Firstly, let me say I am not, nor have I ever been, a TBI hater!
Maybee I need to join the SAE, but with the best Weber carb, the Sauter mean volume diameter of sprayed fuel is 60 to 50 microns. Dellorto carbs are even less. In an EFI engine, 60 microns is common. The shape with an injector is very defined, not a nebulae, but a defined roosters tail. It is possible to drop to 26 microns with a anionic or cationaic additive for carb engines, or to 10 microns with a 150 psi line pressure injector and an 15 amp electrical source.
I don't have any Rayleigh scattering or one, two or three dimensional turbulence laser flurosence picutres, or Planar Laser-Induced Fluorescence fuel concentration measurements ...but from all the reports I have seen, it is a certain fact that the injector is quite a poor atomiser.
There are five specific matters to consider.
i) One, a carb or TBI flows a fully wet intake manifold, up to 30 % of the engine capacity is swilling around in the log intake or whatever manifold. So a non independent runner TBI or carb is always worse fuel economy wise than a port EFI system because so much fuel is wetting the surface of a metal intake manifold.
ii) a TBI is very simalr to a carb in its delivey of fuel to the back of a throttle blade, and then onwards through a wet manifild into the port. As stated, there is a high line pressure, and a wide spread of dropplets of fuel. It should logically follow that any injector must atomise better than an old fangled jet!. However, evidence from the early 80's affirms that atomisation is worse with TBI.
The following facts
iii) Look at the 82/84 CrossFire 350, they had to have bath plug flow inducers in each port to re-atomise the mixture.
iv) The delivery of TBI fuel is not even as violent as a conventional carburetor. Carbs can flow at 250 feet per second, with turbulent air rushing into a 90 degree angle jet, issuing fuel at a huge pressure of way above 45 psi after colliding with an airstream air having been a 15 to 40% restriction (the venturi signal). The mixing event is indeed turbulent, and the atomised fuel is often smaller in droplet size than a TBI unit. TBI systems have to use in excess of 22 psi but less than 53 psi pintle injector. Engineers have spent a long time making simple voltage changes simulate a simple jet. The up an down motion of the injector is unable to work as well as a carb, but there is a reason why an injector is supperior. I stress again, no flow interupting venturi exists to create a mixing vortice.
v) TBI set-ups primarily use single plane manifolds due to the fact that bends and turns cause the mixture to pool very quickly, and Ford and Chevy CFI/TBI's need to have small volume, single plane intakes to work well before things settle out and condense. The entry of the air into the manifold is more laminar than turbulent on a TBI engine, while carb flow is much more turbulent.
On a very concilatory note, the real advantage of TBI is the vastly better ability to meter fuel right through the heat range. From cool to sweltering engine and environmental temperatures. All because it's anologue or digitally mapped, and as time went on, the digital mapping was able to operate by altering (fuging) the fuel delivery to react to the oxygen sensor. Injectors are seldom blueprinted to even 10% of the flow of a jet, but you can alter them minutely at any delivery. Althoug a jet is much more accurate , with a 4.5 to even 1.5% accuracy in its metering, it is only metered by its response to air flow by a variatey of systems. TBI is very, very accurate, while carbs are a hit an miss device which require a compromise as operating conditions vary.