Slightly OT: Generally speaking, total neutrality is the ideal goal, but also exceedingly difficult to get perfect. So how big a deal is it? Is not the purpose of a calibration to correct for such differences? I understand a big push in one colour may be too big to correct for, but little ones should be no big deal. Could you actually have a screen surface that is a better overall performer in other respects but exhibits push in one colour. A good calibration can correct for the push, leaving you with a better screen than one that was perfectly neutral but less well performing in other respects?
I'll leave the calibration and colorimetry stuff to Mech and Bill, but I'll weigh in on this. A
perfectly color neutral screen is very hard to get, as is a
perfect anything in the real world; but luckily we don't need true perfection. Some time ago Bill and Dr. Mark D. Fairchild discussed the matter and came up with a set of neutrality standards for projection screens (which is what we currently use to determine levels of screen neutrality). Most projectors can compensate for small screen neutrality variances, but the more a PJ has to compensate the less chance it can do so properly and the greater chance it will throw other colors off at the same time; thus the desirability for as neutral a screen as possible. Also, the further a screen is from neutral the greater image colors will vary as the projected image drifts away from D65 (an effect called metamerism). Very few screens are perfectly neutral, very few PJ's are perfectly D65 - thus the need for calibration.
Yes, if you need a screen of a certain brightness and that brightness could not be attained with a neutral screen you would have no choice but to use a screen that would achieve that brightness even if it were less neutral. The thing is, screen neutrality (at least with a painted screen) isn't all that hard to achieve if it is approached from a scientific aspect.
The case above where someone hypothetically had to settle for using a non-neutral screen to achieve a viewable image is a case where the wrong projector was being used. Bill has said for quite some time that home theaters should be designed by determining what lighting conditions the screen will be viewed in (no ambient light or lots of it) and
then go shopping for a PJ and screen. This makes so much sense! Unfortunately, most people either get whatever PJ is on sale, or buy one that a friend says works well, without taking
their viewing conditions into consideration. That leaves them trying to find a screen that will compensate for using the wrong PJ. Many times this can be done, but has definite limitations. No screen can totally compensate for not having enough lumens to start with.
I understand as screen formula developers that the holy grail is totally neutral, but is there a risk that goal is followed to the detriment of other goals? For example, on other forums it seems that gain or reflectivity is the goal, to the detriment of neutrality. I'm not claiming to understand the complexities of the art/science, just curious...
I would say the simple answer is no. Neutrality doesn't affect gain. The color of a screen has nothing to do with it's reflective properties other than how much light it absorbs rather than reflects. The darker the screen, the less light it reflects.
Screen gain is a subject that is almost universally misunderstood. There is a general feeling that more gain is better no matter what. This is just the opposite of what it true. Think of screen gain as the ability of the screen to focus the reflected light toward the central viewer. To do this it needs to redirect light away from somewhere else to point it straight toward the audience. This means that those light rays are
not being reflected to areas distant from the central area on-axis with the PJ. This makes for a
darker image for those side seats. No screen generates light, it can only reflect what light it receives from the PJ. If a screen has too much gain it will also hotspot, which is when portions of the screen are visibly brighter than other areas of the screen which should be the same brightness.
We only need gain when the projector isn't bright enough for the job. In the past this had been the case more often than not, thus the need for screens with as high a gain as practical. With todays PJ's approaching the 10,000 lumen mark I personally think that the days of high gain screens are just about done.
The other mixes you are probably referring to were designed with gain as the primary goal and neutrality was deemed not to be all that important (and still doesn't seem to be). Back in the day they were designed PJ's had way fewer lumens than they do today.
The perfect screen is one that can't exist in reality. It would be totally black so it would absorb all ambient light hitting it yet it would still reflect 100% of the projected image to a wide audience.
Check out the Silver Fire thread. Check out the readings there. That can tell what gain from heavy doses of mica and polyurethane will get you. 
