The recent buzz about PBL has caused me a bit of agitation, mostly about the misperception, hence this post.
1. PBL in run-time solutions does not mean physically accurate.
2. The same result could have been achieved without changing shaders solutions*.
*If you had previously handled gamma correction.
The difference in PBL shaders and non-PBL shaders is essentially lock stepped parameters that artist previously had individual control over; spec power, spec intensity, and diffuse intensity. This result is a constraint on artist. While constraints are sometimes good, like when the constraint focuses infinite possibilities to a set of rules to work within, game design for example, I debate the usefulness of this on shading/lighting for artist. True it does simplify the number of parameters the artist has to tune. It makes the tuning parameters less intuitive and slower to iterate on since they're often driven by textures instead of single float values. If artist viewed their work in a whole environment more often we would see the where objects seem out of place and adjust accordingly, negating the need for PBL shaders. Time will tell if this is a true benefit or not.
UPDATE: May 20, 2014
Well, the doctrine of physically based whatever has been taught far and wide and implemented in a couple released games. The result of which is hard to argue with. In short, I am swayed, convinced it is a proper improvement to workflow and end results. The post mortems I've heard have revealed that a large amount of fudge magic is still required to the shader/post process/texture value range/etc to reach the desirable results, so nothing new there. Still the end results are quite an improvement over previous practices.
