If you email me the learned commands for BMC9012, I can probably decode them. If my theory is correct, that's probably all I'll need. My email is in my profile.
As far as the NEC projector I recall seeing some long threads in the Pronto forums but IIRC, the methodology was to paste on a whole string of repeat commands and adjust the parameters to fool the Pronto. The real question will be how long of a Pronto hex string will the RTI's editor allow. The newer ProntoNG's apparently won't work with the trick that worked on the original 1000's and 2000's because of changed software/firmware.
The second issue is that John Fine who decoded the cml files about a year ago came to realize that the Pronto hex conversion in the RTI-TT was flawed and would only convert a maximum value of 65,536 microseconds (uS) (or ~66 microseconds(mS)). Worse yet when it exceeds that value it starts over.
This is important because the NEC1 protocol used by the NEC projectors has a first frame that includes 32 data bits that is sent once and then, for as long as you hold down the button, it repeats what is known as a null repeat frame that contains no data and is the same for every command. Both frames must be pretty close to 108 mS long. If the IR receiver sees the null repeat frame within 108 mS from the start of the previous frame, it interprets it as a held button.
The signal for the repeat frame is +9000, -2250, + 564, -96168 (+ means carrier On, - means Off). So unless that conversion program has been fixed the RTI-TT would put out +9000, -2250, + 564, -30650 and I don't think that would be recognized as a valid frame. All timings in uS.
So this is probably more than you ever wanted to hear about the subject, and I hope someone else with more experience with the hardware has a better workaround for you.
