Tue, 04/08/2008 - 17:12
Forums:
Hi all,
please take a look on this snapshot http://www.benuzzi.net/forum01/zoom.png
Why the quality of surfaces in viewer is bad with high zoom?
There are "holes" and the shadow surface is different than the wireframe surface.
There is a way to control the quality when zooming?
My configuration is:
- Linux Fedora Core 8
- nVidia Quadro NVS 110M
Any suggestion?
Tue, 04/08/2008 - 19:32
Hi, Marco, I am not very familiar with this problem, but I saw it a lot of times. Consider a investigation of these methods of AIS_Shape
OwnDeviationAngle
OwnDeviationCoefficient
OwnHLRDeviationAngle
OwnHLRDeviationCoefficient
SetAngleAndDeviation
SetHLRAngleAndDeviation
SetOwnDeviationAngle
SetOwnDeviationCoefficient
SetOwnHLRDeviationAngle
SetOwnHLRDeviationCoefficient
Also:
AIS_Drawer::HLRAngle
AIS_Drawer::SetHLRAngle
AIS_InteractiveContext::HLRAngle
AIS_InteractiveContext::SetHLRAngle
Regards!
Wed, 04/09/2008 - 12:12
Thanks.
Applying the following settings the quality is really better without any notable performance or memory degration.
myAISContext->SetDeviationAngle(myAISContext->DeviationAngle() / 10);
myAISContext->SetDeviationCoefficient(myAISContext->DeviationCoefficient() / 10);
myAISContext->SetHLRDeviationCoefficient(myAISContext->HLRDeviationCoefficient() / 10);
myAISContext->SetHLRAngle(myAISContext->HLRAngle() / 10);
Wed, 04/09/2008 - 15:52
Yes, the modern graphics hardware is maybe 10-100 times more powerfull since the first version of OpenCascade. I've experienced the same problem as you and possibly will test this code on our project.
Wed, 01/14/2009 - 19:19
I don't agree with you marco:
"any notable performance"
If the setting are following:
SetDeviationAngle(1);
SetDeviationCoefficient(1);
SetHLRDeviationCoefficient(1);
SetHLRAngle(1);
I have a bad visualisation in Shading mode but good one in Wireframe (How can it be ?)
With setting:
SetDeviationAngle(10);
SetDeviationCoefficient(10);
SetHLRDeviationCoefficient(10);
SetHLRAngle(10);
Thje visualization is good, but performance is very bad (much more time to load the Shapes in AIS Context.)
The Graphic Card is fast....
Thu, 01/15/2009 - 09:39
A graphics card being fast is related to its ability to draw more quickly on screen, or to render more frames per second, than other cards or, worse, software emulation.
Creating the triangulation and feeding the graphics card with it is a different story. It's not like you are passing the internal TopoDS_Shape of the AIS_Shape to the graphics card and it'll do all triangulation using its GPU. It's an operation that is still CPU bound.
Thu, 01/15/2009 - 11:31
You are right.
But One question:
I have a good quality of IsoLines, but a bad quality of shading model.
For example a cylinder:
1. With a lot of Iso-Lines I see one cylinder, ever with the very small quality's parameter.
2. In Shading mode I see crooked cylinder. To see a "good" cylinder i have to set a realy big quality's parameter
The different is realy big.
Is it possibly to do something. Is it complett different algorithmus to calculate ISOLines and Triangulation?
Thu, 01/15/2009 - 11:38
Is it deppend on format of file ?
1. To Transfer the file to the Document ?
2. The time, that OCC needs to pass a Shape to Graphic Card ?
Thu, 01/15/2009 - 11:45
The slowdown would occur when OCC has to convert the shape to something that can be used to send to the graphics card. It should happen when OCC computes the presentation of the shape, which happens once for every mode (and for every shape), unless you force a re-computation.
The way it's sent to the graphics card may also hurt performance. I haven't checked how OCC does that.
Thu, 01/15/2009 - 11:39
I haven't checked into it, but maybe, just maybe, some of those parameters may influence a single presentation mode. Another option would be to write your own version of AIS_Shape that uses different parameters when calculating the different presentation modes.
That's all I can think of. I hope that helps.
Thu, 01/15/2009 - 11:47
Thank you !
have a good day
Thu, 01/15/2009 - 22:15
Does somebody know the Program CADMai ?
http://www.cadmai.com/
The program based on OpenCascade Version 5.0.
The CadMai need less than 20% of time to import step fiel and to show the imported shape, than OCC 6.3 with standard methods. The quality is not bad.
Is the Version 5.0 so fast ? Has somebody some expirience with 5.0 ?
Fri, 01/16/2009 - 09:41
I see the logo of OpenGL in their website. It could happen, as some people have already done it, that they are using another visualization library instead of the services provided in OCC. I think the author of HeeksCAD did that (you can find about it in the forums), and I also had a conversation with someone who told me they developed a visualization module using an existing library instead of using OCC's AIS and friends for a project that uses OCC.
Mon, 01/19/2009 - 11:25
After the study of HeeksCAD and some other projects, I have detected in my program that:
The behavior of the transformation of TopoDS_Shape to the Graphic card is a little bit curious :
such way to show a TopoDS_Shape:
...
create TopoDS_Shape and aAISObject
...
aAISObject->SetDisplayMode(1);
theInteractiveContext->m_Context->Display(aAISObject);
leads to bad quality of the shown Shape. I have had to increase the quality factor of the AIS_Interactive context in order to increase the quality->very slow.
BUT
If you load the TopoDS_Shape / ASI_Shape AT FIRST in the Wireframe-mode and AFTER THAT switch it to the Shaded mode:
InteractiveContext.SetDisplayMode(1);
the quality is good and it happens more faster (not very slower than bei HeeksCAD).
What is the reason ? Is it is depend on the mode, how the Shape will be transformed to the GraphicCard ....?