Well, my situation is a bit different, but you can benefit from what I'm doing, but alas, in the end, you might run into the same problem.
My difference is that I still love the MyDMX product, even with this issue.
First, I think you need Compu Show. This would be ideal. But if you could have a computer run say, maybe Sony Vegas, you could have that run MTC, send it to a MIDI interface on your lighting computer, then Compu Show would play your stuff in perfect sync once you got your offset configured properly.
But, short of that, let me describe my situation:
2 hour show, involving LIVE actors and video. Using MyDMX, it was programmed using a stopwatch(for lack of a better term). To ensure I get my act back together due to drift issues, I have many "correction" scenes where there are cue points I can trigger and ensure I stay on track in case something WILL and DOES go off. Each scene is told where to go to next, so if in an ideal world, if I had to, I could go hit the crapper during the show and it theoretically should go well. The reality is that I do have to hit those scenes, as well as deal with errors. I know, it's frustrating.
The bottom line was I was trying to use a product that was not meant for what I am trying to accomplish.
Even if you scale back Compu Show a bit, you can pull in videos and have it sync to your lights perfectly. Toss the video window to a secondary display and it will be broadcast to your projector or whatever. But since you're doing audio only, divide it up into smaller chunks and then import it into Compu Show. Same end results, just no need for a secondary display if you don't want it.
Second, it sounds like you need not only a DMX merger/combiner, but you should be using an Elation Opto Branch/4. Elation makes one. You can set the priority and where to look. Reason being is that I've had MyDMX freak out when I've connected or disconnected DMX signal live.
What I do is: Power on the computer and walk away as I'm stil doing other stuff, so waiting for a start up is nothing. DMX is connected to the dongle, then the dongle is plugged into the computer. OR, you can reverse that by plugging in the dongle and then plugging in the DMX chain to the dongle. Either way, make sure your chain is connected before you run MyDMX.
Even so, following my own advise still has resulted in problems.
As far as people who don't know crap, yeah, welcome to reality. Sucks! I was once given the task of designing and implementing a world wide network for a sales team. They gave me 30 minutes to not only do it, but to have the project DONE and in place. The logic: I don't understand it, therefore it must be simple. Plus, they figured they already had the equipment spread out sufficiently throughout their global business operations that chances are they didn't need any hardware.(much less data circuits!) While I did do the design in 30 minutes(that's easy), deployment never took place since they refused to budget for it. I left the job, the manager got terminated. I left because "If you refuse to purchase equiment and services to make it happen, there's no longer anything for me to do here".
Yes, I like MyDMX. I set it up to run rock shows and small concerts, so it works ideally for me because of how I designed my scene sets. You are using the wrong product for what you're doing.
The justification is easy:
MyDMX retails in most places for around $300. I have no clue what Compu Show retails for as I said "I need it" and got a good deal on the product, so I won't comment on cost other than saying it's gonna be a lot more. And if they had spent the money right the first time, you could have done the job RIGHT the first time, and hence saved them money to do it right the first time.
I spent so many hours redoing my main show in COmpu Show and it took forever because I was learning as I went. Nothing like being forced to use a product to learn it best. I can't say I'm complaining, because in the end, the results were flawless and amazing.
I've just been informed I have to reprogram a smaller show(same producer). It only have 42 scenes, but at least 10 of those are the same, and most of the rest have many common elements. I suspect my time to reprogram this 45 minute show will be under 5 hours. Chances are I'll send 2 hours just figuring out what to do based on what I have to work with, and then the next 2-3 hours getting it done.