Author Topic: Making optimized pictures for normal map  (Read 1499 times)

Hello all,

I started modelling the Basilica of Vezelay. Rather then modelling everything I am planning to fake some parts using normal maps.
At the moment I'm not to far from Vezelay so I can go there and make the pictures myself the way I need them to be. That brings me to my question. What can I do to take the best out of B2M in terms of taking the pictures?

Thanks a lot regards, Robert

Last Edit: December 12, 2014, 11:12:40 am

Hi (or "Bonjour" if you're near Vézelay :) )

It mainly depends on what B2M version you use, here are some recommendations:

- B2M2:
Try to take pictures with an "ambient" lighting (no directional lights). The more "diffuse" the lighting is the better results you'll get.
For exterior elements, a cloudy sky is preferable (avoid bright sun/sharp shadow).
In the filter, the "Shape Recognition" parameters can help to "simulate" details depending on the angle/intensity you set.

- B2M3:
We have 2 methods for the outputs generation for this one.
If you set the "Method" parameter (in "Global" group), on Slope Based, you can use input with a Directional lighting (clear sky). Just be sure to set the "Light Angle" parameter to something similar to the light direction in your source image.
If you set it on "Luminance Based", try to take your pictures with an "ambient" lighting (as for B2M2).

In both case, avoid too important directional lighting: if you get "out of range" areas (completly black, or white), you'll not have the possibility to generate a lot of information from it.
Don't use the flash or, if you need to, diffuse the lighting using a softbox to avoid non homogeneous lighting.
Try not to go too high in ISO to avoid noise caused by the high sensibility, use a longer exposure time (lower shutter speed) with a tripod is preferable.
Last Edit: December 12, 2014, 12:04:35 pm
Lead technical artist

Thanks a lot Gaetan,

This is such a helpful comment, much appreciated.
I will do some testing, is good to have your images as a reference.
Many thanks again.

Regards, Robert