[Gardus@Supersprint]

Looks like a very good description of the exhaust workings, I'd like to add a few things.

About headers:

1st and foremost, on NA engines like the GT86, and on midly Supercharged cars (so we're not talking on drag racing SC engines with huge overlaps) the headers and the 1st cat have the main influence in the power delivery
This is why I'm always skeptic when companies claim big power increases on catback or even axleback systems.
This is obvious, the primary pipes (and secondary in case of 4-2-1 manifolds) collectors are the primary source of returing wave pressure to the exhaust ports, and the second is the cat.
The diameter and routing of these pipes are also very important for flow, as well as the efficiancy of the cat, that can be a heavy restriction on the flow.

It's also important to avoid steps between the exhaust port and the pipes.
If you want bigger primaries either increase the ports size or make stepped primaries.

A T junction (instead of a Y) between two primaries is not a good thing but is actually less important than most people think. The ones you have in the stock GT86 header joins primaries coming from exhaust valves that opens not in sequence but at opposite times of a cycle, and these junction are close to the exhast ports where the waves are very defined.

It's more important that the secondary pipes collector don't have a steep angle, and it's even more important in 4 into 1 collectors. Two or four gas flow must enter at the same time, the compressible nature of the gas means the arrival of the wave pressure is not like "single persons" arriving at the different times, which means they all can pass by the same door, but it's more a flow of many people, with the density of them changing at regular intervals. While you don't have the 4 flows arriving all with the highest density "of people", they cannot pass between a door large enough just for a single flow, and you don't want the flows to collide with each other...

Talking about 4 into 1: in this type of headers is very important that all the primary pipes are equal length, otherwise you seriously risk to have the highest quantity of gas caused by the waves arrive at the same time at the collector.
This is the reason why many OEM headers are 4 in 2 in 1, they are cheaper to make as they can be stamped instead of being tubular, they occupy less room in the engine bay and collectors can be at a steeper angle.

The precence of tight bends in the primary pipes doesn't have much of an effect as the drop in flow is much less important than the effect of the pressure waves.
After the collector the eventual drop in flow (brought by a tight bend, a restriction, a chamber-type silencer etc) is to be avoided to keep the gas velocity high.


Talking about tuning the pressure waves to improve the scavenging effect, you can do this for a single rpm point, and less so on multiples, but on a road engine is quite hard to obtain huge improvements.
An interesting idea is to use this system on pure hybrids, where the engine works at a constant speed!
As already said many engineers have tried various methods with exhaust valves, with multiple "tuned" wave reflectors on the line etc.
They can be effective but they're just one of the factors.

On the intake instead variable lenght systems work very well!
It's obviously almost impossible to make a continously variable lenght exhaust system, so you can make a continously variable flow one, or you can make a bypass valve that shorten the exhaust, but that is for different reasons...



Less talk about DIAMETERS
Some say that bigger the better, if you can keep the correct A/F ratio etc.
Then you fit a 3,5" system on a 200 hp car and it lose power...why?
Many people talk about insuffient backpressure. This, why it is a easy to comprehend argument, is quite improper.

The two main factors are:

-Sudden variation of the diameter: you usually see this when you have flanges, and the pipe before the connection is 2", and the one after that is 2"5 or more. You can imagine what happens when the mostly linear flow enter the bigger pipes: it gets turbolent, you get drops in pressure, and the speed of the gas suffers. Also you introduce unwanted returing pressure waves.
This is the worst when you have it at the exhaust ports.
If you don't have these steps but all the diameter variation are gentle slopes, there's the second factor:

-Inertia: the gas column inside the exhaust has weight, and with weight you get inertia. To get this column moving you spend energy, especially when you drasticaly increase the speed in sudden acceleration, opening the tap!
This energy comes from the gas explosions in the chamber, that you want to be trasferred to the wheels...
You can easily understand that a bigger column (a bigger bore exhaust) weights more...but it's not a direct proportion!
As you increase the diameter the surface of the exhaust increases 3,14 times more. You have hear exchange between the hot exhaust gas and the normal air around. The exhaust gas get colder, more dense, heavier.
Also, to expand in a large volume the exhaust loose heat.
As it gest colder (and heavier) it slows down and it takes more time to get our of the exhaust, and the more time it stay inside the more it gest colder...

This is the reason why you get extremely short exhausts on race cars
This way you can have large bores without any issue!
On road cars you must reach compromises: the exhaust must exits from the back (usually), you need silencers that take space, the packaging of the car is complicated (suspension parts, braces and other components are in the way), and the design is important too.

The bypass valve on some cars works well in this way as on low rpm when the car bust me quiet and performance are less important you can have a longer route, and you have also drops in flow brought by silencers.
In the upper range, when you want the best performance, you open the bypass valve, shortening the route and bypassing the silencers, keeping as much energy as possible