VRayBitmap vs Bitmap render speed

[DCHA34]

Hi,
My simple test scene consists of a plane on which I distribute a sphere with Chaos Scatter.
Simple material with a "Bitmap" JPG file: rendering time = 3min16sec
The same material with the same JPG file in "VRayBitmap": Rendering time = 3min31sec

I've had fun doing other little tests and it happens all the time.

Is it normal that the rendering time is longer with "VRayBitmap" than with "Bitmap"?

Or have I missed something in the settings...

[Joelaff]

If you are viewing it at an oblique angle then it could be the different MIP mapping used. The VrayBitmap will likely be a little sharper in the distance, and this would take more sampling, thus have a higher render time. But it would be sharper in a good way.

Post your scene for more investigation.

[vladimir_krastev]

Hi, could you try a much longer render? A tree minutes render time and sixteen seconds difference could be a misleading result. Are you using CPU or GPU?

[DCHA34]

Hi, i've just redone a test with a slightly more complex scene.

CPU Mode
Bitmap = 5m49s
VRayBitmap = 6m46s

GPU Mode
Bitmap = 3m59s
VRayBitmap = 3m22s


No visible quality difference in VFB compared to A/B.

PS: same render time with Vantage LiveLink between Bitmap and VRayBitmap

[DCHA34]

I'm going to launch a 2k or 4k rendering of the same scene.

[Joelaff]

Can you post your test scene?

[DCHA34]

Can you post your test scene?

Impossible. This is a customer's scene.

[Joelaff]

Can you reproduce it with a simple scene? Could be other factors in that scene, but those numbers are strange. I gotta say it's been a while since I time tested it, but last time I did they were the same or VRayBitmap was a tiny bit faster, and VRayBitmap most definitely does MIPmap differently, I posted a tone of tests a while back.

You're doing anything silly like setting the blur to a low value, right?

[DCHA34]

Can you reproduce it with a simple scene? Could be other factors in that scene, but those numbers are strange. I gotta say it's been a while since I time tested it, but last time I did they were the same or VRayBitmap was a tiny bit faster, and VRayBitmap most definitely does MIPmap differently, I posted a tone of tests a while back.

You're doing anything silly like setting the blur to a low value, right?

As written in my first post, my first test is a simple scene with a default material and all the default rendering parameters. It couldn't be simpler.

[vladimir_krastev]

As written in my first post, my first test is a simple scene with a default material and all the default rendering parameters

Could you run this scene with a lot higher resolution and render settings and share the result. Too short render times in the minutes range are not reliable for comparison.
If you are concerned about confidential scenes you could send them to the support contact form. The whole ticketing system and it's storage folders are well secured and the scenes are used only for troubleshooting purposes.

[DCHA34]

Hi, i've just redone a test with a slightly more complex scene.

CPU Mode
Bitmap = 5m49s
VRayBitmap = 6m46s

GPU Mode
Bitmap = 3m59s
VRayBitmap = 3m22s


No visible quality difference in VFB compared to A/B.

PS: same render time with Vantage LiveLink between Bitmap and VRayBitmap

Same scene in 2K :

CPU Mode
Bitmap = 15m25s
VRayBitmap = 17m19s

No time for GPU testing

[DCHA34]

Here's another test with a simple scene rendered in 8K

CPU Mode
Bitmap = 37m15s
VRayBitmap = 39m1s

[Joelaff]

As written in my first post, my first test is a simple scene with a default material and all the default rendering parameters. It couldn't be simpler.

As written in my first post, can you share the test scene

Oh, there it is now.

[vladimir_krastev]

Thanks for the scene. I've ran a few quick tests. There is indeed some delay in render times with the VRayBitmap. We will investigate the issue further and get back to the thread.

[DCHA34]

Hi,
I'm coming back to you to find out if the rendering time problem has been fixed or not?

Because I have another rendering time “problem”.

I did some tests with the file “Evermotion_Tutorial_01_Final.max” provided by Evermotion in their tutorial here : https://www.youtube.com/watch?v=QPjo...RaBWZw&index=2

It's an old file containing Skylight Portal.

Here's my 3DS MAX 2024 + VRay 6 workflow:
1st Test: Open the file, I don't touch anything except that I change Progressive Render to Bucket Render and run the render.
Render time in CPU = 4m46s
2nd Test: Open the file, remove the Skylight Portals, change Progressive Render to Bucket Render and render.
Render time in CPU = 6m15s

I ran the test with 3DS MAX 2025 + VRay 7 on CPU: identical results.



On the other hand, GPU rendering time is identical.

Where does this difference in rendering time come from?