The top pic shows a typical head runner layout while the bottom pic shows the ideal poppet valve port layout. It makes a good reference to runner diameter versus valve diameter.

Looking at the top pic again, it seems like the air that flows through a typical head (LSx excluded) would flow from left to right (towards the valve) and then be forced to change direction as it enters the chamber. Since the air doesn't want to change direction, it would make sense that the air would want to stay flowing from left to right. This idea is supported by the notes that say there is more loss at point 8 than 7.

What would happen if lateral air flow were reinforced? For example, grind down point 3, fill point 4 and upstream of the valve to smooth the ceiling, and then grind out point 8 as much as possible. Keeping in mind that expanding the runner is bad and that the cross area needs to be kept in check.

 

I'm not sure where you found this information, but I suggest some folks who know a lot more about it than I will not agree with all of it. Operating engine airflow is "unsteady" flow, so some of the (simpler) principles of steady-state gas dynamics don't really apply. There has been tons of literature published on this, and mega-tons of internet discussion, of which only a little bit has credence.

If you read stuff written by Darin Morgan or Larry Meaux you should take it to heart. Both of them have much empirical data to validate what they say about engine airflow and its relationship to power. Additionally they are resonsible for many high-end race-winning engines.

 

In a sense you are right. You open up the main runner, and widen the valve opening to keep air moving smoothly (i use smoothly as a loose term), or as smooth as it can get. Really the air is pulsating insanely fast from the valve opening and closing. You would not have to fill point 4 since the side profile picture does not show what it is really like.

http://fawkit.org/kc104/lt1/welds/DCP_0737.JPG

if you look down the runners on that picture you'll see where the valve stem comes out. Going back to your picture, where point 1 goes in to the valve stem then becomes point 4 on the other side, is really just one continuous curve on the side of the stem (go back to my picture and see how it looks on the side of the stem), and really would hurt the air flow to fill that area.

 

Which is more important though, airflow or getting the air to flow at the right angle to the valve?

 

I think building up point 3 would increase velocity enough to outweigh reducing volume.

 

It goes without saying that you need to increase velocity at peak rpm to fill the chamber the best. What is more at question here is the angle at which the air hits the valve. Ideally it would come straight down perpendicular to the valve face, but since that can't be met, is it worth it to lose CFM for angle?