A Few Questions regarding ressources

[PdZ]

Dear Chaosgroup,

wow! What a great update! Great Work! Iam impressed!

But now I have some questions regarding V-Ray in General and the Node-Implementation.


1. "old" V-Ray Used for example a Standard Material. In this Standard Material in the Diffuse, Reflection and Glossyness Channel are all loaded the same Texture.
Was "old" V-Ray clever enough not to copy all 3 the same textures to the ram or are they always loaded as copys into the Ram?
2. Is "new" Node-Material with One Bitmap-Node in the aforemention three channels able to use now just one copy of that texture or will it be internally loaded as three copies?
3. Except the mentioned missing gradient implementation: Are there any downsides or missing features using node materials?
4. Long time ago I asked Stefan_Laub about how V-Ray handles assets. Is it true that V-Ray internally uses uncompressed TIFF-Images in at best 1024*1024 Format, I mean in that Byte-Format, and if not, it has to convert (for ex.) Images as Jpegs loaded in the shaders internally to Tifs? So it would be wise to use Tifs instead of any other image-format?

[PdZ]

Also some further questions:
1. If you use a Texture/Bitmap while it is in a TexLayeredShader or Cinema 4D Layer: What is V-Ray internally doing? Is it converting the whole Layered-Element to a .TIFF internally, send it to the V-Ray-Renderer so it can use it?
So when using Layer it means V-Ray has always to convert the whole Element so it can use it?

I think I have to test some things now!

[deyan.hadzhiev]

Hi PdZ, here are some answers on your questions:

1. "old" V-Ray Used for example a Standard Material. In this Standard Material in the Diffuse, Reflection and Glossyness Channel are all loaded the same Texture.
Was "old" V-Ray clever enough not to copy all 3 the same textures to the ram or are they always loaded as copys into the Ram?
2. Is "new" Node-Material with One Bitmap-Node in the aforemention three channels able to use now just one copy of that texture or will it be internally loaded as three copies?

In case the used texture file is the same, V-Ray always reuses the handles and the textures are shared and allocate memory only once. The benefit from the Bitmap node in the node system, is that you can now change the additional parameters in the bitmap from a single node instance, affecting all connected inputs in other nodes.
Note that this is not the case with baked textures, and there each texture will use its own memory.

3. Except the mentioned missing gradient implementation: Are there any downsides or missing features using node materials?

Apart from the missing gradient node, there are still some nodes that will be added later (Scanned material, MultiSubTex), but since the node material system is much more flexible (and resembles closer the V-Ray core material layout) it will gradually surpass the possibilities of classic materials. Maybe the biggest downside at the moment is that it is missing the baking functionality for unsupported nodes from other node spaces.

4. Long time ago I asked Stefan_Laub about how V-Ray handles assets. Is it true that V-Ray internally uses uncompressed TIFF-Images in at best 1024*1024 Format, I mean in that Byte-Format, and if not, it has to convert (for ex.) Images as Jpegs loaded in the shaders internally to Tifs? So it would be wise to use Tifs instead of any other image-format?

Short answer - you can use any image-format that is supported by V-Ray and suits your needs and quality requirements. V-Ray uses different formats internally based on render options and whether CPU or GPU is used, and what format you are using as input only depends on tradeoffs between file size and compression artifacts based on the format used. Generally floating point formats (uncompressed TIFF and EXR) will yield better render results, but may require more disk space and may load more slowly.

[deyan.hadzhiev]

To answer your additional question:

1. If you use a Texture/Bitmap while it is in a TexLayeredShader or Cinema 4D Layer: What is V-Ray internally doing? Is it converting the whole Layered-Element to a .TIFF internally, send it to the V-Ray-Renderer so it can use it?
So when using Layer it means V-Ray has always to convert the whole Element so it can use it?

There is no difference in using a Texture/Bitmap in a layered node or shader than using it without a layer, regarding to conversion or baking. The difference is that during rendering V-Ray will evaluate the different layers based on the masks, blend modes and opacity.

[PdZ]

Thank you very much Deyan!

So there is no real benefit in kind of what V-Ray is doing with the textures? V-Ray has to mip-map anyway internally, depending on the distance of camera or visibility -> reducing Tile-Size of that texture? I try to understand where to optimize so V-Ray is as fast as possible. Thx in advance!

[PdZ]

To answer your additional question:


There is no difference in using a Texture/Bitmap in a layered node or shader than using it without a layer, regarding to conversion or baking. The difference is that during rendering V-Ray will evaluate the different layers based on the masks, blend modes and opacity.

So V-Ray is coded THAT intelligently? Wow. Thats a neat insight. Because I always thought V-Ray is converting/baking all textures. But if V-Ray is THAT intelligent and understand the structure of a Layer and just evaluating masks, blendmodes and opacity: Great!

[hsz]

Where is the gradient? still not present even on the latest build? and nightly? any word on this?

[aleksandar.kasabov]

Hello hsz,

The gradient node and some other nodes are currently missing as we had some technical difficulties and those will be added later. This was also mentioned in this thread. We will let you know once we have updates on the matter.