Note that this latter discussion has absolutely nothing to do with the NeoPixel problem. 
The problem was solved already, but we can keep babbling on without the annoying message that it has been. Which latter discussion do you mean ? Like the Z in the XY problem. There is a possible practical application for Y, and now there is also possibly a viable solution for Y. The result may end up as Z, but that has nothing to do with the OP's problem, which was already solved.
Oh, I'm not complaining, tangential discussion is hardly unusual here. 
I was merely concerned that the OP did not get confused that there might be some other meaningful way of approaching his objective.
Well i would hope not. As long as the output can be the multi-channel PWM that the WS2811 provides, that is obviously the proper solution. If on the other hand it would need to be some other kind of digital format (i have recently been confronted with DMX-LED strip that had been permanently housed within an object for instance) then there is another approach possible though, and once my current projects come closer to completion i am for sure gonna try and see if i can get reception to work reliably.
I would think it would be relatively easy to decode the bitstream into a clock+data (wouldn't a clock at 1/2 of a bit time (625ns) + the original datastream work?), at which point SPI is a possibility. You could probably even do something with discreet logic (although by the time you implement 32 bits of shift register and some clock delaying stuff, that's a lot of chips compared to just using a couple WS2811 chips.
I'm pretty sure an AVR could do in it software as well; especially if you can sit there and wait for data to arrive. The difficult part would be with trying to figure out how NOT to "sit and wait." The time frames involved are small compared to the time to respond to an interrupt, especially if there are other interrupts going on.
Discreet logic would no doubt do it very quietly.