So I had a similar symptom a few days ago and discovered that I was so overexposing the scene so severely, everything was washed out. That is not the case this time.
In this case, I have a complex archviz interior. I placed extensive light fixtures containing Octane lights. The scene renderers, but without the light or the fixtures showing. Why would that be? The light fixtures themselves are not luminous. There is also daylight via portals. That appears to be the only source of rendered illumination.
I am stumped.
Whole or partial scene not rendering
Forum rules
Please post only in English in this subforum. For alternate language discussion please go here http://render.otoy.com/forum/viewforum.php?f=18
Please post only in English in this subforum. For alternate language discussion please go here http://render.otoy.com/forum/viewforum.php?f=18
Please post your images.gbambo wrote:So I had a similar symptom a few days ago and discovered that I was so overexposing the scene so severely, everything was washed out. That is not the case this time.
In this case, I have a complex archviz interior. I placed extensive light fixtures containing Octane lights. The scene renderers, but without the light or the fixtures showing. Why would that be? The light fixtures themselves are not luminous. There is also daylight via portals. That appears to be the only source of rendered illumination.
I am stumped.
I would be happy to share the scene.
But I do not think it is light power, as I used a black-body emitter and defined power as 121W per fluorescent tube. There are three tubes in each luminaire and the luminaire reflects off a celing only a foot above. I did extensive render testing outside of the scene before importing the fixture.
Remember, the fixture itself is not rendering at all, despite the rest of the scene rendering fine. I do not understand how one object can not render when every other object does render.
What does "scene scale is too large" mean? Are you referring to the scene units, which are decimal feet, or to the size of the scene measured in the scene units, which is circa 50 x 75 feet?
When I tried to render an aerial shot earlier today, I got a soft off-white frame that was evenly filled except for the vignetting at the edges. But there were no objects rendered. Yet I had rendered the same scene with ever all geometry exposed from interior angles immediately before hand. That resulted in about 1700mb of vram usage. The aerial shot, at the same resolution, with much of the geometry hidden, consumed 3900mb of vram, which may explain the featureless frame. But given that it was the same scene, why was memory usage so different?
At first I thought it might be owing to back face culling, but elsewhere in the forum it is stated that Octane does not use back face culling.
But I do not think it is light power, as I used a black-body emitter and defined power as 121W per fluorescent tube. There are three tubes in each luminaire and the luminaire reflects off a celing only a foot above. I did extensive render testing outside of the scene before importing the fixture.
Remember, the fixture itself is not rendering at all, despite the rest of the scene rendering fine. I do not understand how one object can not render when every other object does render.
What does "scene scale is too large" mean? Are you referring to the scene units, which are decimal feet, or to the size of the scene measured in the scene units, which is circa 50 x 75 feet?
When I tried to render an aerial shot earlier today, I got a soft off-white frame that was evenly filled except for the vignetting at the edges. But there were no objects rendered. Yet I had rendered the same scene with ever all geometry exposed from interior angles immediately before hand. That resulted in about 1700mb of vram usage. The aerial shot, at the same resolution, with much of the geometry hidden, consumed 3900mb of vram, which may explain the featureless frame. But given that it was the same scene, why was memory usage so different?
At first I thought it might be owing to back face culling, but elsewhere in the forum it is stated that Octane does not use back face culling.
4-core 6gb | Win 7 x64 | nVidia GeForce GTX 670 4gb | 3ds Max Design 2013 | OctaneRender Plugin 1.0 Beta 2.58e Kepler build
Can you send it to [email protected]?gbambo wrote:I would be happy to share the scene.
Thank you.
I am especially confused at the ballooning memory requirement. Without doing any work on the scene I went from less than 2gb vram to render an interior shot with all geometry set to visible and then, with at least a third of the geometry hidden, it was consuming 4gb on an exterior shot. No changes in materials. No changes in render resolution. And a dramatic decrease in the number of lights from dozens + daylight, to just daylight.
I am especially confused at the ballooning memory requirement. Without doing any work on the scene I went from less than 2gb vram to render an interior shot with all geometry set to visible and then, with at least a third of the geometry hidden, it was consuming 4gb on an exterior shot. No changes in materials. No changes in render resolution. And a dramatic decrease in the number of lights from dozens + daylight, to just daylight.
4-core 6gb | Win 7 x64 | nVidia GeForce GTX 670 4gb | 3ds Max Design 2013 | OctaneRender Plugin 1.0 Beta 2.58e Kepler build
So I just retried the aerial render (view = top). I did this because I saw a post about a memory bug that causes crashes when rendering sequences of individual frames at once. I had such a crash before my failed aerial render.
So I restarted 3ds and retried. This time Scene Statistics showed memory use as P = 1.9gb and V = 4.4gb. Previously those numbers were P = 1.0gb and V = 4.0gb. This is really weird. what do P and V stand for? I though V was vram, but the context makes me doubt it. P is system memory? How could I use 4.4gb of memory on a 4gb GTX 670?
The render result was a frame uniformly filled nearly black.
So I restarted 3ds and retried. This time Scene Statistics showed memory use as P = 1.9gb and V = 4.4gb. Previously those numbers were P = 1.0gb and V = 4.0gb. This is really weird. what do P and V stand for? I though V was vram, but the context makes me doubt it. P is system memory? How could I use 4.4gb of memory on a 4gb GTX 670?
The render result was a frame uniformly filled nearly black.
4-core 6gb | Win 7 x64 | nVidia GeForce GTX 670 4gb | 3ds Max Design 2013 | OctaneRender Plugin 1.0 Beta 2.58e Kepler build
V - virtualgbambo wrote:So I just retried the aerial render (view = top). I did this because I saw a post about a memory bug that causes crashes when rendering sequences of individual frames at once. I had such a crash before my failed aerial render.
So I restarted 3ds and retried. This time Scene Statistics showed memory use as P = 1.9gb and V = 4.4gb. Previously those numbers were P = 1.0gb and V = 4.0gb. This is really weird. what do P and V stand for? I though V was vram, but the context makes me doubt it. P is system memory? How could I use 4.4gb of memory on a 4gb GTX 670?
The render result was a frame uniformly filled nearly black.
P - phisical.
both are not GPU memory.
For what it's worth, these are my hard learned lessons on this subject. Mostly they seem obvious ... until they happened.
Ways that an object or objects fail, or appear to fail, to render.
Wild under exposure
Wild over exposure
Intervening transparent object with broken material so it previews as translucent, but renders as opaque
Rendering a scene from a ortho viewport, while Octane only supports perspective views
In this circumstance, the perspective view picked by the renderer is not always pointed in the same direction as the viewport
which is counter to expectation, since, when switching between ortho and perspective, 3ds Max chooses views that are aimed the same but differ only in zoom
In the layers menu, the render column for the object(s) is unchecked
GPU runs out of memory (uncertain what happens when only texture limit is violated)
A frame not be displayed in the viewport is dependent on path constraint that has broken
for example, path constraint was applied to a group containing camera and target, which is then ungrouped to allow adjusting the settings of the camera
A path or object not meant to render, but set to preview in the viewport, which is the subject of a constrained camera, is mistakenly set to render and is located in front of the camera.
Hopefully, this will help someone else out of a bind!
Additions to this list are welcome.
Ways that an object or objects fail, or appear to fail, to render.
Wild under exposure
Wild over exposure
Intervening transparent object with broken material so it previews as translucent, but renders as opaque
Rendering a scene from a ortho viewport, while Octane only supports perspective views
In this circumstance, the perspective view picked by the renderer is not always pointed in the same direction as the viewport
which is counter to expectation, since, when switching between ortho and perspective, 3ds Max chooses views that are aimed the same but differ only in zoom
In the layers menu, the render column for the object(s) is unchecked
GPU runs out of memory (uncertain what happens when only texture limit is violated)
A frame not be displayed in the viewport is dependent on path constraint that has broken
for example, path constraint was applied to a group containing camera and target, which is then ungrouped to allow adjusting the settings of the camera
A path or object not meant to render, but set to preview in the viewport, which is the subject of a constrained camera, is mistakenly set to render and is located in front of the camera.
Hopefully, this will help someone else out of a bind!
Additions to this list are welcome.
4-core 6gb | Win 7 x64 | nVidia GeForce GTX 670 4gb | 3ds Max Design 2013 | OctaneRender Plugin 1.0 Beta 2.58e Kepler build