Air launch to LEO works better if you take into account the three top items in their order of importance: (1) staging speed, (2) path angle at staging, and (3) staging altitude. These three items are not even close to equal importance, speed is simply everything (an unfortunate technological fact-of-life). You get do-able second stage mass ratios starting about Mach 5 to 6, and reasonable second-stage mass ratios at about Mach 10, for two-stage vehicles.
Speed at staging figures directly into the delta-vee requirements associated with the rocket equation for the second stage. Path angle at staging gets you to a second stage trajectory that needs no lift (and no drag-due-to-lift) for a simple non-lifting ballistic trajectory to orbit. Altitude at staging gets you some extra energy in the flight vehicle energy-management equations, but is the weakest of the three effects, by far.
An all-rocket first stage airplane can do this job, but it is of enormous size, with some very serious structural issues having to do with landing gear loads that start to resemble a water balloon resting on nails. That’s a common sense thing, as well as a structural design thing. On the other hand, you can save weight by using airbreathing propulsion to the greatest extent possible in that first stage. “Greatest extent” simply means the widest-possible range of speeds. That’s just common sense.
Scramjet is neither ready-for-prime-time technologically, nor a wide speed-range type of propulsion just yet (X-51 flew in scramjet at Mach 5 +/- 0.1). It takes over at about Mach 4 , minimum, so going for Mach 10 capability, you cover only a delta-Mach of 6. If your scramjet system really only gets you to Mach 7 or 8 as currently seems practical, well, the delta-Mach drops to 3 or 4. That’s a lot more likely outcome for the next few decades.
The fastest turbines that ever were (those powering the SR-71), were good only to about Mach 3.6-ish, which is woefully short of an adequate staging speed. So such a first stage would require rockets as well, and also some protection for the engines against overheated inlet air, at speeds above Mach 3.6-ish, which the SR-71 never had. If the turbine-powered stage can take off on only-turbine power, then the airbreathing delta-Mach is about 3.6-ish at maximum, since turbine supplies static (Mach 0) thrust.
Plain ramjet technology has been well-proven in flight for many decades now, and can be arranged to work from Mach 1.8-ish to Mach 6 quite easily. That's a delta-Mach of 4.2-ish. That figure is at least as good as the more realistic scramjet concepts, and better than any of the turbine concepts. So, of the choices, it’s quite simply the best, especially when you look at technological readiness-for-application. That’s just basic common sense plus the facts of technological readiness.
If one were to do a combined rocket-ramjet propulsion airplane for a first stage, we would take off on rocket at Mach 0 and accelerate to about Mach 1.8, then transition to ramjet and climb-and-accelerate to Mach 6 as high as is feasible, then go back on rocket power, pull up sharply, and accelerate exoatmospheric to speeds that would have been near Mach 10 in the air. And, with ramjet, there’s less susceptibility to overheated inlet air (no turbomachinery to damage).
That switching back and forth between rocket and ramjet requires either combined-cycle or parallel-burn (with separate engines). Combined cycle is quite simply not technologically ready for prime time, and has always very seriously compromised the performance of both propulsion cycles, because of the incompatible engine geometries. So parallel-burn with separate engines really is the way to go! Basic common sense.
The hardest part of the design is packaging the rocket engines somewhere in the airframe, because the ramjet will essentially fill the fuselage (a hard technological fact-of-life). But the rocket thrust chambers are actually quite small, and will fit inside the aft portions of the wing strakes or fillets. Plus, there is now aerospike nozzle technology to eliminate drag-inducing huge engine bells. The next hardest part of the design will be inlet and combustor heat protection-as-reusable devices, not one-shot ablatives as in missile work.
This could have been done at least 2 decades ago. It was not.