Well, I think you have reasons to use a VL70-m and this reasons are usually sound quality and attack. Regarding the attack, I recommend not to use a MIDI interface in front of the VL70 but rather buy a MIDI merger. This little box has two inputs and one output. One input would be for the EWI and the other one for the computer (via a MIDI interface). The output goes to the VL70. That way you play directly to the synth and avoid the latency of the MIDI interface which can be annoying. You can check this when you select a short sound with a good attack (a dry guitar probably) and try both connections: directly to the VL70 and via the MIDI interface. If you feel a difference, the MIDI merger is a solution for you.
Here are a couple of setups:
A) Straight connection:
This is the obvious setup, fast, does not allow to control the synth from within the computer.
B) This setup seems to be what you have now:
If you want the fastest attack possible, this is not the optimal connection for you. You lose some time via the MIDI interface.
C) This is better:
In this setup, your MIDI reaches the synth as fast as possible. And you can control the synth by software and hardware controllers.
This is the setup I work with, and I do about the same you want to do: layering sounds with effects and software instruments.
Another possibility:
This last example looks a bit overdone: first splitting, then merging. But it may make sense under certain circumstances.
In any case, watch that MIDI out from the VL70, it sends you avalanches of SysEx. You may either not connect the cable permanently or filter SysEx out in the Logic environment. Mainstage does not have problems with too much of SysEx so far I've seen. Maybe it filters it out anyway, don't know.
---
Talking about audio ...
I am not sure what your mixer does in this position, right after the VL70 and the EWI. It would make sense if you want an amount of direct signal without processing and almost without latency (which is a good idea). If this is not the reason and you have several inputs on your audio interface available, you rather want to go directly to the interface and keep the VL70 and the EWI apart to process them independently.
Optimal:
2 interface inputs for the VL70-m
1 interface input for the EWI
MIDI from EWI and Mac to the VL70 via a merge box
I don't know if the EWI has stereo audio, if yes, check if it it makes really sense. Otherwise mono is enough for further processing. The VL70 should play in stereo.
Layering and software instruments
Ideally you have all incoming audio signals and the MIDI stream in parallel in Logic or Mainstage. Then you can process and layer them as you like and finally mix them together to two or more output channels. You can control that mix by a a foot controller, pedals etc.
As we already know, not all software instruments work properly with a wind controller. You control the tone by your breath volume and you got to tell the instrument what this means. Controlling the channel volume often results in a "jerky leaping" sound. You can tame it a little by some processing in the Logic environment. In Mainstage you have the curve adjustments. For me, I don't like different processing depending on the program, therefore I use additional software.
However, the first step is to choose instruments you like and which respond in a way you like before you start a processing orgy. This requires some testing. Try to map breath to the plugins volume, modulation, filters etc. Sometimes it is a nice experience to map everything to everything
and enjoy new dynamic sounds.
I use mainly the following software instruments (in random order):
Logic Sculpture
Sonic Charge Synplant
u-he Zebra
QuikQuak Glass Viper
Cakewalk Rapture
This is a personal selection for my own music and you may miss the Logic ES2 and EXS which work pretty good with wind controllers. If you want good standard brass and woodwinds, check the Wallander Instruments, people say they respond very quickly but may not be optimal for live playing.
Remember: The fastest and most dynamic sound comes from the VL70-m, especially when you use the Patchman library. The EWI sounds may not be that good but they are also very accurate when you take them directly from the instrument. If you need this attack, take at least a little of the original sounds by splitting the line signal or via your mixer, because you will never get this attack again once your signals are in the software.
Effects and software instruments have necessarily some latency because they go through the computer and software and back out. The response is not as good as you would expect it from a wind controller. Recently a musician in another forum called that "walking through honey" and this describes the feeling very well. You can turn this into an advantage: use the direct sounds for the attack and take the effects from the software. Due to the latency of the effects the overall sound becomes automatically deeper. No wonder, because the computer adds some delay.
If you need some effects as accurate as the wind controller, you have to use hardware effects. I use both, hardware and software.
Edit:
Mhmm ... re-reading this, I apologize for my sloppy usage of the word "Attack". The attack of the sound itself does of course not change with the type of connection or in the software. I rather mean the "attack feeling", the time a sound needs to emerge when you blow into the mouthpiece. I want that instantly, with no latency. Effects are allowed come later.
