[yt-users] Interactive Volume Rendering Screenshot
Matthew Turk
matthewturk at gmail.com
Wed Jan 3 08:44:28 PST 2018
I think this would be straightforward; it would use a colormap like MIP
does elsewhere. Currently the ProjectionSampler class in
image_samplers.pyx is the only place that we do the projections in the
volume renderer. We would have to make a new one and have it set
im.rgba[i] = max(im.rgba[i], vc.data[i][di]) rather than doing the
accumulation. I am looking now and don't see where we would have to change
the blending in the python code, but I suspect it's somewhere. My guess is
this would be a handful of lines of code to add MIP in software VR.
-Matt
On Wed, Jan 3, 2018 at 10:35 AM, Nathan Goldbaum <nathan12343 at gmail.com>
wrote:
> We should also look into adding support for MIP projections for off-axis
> projections with the software volume renderer. It seems like it should be
> doable although I bet it will require some poking around in the volume
> renderer's internals since I doubt there's a way to paramaterize a MIP
> projection in terms of a transfer function.
>
> On Wed, Jan 3, 2018 at 10:32 AM, Matthew Turk <matthewturk at gmail.com>
> wrote:
>
>>
>> Hi Nicole,
>>
>> On Tue, Jan 2, 2018 at 3:28 PM, Nicole DyAnn Melso <ndm2126 at columbia.edu>
>> wrote:
>>
>>> Hello,
>>>
>>>
>>>
>>> I am attempting to create a series of volume rendered images of my
>>> simulation at different time steps so that I can compile them into a movie.
>>> I want to use the maximum intensity fragment shader, which I believe is
>>> only available in the yt interactive volume rendering with OpenGL (I've
>>> looked into simply using a maximum intensity projection, but this is only
>>> available on-axis). I would like to be able to take a screenshot of the
>>> OpenGL window created when I run the yt interactive volume rendering, but
>>> I'm having difficulty adding this feature to the volume rendering source
>>> code in my local yt directory. Below is the modification I’ve made to the
>>> code as well as a screenshot of the problem. I have no prior experience
>>> with OpenGL.
>>>
>>>
>>>
>>
>> Ah! This looks pretty close, I think. What might be happening is a
>> strangeness with how the bytes are coming back, and maybe even
>> double-writing based on the start/end and which buffer is being read. I
>> can't quite say for sure.
>>
>> But! Since you're experimenting with the OpenGL stuff anyway, I would
>> suggest checking out the branch for interactive VR that I have in an open
>> pull request:
>>
>> https://github.com/yt-project/yt/pull/1452
>>
>> This isn't *completely* working, but it *should* work precisely for what
>> you're doing. The API is considerably different, but also much simpler.
>> This is an example script that gets your started:
>>
>> import yt
>> import yt.visualization.volume_rendering.interactive_vr as ivr
>> ds = yt.testing.fake_amr_ds()
>>
>> data_source = ds.sphere("c", 0.5)
>>
>> ivr.SceneGraph.from_ds(data_source, ("index", "x"))
>>
>> This will dump you into an IPython shell that has access to variables
>> such as "scene" which has an "image" property. I'm hoping to have the time
>> to finish this up in the next month or so and get it accepted upstream in
>> yt.
>>
>> This also has the ability to do offscreen rendering, but the API is
>> (currently) slightly more complicated. Hope that helps!
>>
>> -Matt
>>
>>
>>> */yt/visualization/volume_rendering/interactive_vr.py*
>>>
>>>
>>>
>>> 846 def _retrieve_framebuffer(self):
>>>
>>> 847 ox, oy, width, height = GL.
>>> glGetIntegery(GL.GL_VIEWPORT)
>>>
>>> 848 + my_buffer = GL.glReadBuffer(GL.GL_FRONT)
>>>
>>> 849 - debug_buffer = GL.glReadPixels(0, 0, width,
>>> height, GL.GL_RGB, GL.GL_UNSIGNED_BYTE)
>>>
>>> 849 + debug_buffer = GL.glReadPixels(0, 0, width,
>>> height, GL.GL_RGB, GL.GL_UNSIGNED_BYTE, my_buffer)
>>>
>>> 850 arr = np.fromstring(debug_buffer, “unit8”, count,
>>> width*height*3)
>>>
>>> 851 return arr.reshape((width, height, 3))
>>>
>>>
>>>
>>>
>>>
>>> */yt/visualization/volume_rendering/interactive_loop.py*
>>>
>>>
>>>
>>> 187 def __call__(self, scene, camera, callbacks):
>>>
>>> 188 while not glfw.WindowShouldClose(self.window) or
>>> self.should_quit:
>>>
>>> 189 callbacks(self.window)
>>>
>>> 190 if callbacks.draw:
>>>
>>> 191 camera.compute_matrices()
>>>
>>> 192 scene.set_camera(cameara)
>>>
>>> 193 scene.render()
>>>
>>> 194
>>> glfw.SwapBuffers(self.window)
>>>
>>> 195 callbacks.draw = False
>>>
>>> 196 + arr =
>>> scene._retrieve_framebuffer()
>>>
>>> 197 + write_bitmap(arr,
>>> “test.png”)
>>>
>>> 198 glfw.PollEvents()
>>>
>>> 199 yield self
>>>
>>> 200 glfw.Terminate()
>>>
>>>
>>> The simulation is a single rectangular box, but the screenshot I get
>>> with the above code is multiple boxes that look like they should be
>>> overlaid? Maybe there is a simpler method of accomplishing this? I'd
>>> appreciate any suggestions.
>>>
>>> Thanks,
>>> Nicole Melso
>>>
>>> --
>>> Nicole Melso
>>> NSF Graduate Research Fellow
>>> Columbia Astronomy Department
>>>
>>> _______________________________________________
>>> yt-users mailing list
>>> yt-users at lists.spacepope.org
>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>
>>>
>>
>> _______________________________________________
>> yt-users mailing list
>> yt-users at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>
>>
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-users-spacepope.org/attachments/20180103/ee50fe49/attachment-0002.html>
More information about the yt-users
mailing list