I'm working on another engine that is a much more high-performance engine from original equipment (Kawasaki ZX10R). This is a 4 valve per cylinder DOHC engine. It has very good CNC-machined ports as original equipment. Upon crunching some numbers, I have established that based on the calculation given in post #107, the Mach number at the minimum cross-sectional area in the port is around 0.6 at the peak-power RPM and the peak piston speed. Seeing that this is in a good range, the only thing I've done was to blend the entrance from the throttle body into the port ... and I've done it by adding epoxy to the head, not by increasing the diameter of the adapter.
If the RPM of the engine is fixed then ports that are a little too big seem to kill power quicker than ports that are a little too small ...
If you plug in the formula at post #38 to a TDI engine's bore, stroke, and target peak-power RPM (4000) with a target Mach number of 0.6 then the target runner diameter is only 25mm, and if you don't want excessive losses from velocity speeding up (through the minimum cross-section) then slowing down (into the port above the valve) then speeding up (through the valve curtain area) then slowing down again (into the cylinder) that cross-sectional area has to extend all the way to the valve seat. The last little bit needs to be a slightly bigger diameter to account for the valve stem and the guide.
There is not fixed "best" flow speed for all engines.
Many things affect optimal flow speed.
low duration camshafts tolerate and like less flow speed for lower rpm use.
Flow speed is all about volumetric efficiency and inertia filling after BDC.
If IVC is not much after BDC, then engine can not make much use of flow speed, but needs more area/flow to fill cylinder during less time.
Flow speed is basically made by some restriction in runner and it reduces cylinder pressure / filling during intake cycle before BDC, but inertia continues to fill cylinder with made flow speed after BDC.
If valve shuts too soon (lets say at BDC) then all that flow speed made just generated less pressure (air mass) in cylinder and before mass inertia does it job, valve shuts already.
after ~0.6mach there is more pumping losses when making flow speed, than what can be gained from increased mass inertia.
All latest OEM 4-valve (what i have seen), or new GM LS3+ engines have HUGE ports, lots of flow and runner area and small camshafts for the intended RPM.
This does not yield to good VE%, but gives broad RPM band and good RPM capability as cylinder pressure is high due early IVC, but due big cross section areas, it can carry to higher RPM.
Option would be to take more RPM, but that would need more camshaft, or to less runner area and more camshaft to maintain same max hp rpm, but both would make emissions to fail.
Without flow speed there is no inertia filling to allow good volumetric efficiency and camshaft / runner areas must be specified for each other.
But like you say, .6mach is close to perfect for any high rpm performance engine.
The more VE%, the more sensitive engines are for everything.
RPM limitations make some racing classes really interesting when HP must be get before certain rpm.
Beyond 120VE% things get hard and approaching 130VE% (above 170hp@10000rpm with ZX10R).
But with more rpm and less VE% things are usually easier before RPM gets crazy.
Usually slightly to big is not as bad as slightly to small =)
When speed starts to go beyond .6mach, usually power drops drastically.
Depending of pulses and overlap scavening, flow speed can be easily 0.1 more than less than what calculated, so you need to use simulation to get correct flow speeds and runner areas.
Most flow losses come from air deceleration / expansion.
Depending on valve sizes, I usually start taking speed down slightly before valve guide to make it turn better and expand more gradually to combustion chamber.