Locking view vection or 'self-vection' (pitch, bob, roll, yaw) in games is not the best idea; foisting your idea of what view vection should look like is not just an issue of a 'turn-off' to some players -- it can make games unplayable due to simulator sickness.
The simplistic id engine view vection physics are unable to achieve anything remotely kinesically correct or realistic and the effect is individual; some people bob and roll when walking like trolls, some people (typically women and smaller framed men) barely bob and do not roll at all, and visual accommodation is also individual further and radically altering the perceived real world effect.
Most people do not perceive any view self-vection unless they are walking very near a static object(s) e.g. along a wall or fence -- some will not perceive it even in these circumstances.
View vection mechanics in games in general also assumes an technically incorrect perspective where the camera is locked in the players eye sockets -- where in reality a persons eyes are able to instantly react and compensate for real self and induced vection.
While the number of gamers that react nauseogenic to view self-vection is statistically small the number of people in the population that react is relatively high (around 30%); so if you want your game to appeal to the largest audience you may want to follow the path wise commercial developers have followed after reading up on studies of view vection based simulator sickness.
A few references about the topic below, one of the best and most thorough was done by the Navy is no longer on-line but can be found in most libraries government documents departments:
http://www.resonant.org/text/simulator_sickness_report.html
http://www.hitl.washington.edu/publications/r-97-12/
http://www.hitl.washington.edu/publications/r-97-14/
http://citeseer.nj.nec.com/393186.html
.