[vtkusers] vtkVolumeTextureMapper3D in VTK 7.1?
Elvis Stansvik
elvis.stansvik at orexplore.com
Wed Mar 8 13:35:02 EST 2017
Den 8 mars 2017 2:31 em skrev "Simon ESNEAULT" <simon.esneault at gmail.com>:
>
> Same here, for the still rendering I use the minimal spacing value in all
3 directions divided by 2 (something to respect Shannon theorem but I'm not
sure if it's relevant here ...)
> During interaction I ended up hacking
vtkOpenGLGPUVolumeRayCastMapper::ComputeReductionFactor() method to compute
a sampling distance that really honor the wanted FPS. That make it easier
to ensure a target FPS
>
> Here is the diff from VTK 7.1 release
Thanks for sharing. In my case what I found was that the default
computation was too aggressive in reducing the quality. At least it
looked/felt like that.
I also know that adding support for respecting the image sample distance is
on the radar for the VTK devs (see recent post by me on vtk-developers, I'm
on my phone so don't have it handy atm).
Elvis
>
>
//------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
> ---
C:/VTK-7.1.0/Rendering/VolumeOpenGL2/vtkOpenGLGPUVolumeRayCastMapper.cxx
Mon Nov 14 19:58:01 2016
> +++
C:/VTK-7.1.0/Rendering/VolumeOpenGL2/vtkOpenGLGPUVolumeRayCastMapper.cxx
Wed Mar 08 14:21:13 2017
> @@ -2116,26 +2116,9 @@
> else
> {
> input->GetSpacing(this->CellSpacing);
> - vtkMatrix4x4* worldToDataset = vol->GetMatrix();
> - double minWorldSpacing = VTK_DOUBLE_MAX;
> - int i = 0;
> - while (i < 3)
> - {
> - double tmp = worldToDataset->GetElement(0,i);
> - double tmp2 = tmp * tmp;
> - tmp = worldToDataset->GetElement(1,i);
> - tmp2 += tmp * tmp;
> - tmp = worldToDataset->GetElement(2,i);
> - tmp2 += tmp * tmp;
> -
> - // We use fabs() in case the spacing is negative.
> - double worldSpacing = fabs(this->CellSpacing[i] * sqrt(tmp2));
> - if(worldSpacing < minWorldSpacing)
> - {
> - minWorldSpacing = worldSpacing;
> - }
> - ++i;
> - }
> + // Approximate Shannon -> half the miniaml spacing
> + double minWorldSpacing = std::min(std::min(this->CellSpacing[0],
> + this->CellSpacing[1]),this->CellSpacing[1])
/ 2.;
>
> // minWorldSpacing is the optimal sample distance in world space.
> // To go faster (reduceFactor<1.0), we multiply this distance
> @@ -3113,7 +3096,6 @@
>
> if ( this->TimeToDraw )
> {
> - double oldFactor = this->ReductionFactor;
>
> double timeToDraw;
> if (allocatedTime < 1.0)
> @@ -3137,30 +3119,17 @@
> timeToDraw = 10.0;
> }
>
> - double fullTime = timeToDraw / this->ReductionFactor;
> - double newFactor = allocatedTime / fullTime;
> -
> - // Compute average factor
> - this->ReductionFactor = (newFactor + oldFactor)/2.0;
> -
> - // Discretize reduction factor so that it doesn't cause
> - // visual artifacts when used to reduce the sample distance
> - this->ReductionFactor = (this->ReductionFactor > 1.0) ? 1.0 :
> - (this->ReductionFactor);
> -
> - if (this->ReductionFactor < 0.20)
> + if ( timeToDraw == BigTimeToDraw )
> {
> - this->ReductionFactor = 0.10;
> + this->ReductionFactor = 1;
> }
> - else if (this->ReductionFactor < 0.50)
> - {
> - this->ReductionFactor = 0.20;
> - }
> - else if (this->ReductionFactor < 1.0)
> + else
> {
> - this->ReductionFactor = 0.50;
> + double l_wanted_fps = 1. / ( allocatedTime * 3. );
> + double l_current_fps = 1. / timeToDraw;
> + this->ReductionFactor *= l_current_fps / l_wanted_fps;
> + this->ReductionFactor = vtkMath::Round( this->ReductionFactor *
50. ) / 50.;
> }
> -
> // Clamp it
> if ( 1.0/this->ReductionFactor > this->MaximumImageSampleDistance )
> {
>
//------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
>
> Simon
>
> 2017-03-08 14:04 GMT+01:00 Elvis Stansvik <elvis.stansvik at orexplore.com>:
>>
>> 2017-03-08 11:16 GMT+01:00 Simon ESNEAULT <simon.esneault at gmail.com>:
>> > Hello Alvaro,
>> >
>> > That's real good news to hear (about the down-sampling in X and Y), I
find
>> > it really difficult to tweak volume rendering only with sample
distance,
>> > especially during interaction.
>> > Has the development began already ? Is there some branche where we can
test
>> > that feature ?
>> >
>> > Also in that case would it be possible to add a final "gaussian
smoothing
>> > pass" or equivalent only on the 2D generated image (maybe only as an
option)
>> > ? I look like it is the case with FixedPointVolumeraycastMapper, the
image
>> > looks a bit blurred but sometimes it helps to see the structures
>>
>> I'm also very interested in this. I'm using vtkGPUVolumeRayCastMapper,
>> and I'm currently turning off the automatic adjustment of sample
>> distance during interaction and resorted to doing my own simpler
>> adjustment. I'm simply quadrupling the sample distance during
>> interaction (for still rendering I use
>> (spacingX+spacingY+spacingZ)/6). I'd love to have another quality
>> controlling knob to turn.
>>
>> Is there a MR up for it already? Will it be in 7.2?
>>
>> Elvis
>>
>> >
>> > Thanks
>> > Simon
>> >
>> >
>> > 2017-03-06 18:17 GMT+01:00 Alvaro Sanchez <alvaro.sanchez at kitware.com>:
>> >>
>> >> Hi Shark,
>> >>
>> >> unfortunately there is no current plan of porting
vtkVolumeTextureMapper3D
>> >> to OpenGL2
>> >> , as far as I know. We will soon however support image down sampling
in X
>> >> and Y which
>> >> should serve as additional knobs to tweak the performance of
>> >> GPURayCastMapper in
>> >> OpenGL2 (only sample distance can be currently adjusted).
>> >>
>> >> Could you provide more detail on the size of the dataset you are
rendering
>> >> and your GPU
>> >> specs?
>> >>
>> >> Thanks,
>> >> Álvaro
>> >>
>> >> On Mon, Mar 6, 2017 at 10:30 AM, Shark <m.nunes at fratoria.com> wrote:
>> >>>
>> >>> Hello,
>> >>>
>> >>> I have been having some issues with vtkVolumeTextureMapper3D. I see
it
>> >>> has
>> >>> been deprecated, however it is still available if I use the old
openGL.
>> >>>
>> >>> Since for software rendering with texture mapping is much faster
than ray
>> >>> casting, I would like to know if you are planning to have a Volume
>> >>> Texture
>> >>> Mapper for openGL 2, or if there is actually a new one that I am not
able
>> >>> to
>> >>> find.
>> >>>
>> >>> Best regards,
>> >>> Shark
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> View this message in context:
>> >>>
http://vtk.1045678.n5.nabble.com/vtkVolumeTextureMapper3D-in-VTK-7-1-tp5742368.html
>> >>> Sent from the VTK - Users mailing list archive at Nabble.com.
>> >>> _______________________________________________
>> >>> Powered by www.kitware.com
>> >>>
>> >>> Visit other Kitware open-source projects at
>> >>> http://www.kitware.com/opensource/opensource.html
>> >>>
>> >>> Please keep messages on-topic and check the VTK FAQ at:
>> >>> http://www.vtk.org/Wiki/VTK_FAQ
>> >>>
>> >>> Search the list archives at: http://markmail.org/search/?q=vtkusers
>> >>>
>> >>> Follow this link to subscribe/unsubscribe:
>> >>> http://public.kitware.com/mailman/listinfo/vtkusers
>> >>
>> >>
>> >>
>> >>
>> >> --
>> >> Alvaro Sanchez
>> >> Kitware, Inc.
>> >> Senior R&D Engineer
>> >> 21 Corporate Drive
>> >> Clifton Park, NY 12065-8662
>> >> Phone: 518-881-4901
>> >>
>> >> _______________________________________________
>> >> Powered by www.kitware.com
>> >>
>> >> Visit other Kitware open-source projects at
>> >> http://www.kitware.com/opensource/opensource.html
>> >>
>> >> Please keep messages on-topic and check the VTK FAQ at:
>> >> http://www.vtk.org/Wiki/VTK_FAQ
>> >>
>> >> Search the list archives at: http://markmail.org/search/?q=vtkusers
>> >>
>> >> Follow this link to subscribe/unsubscribe:
>> >> http://public.kitware.com/mailman/listinfo/vtkusers
>> >>
>> >
>> >
>> >
>> > --
>> > ------------------------------------------------------------------
>> > Simon Esneault
>> > Rennes, France
>> > ------------------------------------------------------------------
>> >
>> > _______________________________________________
>> > Powered by www.kitware.com
>> >
>> > Visit other Kitware open-source projects at
>> > http://www.kitware.com/opensource/opensource.html
>> >
>> > Please keep messages on-topic and check the VTK FAQ at:
>> > http://www.vtk.org/Wiki/VTK_FAQ
>> >
>> > Search the list archives at: http://markmail.org/search/?q=vtkusers
>> >
>> > Follow this link to subscribe/unsubscribe:
>> > http://public.kitware.com/mailman/listinfo/vtkusers
>> >
>
>
>
>
> --
> ------------------------------------------------------------------
> Simon Esneault
> Rennes, France
> ------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtkusers/attachments/20170308/b53c0370/attachment.html>
More information about the vtkusers
mailing list