm. I think I recall we compared them against Lightwave at various
> anti-aliasing settings as well, and we found that LW with no AA was
> faster, since we always antialias. As I recall, when you turned LW
> up to a comparable quality of AA, we were faster.

Okay - I can't find that original message thread in my archives. I
distinctly remember doing tests on the render speed of different 3DO
settings within VNS...

>> I suspect that he was referring to the generation of shadow maps for
>> large numbers of 3DOs, which can add considerably to render time. In the
>> case of cached shadow maps in scenes with static lighting, this only
>> occurs once, but still represents a time-consuming process.
>
> Well, true, but I don't think any other program can do better. You
> can't efficiently raytrace shadows with the amount of geometry we
> typically render, and that many shadowmaps are going to take a bunch
> of time to render in any program.
>
>> or two in various programs if that would help - I suspect you'd find
>> that VNS was slower to render raw 3DO geometry but came into its own
>> when dealing with large numbers of foliage instances and large terrain
>> areas.
>
> I think my statement would be "When you compare apples to apples, WCS/VNS
> should be no slower than anything else."
>
> I just don't want a statement like "VNS renders 3D objects slow" to go
> unchallenged here, if we're not talking about a valid comparison.

Agreed.

That said, without doing some comparative renders, I'm concerned that
I'll come across as a fanboy ;)

Adam Hauldren wrote:
> Chris 'Xenon' Hanson wrote:
>> Adam Hauldren wrote:
>>> I believe the test I did was between different quality settings of 3DO,
>>> not comparative between different render programs.
>> Hmm. I think I recall we compared them against Lightwave at various
>> anti-aliasing settings as well, and we found that LW with no AA was
>> faster, since we always antialias. As I recall, when you turned LW
>> up to a comparable quality of AA, we were faster.
>
> Okay - I can't find that original message thread in my archives. I
> distinctly remember doing tests on the render speed of different 3DO
> settings within VNS...
>
>>> I suspect that he was referring to the generation of shadow maps for
>>> large numbers of 3DOs, which can add considerably to render time. In the
>>> case of cached shadow maps in scenes with static lighting, this only
>>> occurs once, but still represents a time-consuming process.
>> Well, true, but I don't think any other program can do better. You
>> can't efficiently raytrace shadows with the amount of geometry we
>> typically render, and that many shadowmaps are going to take a bunch
>> of time to render in any program.
>>
>>> or two in various programs if that would help - I suspect you'd find
>>> that VNS was slower to render raw 3DO geometry but came into its own
>>> when dealing with large numbers of foliage instances and large terrain
>>> areas.
>> I think my statement would be "When you compare apples to apples, WCS/VNS
>> should be no slower than anything else."
>>
>> I just don't want a statement like "VNS renders 3D objects slow" to go
>> unchallenged here, if we're not talking about a valid comparison.
>
> Agreed.
>
> That said, without doing some comparative renders, I'm concerned that
> I'll come across as a fanboy ;)

Another good point is that other rendering engines are often
multithreaded, giving the impression that they are faster to render,
whilst leveraging the power of more than one CPU to get that speed increase.

Adam Hauldren wrote:
> Another good point is that other rendering engines are often
> multithreaded, giving the impression that they are faster to render,
> whilst leveraging the power of more than one CPU to get that speed increase.

   Ah well, there is that. Sigh.

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
                 3D Earth: http://3d-earth.net/

What if I use Superconducter, run from a machine with Windows XP, to control
renders on a PC with Vista64? Is there any reason I wouldn't be able to take
advantage of the faster machine and render that way?

Will WCS ever be upgraded, or is that now a dead product to future operating
systems? What about Windows 7? It's kind of a steep price to upgrade to VNS just
to be able to run on a newer OS.

Thanks,

David


--- In wcsml@yahoogroups.com, Chris 'Xenon' Hanson <xenon@...> wrote:
>
> davifi21 wrote:
> > I just installed WCS 6 on a new machine with Vista (64), and some of the
buttons, for instance the crucial "Go" button on the Render dialog box. I saw
there were some posts about this a few months ago. Has there been any solution
or bug fix for this problem?
> > Is there any way to start a render besides hitting that "Go" button?
>
>   WCS 6 and VNS 2 are not compatible with Vista64, but will work with Vista32
or XP64.
> It's a side effect of some dark magic Vista64 uses to implement the Vista
Desktop Window
> Manager in a 64-bit thunked environment. Basically, Vista64 sends messages
from a source
> that appears to be a different control class than what they're really coming
from.
> WCS6/VNS2 relied on detecting the origin of events in order to decide how to
process them
> (was it a pushbutton or checkbox or scrollbar?). Unfortunately, the blockage
here is
> within WCS6 code -- the button pushes, but the event is not interpreted by
WCS6, so any
> method you can imagine that could potentially cause the same action will still
get
> disregarded by WCS6's event handler.
>
>   If you can't upgrade to VNS 3 and have to run WCS 6 on Vista64, you might be
able to run
> it under a virtualized WinXP using VMWare, VirtualBox or the like.
>
> > thanks,
> > David Fierstein
>
> --
>      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
>            3D Landscape Design: http://3d-landscape-design.com/
>

--- In wcsml@yahoogroups.com, Chris 'Xenon' Hanson <xenon@...> wrote:
>
> Adam Hauldren wrote:
> > Another good point is that other rendering engines are often
> > multithreaded, giving the impression that they are faster to render,
> > whilst leveraging the power of more than one CPU to get that speed increase.
>
>   Ah well, there is that. Sigh.
>
> --
>      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
>                 3D Earth: http://3d-earth.net/
>

Sorry I've missed the discussion.

Adam was correct that the speed takes a hit when generating shadow maps. I was
doing a render with a landsat image draped over a SRTM DEM. The scene had less
the 20 3d objects. I ended up aborting the 3d object render because it was
taking so long.

davifi21 wrote:
> What if I use Superconducter, run from a machine with Windows XP, to control
renders on a PC with Vista64? Is there any reason I wouldn't be able to take
advantage of the faster machine and render that way?

   I think that will work, yes.

> Will WCS ever be upgraded, or is that now a dead product to future operating
systems? What about Windows 7? It's kind of a steep price to upgrade to VNS just
to be able to run on a newer OS.

   That is undecided at this point. Very few people use WCS these days, so
obviously that's
where we are focusing. I haven't tried Win7-64 but I expect it will have the
same problem
as Vista-64.

> Thanks,
> David

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
            3D Landscape Design: http://3d-landscape-design.com/

>ll WCS ever be upgraded, or is that now a dead product to future
>operating systems? What about Windows 7? It's kind of a steep price
>to upgrade to VNS just to be able to run on a newer OS.
>
>   That is undecided at this point. Very few people use WCS these
> days, so obviously that's
>where we are focusing. I haven't tried Win7-64 but I expect it will
>have the same problem
>as Vista-64.

Kind of been waiting for this to come up. It will be sad if WCS just
drifts away for new OS's. My teaching lab cannot afford multiple
seats for VNS as much as I would love to have them (my annual lab
budget is $4700. and dropping with the annual cuts). Also cannot
afford to maintain a lab of XP machines just to run WCS. If this
happens, instruction in the 3dNature family will drop from UHM.

ev

Ev Wingert wrote:
> Kind of been waiting for this to come up. It will be sad if WCS just
> drifts away for new OS's. My teaching lab cannot afford multiple
> seats for VNS as much as I would love to have them (my annual lab
> budget is $4700. and dropping with the annual cuts). Also cannot
> afford to maintain a lab of XP machines just to run WCS. If this
> happens, instruction in the 3dNature family will drop from UHM.

   Remember, WCS 6 does work under Vista, just not Vista-64. We still haven't
made a final
decision, but right now, we don't sell enough WCS units to make working on WCS
worth the
time. We'll see what happens in the near future.

> ev

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
                 3D Earth: http://3d-earth.net/

UHM is bypassing Vista altogether and waiting to see what happens with 7.


>   Remember, WCS 6 does work under Vista, just not Vista-64. We
> still haven't made a final
>decision, but right now, we don't sell enough WCS units to make
>working on WCS worth the
>time. We'll see what happens in the near future.

I certainly understand the economics of a small business, sharing my
home office with one, but its still sad to see a really powerful
render engine move out of reach of students.

e


> > ev
>
>--
>      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
>                 3D Earth: http://3d-earth.net/
>
>
>------------------------------------
>
>---------------------------------------------------------------------------
>Subscribe/UnSubscribe, Online
>Archives:  http://www.groups.yahoo.com/group/wcsml/
>Yahoo! Groups Links
>
>
>

billisenberger wrote:
> --- In wcsml@yahoogroups.com, Chris 'Xenon' Hanson <xenon@...> wrote:
>> Adam Hauldren wrote:
>>> Another good point is that other rendering engines are often
>>> multithreaded, giving the impression that they are faster to render,
>>> whilst leveraging the power of more than one CPU to get that speed increase.
>>   Ah well, there is that. Sigh.
>>
>> --
>>      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
>>                 3D Earth: http://3d-earth.net/
>>
>
> Sorry I've missed the discussion.
>
> Adam was correct that the speed takes a hit when generating shadow maps.
> I was doing a render with a landsat image draped over a SRTM DEM. The
> scene had less the 20 3d objects. I ended up aborting the 3d object
> render because it was taking so long.

Hmmm...well in defense of the software, that may very well have been due
to your choice of shadow settings.

I can set up pretty much any piece of rendering software such that it
takes an absolute age to render, regardless of how fast the render
engine is meant to be ;)

In those sorts of situations, always worth posting to the list - there
are a number of ways in which shadow map generation can be speeded up.
I've rendered scenes with hundreds of 3DOs casting shadows and not had
rendertimes become unmanageable.

Regards

Adam

--- In wcsml@yahoogroups.com, Adam Hauldren <adam@...> wrote:
>
> billisenberger wrote:
> > --- In wcsml@yahoogroups.com, Chris 'Xenon' Hanson <xenon@> wrote:
> >> Adam Hauldren wrote:
> >>> Another good point is that other rendering engines are often
> >>> multithreaded, giving the impression that they are faster to render,
> >>> whilst leveraging the power of more than one CPU to get that speed
increase.
> >>   Ah well, there is that. Sigh.
> >>
> >> --
> >>      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
> >>                 3D Earth: http://3d-earth.net/
> >>
> >
> > Sorry I've missed the discussion.
> >
> > Adam was correct that the speed takes a hit when generating shadow maps.
> > I was doing a render with a landsat image draped over a SRTM DEM. The
> > scene had less the 20 3d objects. I ended up aborting the 3d object
> > render because it was taking so long.
>
> Hmmm...well in defense of the software, that may very well have been due
> to your choice of shadow settings.
>
> I can set up pretty much any piece of rendering software such that it
> takes an absolute age to render, regardless of how fast the render
> engine is meant to be ;)
>
> In those sorts of situations, always worth posting to the list - there
> are a number of ways in which shadow map generation can be speeded up.
> I've rendered scenes with hundreds of 3DOs casting shadows and not had
> rendertimes become unmanageable.
>
> Regards
>
> Adam
>

Hi Adam, don't get me wrong. I love VNS for working with large GIS data sets. If
you could share the settings that can be tweaked for rendering 3d object shadows
I would greatly appreciate it (or point me to where I can find more information
about them). When I was working on the project with the 3d objects as mentioned
above, I sent emails to 3D Nature support describing the scene and the slow
rendering and asked what would be the best setup for my scenario, but never got
a response. Again, I would very much appreciate any guidance you could give me
that would up the performance of rending shadows for 3d objects in VNS. For what
I was trying to do, VNS was definitely the best solution until I ran into issue
with the 3d object shadows.

--- In wcsml@yahoogroups.com, "billisenberger" <billisenberger@...> wrote:
>
> --- In wcsml@yahoogroups.com, Adam Hauldren <adam@> wrote:
> >
> > billisenberger wrote:
> > > --- In wcsml@yahoogroups.com, Chris 'Xenon' Hanson <xenon@> wrote:
> > >> Adam Hauldren wrote:
> > >>> Another good point is that other rendering engines are often
> > >>> multithreaded, giving the impression that they are faster to render,
> > >>> whilst leveraging the power of more than one CPU to get that speed
increase.
> > >>   Ah well, there is that. Sigh.
> > >>
> > >> --
> > >>      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
> > >>                 3D Earth: http://3d-earth.net/
> > >>
> > >
> > > Sorry I've missed the discussion.
> > >
> > > Adam was correct that the speed takes a hit when generating shadow maps.
> > > I was doing a render with a landsat image draped over a SRTM DEM. The
> > > scene had less the 20 3d objects. I ended up aborting the 3d object
> > > render because it was taking so long.
> >
> > Hmmm...well in defense of the software, that may very well have been due
> > to your choice of shadow settings.
> >
> > I can set up pretty much any piece of rendering software such that it
> > takes an absolute age to render, regardless of how fast the render
> > engine is meant to be ;)
> >
> > In those sorts of situations, always worth posting to the list - there
> > are a number of ways in which shadow map generation can be speeded up.
> > I've rendered scenes with hundreds of 3DOs casting shadows and not had
> > rendertimes become unmanageable.
> >
> > Regards
> >
> > Adam
> >
>
> Hi Adam, don't get me wrong. I love VNS for working with large GIS data sets.
If you could share the settings that can be tweaked for rendering 3d object
shadows I would greatly appreciate it (or point me to where I can find more
information about them). When I was working on the project with the 3d objects
as mentioned above, I sent emails to 3D Nature support describing the scene and
the slow rendering and asked what would be the best setup for my scenario, but
never got a response. Again, I would very much appreciate any guidance you could
give me that would up the performance of rending shadows for 3d objects in VNS.
For what I was trying to do, VNS was definitely the best solution until I ran
into issue with the 3d object shadows.
>

For additional information the scene I was working on consisted of 90 meter SRTM
data covering an area approximatly 470x350 kilometers, a Color Map image
(landsat7 mosaic ECW format). I originally had 300 3d objects, then cut this
back to 10 3d objects. The 3d objects were simple elongated boxes with multiple
material colors. I wanted the 3d objects to cast shadows on themselves and on
the terrain. I didn't want the terrain to cast shadows on itself as they were
inherent in the image.

> -----Original Message-----
> From: wcsml@yahoogroups.com [mailto:wcsml@yahoogroups.com]On Behalf Of
> billisenberger
> Sent: Wednesday, September 09, 2009 6:05 AM
> To: wcsml@yahoogroups.com
> Subject: Re: [WCS] For highway landscape simulation, which is better?
> VNS or 3DMax?
>
>
> [munch]
>
> Hi Adam, don't get me wrong. I love VNS for working with large
> GIS data sets. If you could share the settings that can be
> tweaked for rendering 3d object shadows I would greatly
> appreciate it (or point me to where I can find more information
> about them). When I was working on the project with the 3d
> objects as mentioned above, I sent emails to 3D Nature support
> describing the scene and the slow rendering and asked what would
> be the best setup for my scenario, but never got a response...

I know from many years' experience that the 3DN support team places a high
priority on promptly replying to support requests. If you received no
response, they didn't get the messages. If it happens again, post to the
list that you haven't been able to contact them via support@....

Scott

>> Hi Adam, don't get me wrong.

I don't think I did ;)

>> I love VNS for working with large GIS data sets. If you could
> > share the settings that can be tweaked for rendering 3d object
>> shadows I would greatly appreciate it (or point me to where I
>> can find more information about them).

Each project will have its own requirements. But there are some general
issues to observe:

- vector bound shadows with as small a vector area as possible. This
will allow you to reduce the shadow map resolution (speeding generation)
while maintaining a high detail shadow.
- Only generate shadows where they are going to be seen by the camera
- If the camera is not going to approach the shadows too closely, reduce
the resolution of the shadowmaps as low as possible.
- Use shaodw map caching (allowing you to only generate the shadows once)

> For additional information the scene I was working on consisted
> of 90 meter SRTM data covering an area approximatly 470x350
> kilometers, a Color Map image (landsat7 mosaic ECW format).

That's a reasonably large area of terrain - if you were using one shadow
map component for the whole area, your shadows might take some time to
generate and would be extremely low resolution.

> I originally had 300 3d objects, then cut this back to 10 3d
> objects.

Whoa! That's a big change. Some might ask whether the 3DOs were even
relevant at all if you could cull that many from the scene!

How were they placed? Geographic instances, or vector-bound? Or were
they part of an ecosystem?

Also - be aware that some of preset components (like boulder lake and
boulder creek) have large numbers of 3DOs as ecotypes in the beach
materials. These objects are by default set to a higher render quality
than necessary and will render (and generate shadows) regardless of
whether or not they are visible. If you threw a component like that into
the mix, your 300 3DOs would be the least of your problems! ;)

> The 3d objects were simple elongated boxes with multiple material
> colors. I wanted the 3d objects to cast shadows on themselves and
> on the terrain. I didn't want the terrain to cast shadows on
> itself as they were inherent in the image.

Fine - simply create a shadow map component, attach it to vector bounded
areas around each 3DO, and do some tests from the closest camera
viewpoint to make sure that you have sufficient resolution in the
shadowmaps.

The following should be enabled in the shadowmap component:
- Receive from 3D Objects

The following should be enabled for the 3DOs:
- Cast
- Receive from 3D Objects

Quality will have to be determined on a per-project basis by quick
preview renders.

"Use file" should be enabled, unless you are animating the light position.

I think that's all.

Would be interested to know if changing the project settings will allow
you to get faster rendertimes out of it.

I have just created a project here with the same terrain/drape specs as
above, and 100 3DOs all casting shadows. Shadow generation took around
14 seconds on my laptop (T9400@..., 4Gb Ram, WinXP32).

Hope this all helps...

Regards

Adam

--- In wcsml@yahoogroups.com, "Scott Cherba" <wcsm@...> wrote:
>
> > -----Original Message-----
> > From: wcsml@yahoogroups.com [mailto:wcsml@yahoogroups.com]On Behalf Of
> > billisenberger
> > Sent: Wednesday, September 09, 2009 6:05 AM
> > To: wcsml@yahoogroups.com
> > Subject: Re: [WCS] For highway landscape simulation, which is better?
> > VNS or 3DMax?
> >
> >
> > [munch]
> >
> > Hi Adam, don't get me wrong. I love VNS for working with large
> > GIS data sets. If you could share the settings that can be
> > tweaked for rendering 3d object shadows I would greatly
> > appreciate it (or point me to where I can find more information
> > about them). When I was working on the project with the 3d
> > objects as mentioned above, I sent emails to 3D Nature support
> > describing the scene and the slow rendering and asked what would
> > be the best setup for my scenario, but never got a response...
>
> I know from many years' experience that the 3DN support team places a high
> priority on promptly replying to support requests. If you received no
> response, they didn't get the messages. If it happens again, post to the
> list that you haven't been able to contact them via support@...
>
> Scott
>

Thanks Scott,
This is good to know. Whenever I talked to 3D Nature on the phone they were
always very helpful and responsive.
Bill

--- In wcsml@yahoogroups.com, Adam Hauldren <adam@...> wrote:
>
> >> Hi Adam, don't get me wrong.
>
> I don't think I did ;)
>
> >> I love VNS for working with large GIS data sets. If you could
> > > share the settings that can be tweaked for rendering 3d object
> >> shadows I would greatly appreciate it (or point me to where I
> >> can find more information about them).
>
> Each project will have its own requirements. But there are some general
> issues to observe:
>
> - vector bound shadows with as small a vector area as possible. This
> will allow you to reduce the shadow map resolution (speeding generation)
> while maintaining a high detail shadow.
> - Only generate shadows where they are going to be seen by the camera
> - If the camera is not going to approach the shadows too closely, reduce
> the resolution of the shadowmaps as low as possible.
> - Use shaodw map caching (allowing you to only generate the shadows once)
>
> > For additional information the scene I was working on consisted
> > of 90 meter SRTM data covering an area approximatly 470x350
> > kilometers, a Color Map image (landsat7 mosaic ECW format).
>
> That's a reasonably large area of terrain - if you were using one shadow
> map component for the whole area, your shadows might take some time to
> generate and would be extremely low resolution.
>
> > I originally had 300 3d objects, then cut this back to 10 3d
> > objects.
>
> Whoa! That's a big change. Some might ask whether the 3DOs were even
> relevant at all if you could cull that many from the scene!
>
> How were they placed? Geographic instances, or vector-bound? Or were
> they part of an ecosystem?
>
> Also - be aware that some of preset components (like boulder lake and
> boulder creek) have large numbers of 3DOs as ecotypes in the beach
> materials. These objects are by default set to a higher render quality
> than necessary and will render (and generate shadows) regardless of
> whether or not they are visible. If you threw a component like that into
> the mix, your 300 3DOs would be the least of your problems! ;)
>
> > The 3d objects were simple elongated boxes with multiple material
> > colors. I wanted the 3d objects to cast shadows on themselves and
> > on the terrain. I didn't want the terrain to cast shadows on
> > itself as they were inherent in the image.
>
> Fine - simply create a shadow map component, attach it to vector bounded
> areas around each 3DO, and do some tests from the closest camera
> viewpoint to make sure that you have sufficient resolution in the
> shadowmaps.
>
> The following should be enabled in the shadowmap component:
> - Receive from 3D Objects
>
> The following should be enabled for the 3DOs:
> - Cast
> - Receive from 3D Objects
>
> Quality will have to be determined on a per-project basis by quick
> preview renders.
>
> "Use file" should be enabled, unless you are animating the light position.
>
> I think that's all.
>
> Would be interested to know if changing the project settings will allow
> you to get faster rendertimes out of it.
>
> I have just created a project here with the same terrain/drape specs as
> above, and 100 3DOs all casting shadows. Shadow generation took around
> 14 seconds on my laptop (T9400@..., 4Gb Ram, WinXP32).
>
> Hope this all helps...
>
> Regards
>
> Adam
>
  Thanks for the detailed information Adam. 14 seconds that's incredible. When I
get the time I'll go back to the project and try your suggestions. Right now I'm
under the gun on another project. Again thank you very much.
Bill

Hello List,

                 I'm running an animation that comes up to an area where I
have a lot of vegetation.  When using the distance dissolve, there is a
strong contrast between the vegetation and the place in which it has
dissolved.  The result is a wiping effect of vegetation as the camera gets
closer to the focus area.  Would it be possible to allow for a profile for
the distance dissolve?  The effect I'm looking for would allow the dissolve
to "trickle in" the vegetation.  For example, as the camera moves forward
the dissolve would gradually go from 100% to 0%, eliminating the strong
boundary between dissolved vegetation and full vegetation.

                 Is this already achievable and I've just overlooked it?



Thanks.



David Osti

34 North

310-628-9944

david@...





[Non-text portions of this message have been removed]

David Osti wrote:
>                 I'm running an animation that comes up to an area where I
> have a lot of vegetation.  When using the distance dissolve, there is a
> strong contrast between the vegetation and the place in which it has
> dissolved.  The result is a wiping effect of vegetation as the camera gets
> closer to the focus area.  Would it be possible to allow for a profile for
> the distance dissolve?  The effect I'm looking for would allow the dissolve
> to "trickle in" the vegetation.  For example, as the camera moves forward
> the dissolve would gradually go from 100% to 0%, eliminating the strong
> boundary between dissolved vegetation and full vegetation.
>                 Is this already achievable and I've just overlooked it?

   The renderer actually already does use a dissolve zone rather than a hard-edge
disappear
point. The trick with distance dissolve is to make the dissolve color/texture
similar
enough to the undissolved vegetation that the transition is not that noticeable.
If you're
seeing the transition, it means you can distinguish between the two states,
which means
tuning the color or texture might help.

   Maybe show us a picture of the situation?

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
                    3D Terrain: http://3d-terrain.com/

I will try to create a better dissolve texture.  See attached for some
images.



David Osti

34 North

310-628-9944

david@...



From: wcsml@yahoogroups.com [mailto:wcsml@yahoogroups.com] On Behalf Of
Chris 'Xenon' Hanson
Sent: Thursday, September 10, 2009 8:49 AM
To: wcsml@yahoogroups.com
Subject: Re: [WCS] Distance Dissolve for animations





David Osti wrote:
> I'm running an animation that comes up to an area where I
> have a lot of vegetation. When using the distance dissolve, there is a
> strong contrast between the vegetation and the place in which it has
> dissolved. The result is a wiping effect of vegetation as the camera gets
> closer to the focus area. Would it be possible to allow for a profile for
> the distance dissolve? The effect I'm looking for would allow the dissolve
> to "trickle in" the vegetation. For example, as the camera moves forward
> the dissolve would gradually go from 100% to 0%, eliminating the strong
> boundary between dissolved vegetation and full vegetation.
> Is this already achievable and I've just overlooked it?

The renderer actually already does use a dissolve zone rather than a
hard-edge disappear
point. The trick with distance dissolve is to make the dissolve
color/texture similar
enough to the undissolved vegetation that the transition is not that
noticeable. If you're
seeing the transition, it means you can distinguish between the two states,
which means
tuning the color or texture might help.

Maybe show us a picture of the situation?

--
Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
3D Terrain: http://3d-terrain.com/





[Non-text portions of this message have been removed]

David Osti wrote:
> I will try to create a better dissolve texture.  See attached for some
> images.

   You have to upload it to the files section, the list strips attachments.

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
                 SketchUp Components: http://sketchup-components.com/

Hello,

This email message is a notification to let you know that
a file has been uploaded to the Files area of the wcsml
group.

   File        : /Yolo_Rehab_v400990.png
   Uploaded by : freestyle00072000 <david@...>
   Description : Distance Dissolve Example

You can access this file at the URL:
http://groups.yahoo.com/group/wcsml/files/Yolo_Rehab_v400990.png

To learn more about file sharing for your group, please visit:
http://help.yahoo.com/l/us/yahoo/groups/original/general.htmlfiles

Regards,

freestyle00072000 <david@...>

David Osti wrote:
> I will try to create a better dissolve texture.  See attached for some
> images.

   So, in the image, please explain what is dissolving to what.

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
                 3D World Maps: http://3d-world-map.com/

Chris 'Xenon' Hanson <xenon@...> wrote:

> So, in the image, please explain what is dissolving to what.


About a 1/5 of the way down from the top, to the right of the long
straight canal, is where the dissolve is taking place.  It is dissolving
into the aerial image.

David

freestyle00072000 wrote:
> About a 1/5 of the way down from the top, to the right of the long
> straight canal, is where the dissolve is taking place.  It is dissolving
> into the aerial image.

   And so the grey-ish color is the color for that sort of vegetation as seen in
the
airphoto? yeah, they are quite different. I would use a fractal noise texture
using a
couple of colors sampled from the tans and browns of the actual rendered
vegetation.
you'll need to play with the colors and scale of the fractal noise, but it
usually does
the trick.

> David

--
      Chris 'Xenon' Hanson | Xenon @ 3D Nature | http://www.3DNature.com/
                 Virtual Globe: http://virtual-globe.org/

Chris 'Xenon' Hanson <xenon@...> wrote:

>   And so the grey-ish color is the color for that sort of vegetation as seen
in the
> airphoto? yeah, they are quite different. I would use a fractal noise texture
using a
> couple of colors sampled from the tans and browns of the actual rendered
vegetation.
> you'll need to play with the colors and scale of the fractal noise, but it
usually does
> the trick.

Thanks Chris, I'll give that a go.

David

Chris 'Xenon' Hanson wrote:
> freestyle00072000 wrote:
>> About a 1/5 of the way down from the top, to the right of the long
>> straight canal, is where the dissolve is taking place.  It is dissolving
>> into the aerial image.
>
>   And so the grey-ish color is the color for that sort of vegetation as seen
in the
> airphoto? yeah, they are quite different. I would use a fractal noise texture
using a
> couple of colors sampled from the tans and browns of the actual rendered
vegetation.
> you'll need to play with the colors and scale of the fractal noise, but it
usually does
> the trick.

Alternatively, you might want to try tinting your foliage with the
underlying colors of the aerial image. Add a color control attribute to
the foliage images, then turn off "use color", and apply your aerial
image to the "Replace Gray" attribute in the ecosystem editor as a
georeferenced planar image.

If the colors don't quite match (the foliage is likely to be darker than
the aerial image), you can copy the aerial image and adjust its
saturation and brightness and use the color-corrected image to tint the
foliage.

Assuming you are maintaining the elevation for the camera seen in your
test shot, then this should provide the most accurate color-matching at
the zone of foliage dissolve.

Regards

Adam

--- In wcsml@yahoogroups.com, "billisenberger" <billisenberger@...> wrote:
>
> --- In wcsml@yahoogroups.com, Adam Hauldren <adam@> wrote:
> >
> > >> Hi Adam, don't get me wrong.
> >
> > I don't think I did ;)
> >
> > >> I love VNS for working with large GIS data sets. If you could
> > > > share the settings that can be tweaked for rendering 3d object
> > >> shadows I would greatly appreciate it (or point me to where I
> > >> can find more information about them).
> >
> > Each project will have its own requirements. But there are some general
> > issues to observe:
> >
> > - vector bound shadows with as small a vector area as possible. This
> > will allow you to reduce the shadow map resolution (speeding generation)
> > while maintaining a high detail shadow.
> > - Only generate shadows where they are going to be seen by the camera
> > - If the camera is not going to approach the shadows too closely, reduce
> > the resolution of the shadowmaps as low as possible.
> > - Use shaodw map caching (allowing you to only generate the shadows once)
> >
> > > For additional information the scene I was working on consisted
> > > of 90 meter SRTM data covering an area approximatly 470x350
> > > kilometers, a Color Map image (landsat7 mosaic ECW format).
> >
> > That's a reasonably large area of terrain - if you were using one shadow
> > map component for the whole area, your shadows might take some time to
> > generate and would be extremely low resolution.
> >
> > > I originally had 300 3d objects, then cut this back to 10 3d
> > > objects.
> >
> > Whoa! That's a big change. Some might ask whether the 3DOs were even
> > relevant at all if you could cull that many from the scene!
> >
> > How were they placed? Geographic instances, or vector-bound? Or were
> > they part of an ecosystem?
> >
> > Also - be aware that some of preset components (like boulder lake and
> > boulder creek) have large numbers of 3DOs as ecotypes in the beach
> > materials. These objects are by default set to a higher render quality
> > than necessary and will render (and generate shadows) regardless of
> > whether or not they are visible. If you threw a component like that into
> > the mix, your 300 3DOs would be the least of your problems! ;)
> >
> > > The 3d objects were simple elongated boxes with multiple material
> > > colors. I wanted the 3d objects to cast shadows on themselves and
> > > on the terrain. I didn't want the terrain to cast shadows on
> > > itself as they were inherent in the image.
> >
> > Fine - simply create a shadow map component, attach it to vector bounded
> > areas around each 3DO, and do some tests from the closest camera
> > viewpoint to make sure that you have sufficient resolution in the
> > shadowmaps.
> >
> > The following should be enabled in the shadowmap component:
> > - Receive from 3D Objects
> >
> > The following should be enabled for the 3DOs:
> > - Cast
> > - Receive from 3D Objects
> >
> > Quality will have to be determined on a per-project basis by quick
> > preview renders.
> >
> > "Use file" should be enabled, unless you are animating the light position.
> >
> > I think that's all.
> >
> > Would be interested to know if changing the project settings will allow
> > you to get faster rendertimes out of it.
> >
> > I have just created a project here with the same terrain/drape specs as
> > above, and 100 3DOs all casting shadows. Shadow generation took around
> > 14 seconds on my laptop (T9400@, 4Gb Ram, WinXP32).
> >
> > Hope this all helps...
> >
> > Regards
> >
> > Adam
> >
>  Thanks for the detailed information Adam. 14 seconds that's incredible. When
I get the time I'll go back to the project and try your suggestions. Right now
I'm under the gun on another project. Again thank you very much.
> Bill
>

Adam,
I tried your settings and they made a huge difference (very straightforward I
might add). My 3d objects are elongated square boxes. The side facing the sun
has banding with lower shadow quality settings. Is there a way to deal with the
banding other than the shadow quality?

Here's an image that shows the banding:
http://www.3d-garage.com/forums/VNS/banding.jpg
Thanks for you help, I really appreciate it.
Bill

billisenberger wrote:

> Adam,
> I tried your settings and they made a huge difference

Good to hear.

> (very straightforward I might add).

;)

> My 3d objects are elongated square boxes. The side facing
> the sun has banding with lower shadow quality settings. Is
> there a way to deal with the banding other than the shadow
> quality?
>
> Here's an image that shows the banding:
> http://www.3d-garage.com/forums/VNS/banding.jpg
> Thanks for you help, I really appreciate it.

Just increase the shadow offset value on the shadows page of the 3DO
editor. Your objects are quite large, so I suspect you will need a value
somewhat greater than the default.

Regards

Adam

...and the team at the Manhatta Project:
http://www.themanhattaproject.org

We first saw some of the artwork back in April on the NatGeo site, but I
just received my copy of this month's National Geographic magazine, in
which there's an excellent article on the Manhattan of the 17th Century.
Markley's excellent work is dotted though the article.

Definitely worth grabbing a copy, if you get a chance :)

Best regards

Adam