anybody use this to good effect without losing the car, especially
through corners ?
Lack of peripheral vision in car sims has always been a problem. GPLs'
use of the letterbox view is, I guess, partly an attempt to address
this problem and partly a frame rate saver. It may be better than a
full screen but I think there's still a way to go with this problem.
I find its possible to loose sight of a nearby car only to have it
turn up by contacting you some way down the track. The look-around
buttons will enable you to see much of your blind spot ( but not all )
can aid the situation. But who can drive hard, pushing the edge, while
their view is turned to the side. On the straights its somewhat
possible but when going through a corner, wondering where that nearby
car has gone, its just about impossible to use this system without
loosing your own car.
The solution in NASCAR2 I believe is much better. There they used a
three part rear view mirror where the outer parts showed the side
views more while the centre section showed the rear. Not allowed for
in that game but neither-the-less tweakable through the program files
were ways of turning and modifying the mirrors to suit the particular
tastes of the driver, provided they could be bothered to go to the
trouble this kind of thing entails.
So what are some of the options ?
1) Driver Position Turn-Around Views.
The benefits of this approach is that its very easily implemented and
the number of point-of-view displays on screen at time remains
un-increased - being one drivers view, one right-side mirror view, and
one left-side mirror view.
Its easy to do because all that's required is to alter the drivers'
view angle perspective - which the program has to recalculate
constantly anyway with the normal movements of the car. It probable
only take a few lines of code to achieve and puts no extra load on the
system what-so-ever when activated. Perhaps a good solution if it
works - but I don't think it does.
The downside is simply that it very poorly solves the initial problem
that prompted its existence. Sounds OK and is easily done but in
practice is unusable when you most need them IMHO.
2) Multi-Part Mirrors
The mirrors could each have two parts, the inner parts reflecting the
direct rear view and the outside parts reflecting the outer blind spot
area. Similar to the solution used in NC2 but modified for the side
mirror situation.
The benefits are that the driver would have some constant peripheral
input, and that's what I feel is needed.
The downside here is that this would increase the number of
point-of-view displays on screen at the same time. One drivers' view
and a four mirror views. Each view angle requires its own set of
calculations. The more point-of-view displays shown simultaneously the
greater the load on the system to provide such. With GPLs' already
heavy system demands, albeit justified I hasten to add, extra
simultaneous displays may not be easily accommodated.
3) Convex Mirrors
These may be helpful but possibly with each pixel of such effectively
showing a different perspective it may demand too much extra maths
anyway, despite the number of simultaneous displays remaining at
three.
4) Peripheral Display Areas
Simular to multi part mirrors except the display areas are completely
seperate from the mirrors. As an option, activated by a button say,
the upper corners of the screen could display extra peripheral views.
This would be a good solution as far as the drivers visual information
is concerned but for the reasons already covered puts more load on the
system. Still as a momentarily used option it may be useful.
5) Mirror Position Turn-Around Views.
Given the restraints mentioned above, this is the system I feel would
be best. Not perfect but better overall than the others cited here.
In this case a driver would press a button to cause the desired sides'
mirror to turn its view around to show the peripheral area / blind
spot. In this case the drivers' forward view remains constant. A
mirrors' view ought to be changed to and froe by a toggle action, not
a hold-to-effect procedure.
In this way the driver would exchange some rear view information for
some peripheral view information as required. The number and size of
simultaneous displays remains unchanged at three and the system isn't
doing any extra calculations than it otherwise would be doing. Also as
nothing other than the viewpoint angle is altered this should be very
easy to implement.
6) Adjustable Mirrors
Even without any of the above methods implemented I think the driver
mirrors ought to be angle adjustable as individuals prefer. In this
way some may prefer to angle their mirrors out a bit more to pick up
some extra peripheral information at the expense of rear view info.
Again there'd be no extra system load required and it should be easy
to do.
There's some other methods that come to mind but are either variations
of those mentioned or really aren't viable on a usefulness-for-load
basis.
I don't know if Papyrus intends to release any upgrades, free or
otherwise. ( I wouldn't call them patches as 'patch' implies that
something was done but done wrong and needs fixing. 'Upgrade' on the
other hand implies that what's done works OK as far as it goes but can
be improved upon. I think what's done in this area is simply too
limited and not really useful and can be improved ).
Any other thoughts on the peripheral vision problem ?
Phillip McNelley