LRGB or RGB files for painting with varnish

[jtneubecker]

I did my first RTI image of a Rembrandt painting and noticed a hint of green in the LRGB file; but when changed it to RGB- the color was more accurate; is it better to use RGB when doing most paintings?

[cdschroer]

I moved this over to the processing list - since I think it's a better fit here.

 

There is another post about choosing hsh vs ptm that is likely of interest: http://forums.culturalheritageimaging.org/index.php?/topic/190-hsh-or-ptm-how-to-choose-the-best-fitter/

 

In terms of ptm types - the LRGB is a smaller file size and it is the default. The RGB doubles the file size, but it also renders colors more accurately.  This is particularly noticeable with shiny material, especially gold.  I actually think HSH has some advantages for shiny material.  Read the other post to learn more about your options there and why you would choose one thing over another.  Like most things in life, there are trade-offs.

 

Carla

[Taylor]

I had similar encounters with specular reflections off the varnish when capturing RTIs of a 19th century landscape painting. The degree of specularity would depend on the type and condition of the varnish, the surface roughness, and how flat the surface is; as we know, not all paintings are flat. Concave curvature of the surface can make it more difficult to avoid.

 

Some other suggestions to add to Carla's: Shooting a smaller area can help. It was more of a problem with higher light angles, and I removed those images with larger areas of reflection from the RTI sequence during processing. I didn't use a strict threshold of "no specular highlights allowed," since I didn't want to lose useful data. Keeping the light angles below about 60 degrees helps to minimize the problem. Some have used a snoot on the flash to help, but you want even illumination over the surface, so proceed with caution. Finally, George Bevan points to a paper by Lindsay MacDonald that uses a different method of processing RTI data sets, using many triplets of photometric stereo images, that might (?) help to limit the errors in normals from images with specular highlights. George's discussion and Lindsay's paper can be found at forums.culturalheritageimaging.org/index.php?/topic/220-care-tool-and-surface-normals/