As far as I can tell the program works so far since I am able to check the output in AnalasysIR.

AIRemote.h is just the experimental branch of IRemote which I'm just using to test a few things with at the moment.

In the code above anything labeled btn in those arrays are based on the formula
X * 138 + 758

So the array btn 1 would translate to

896, 1724, 758, 2828, 1310, 1310, 1172, 2690, 896, 2690, 758, 758, 758, 896, 758, 758,

The code then populates an array with 208 on signals and the half (13800) and lead out (80400) timings. (I'd like to point out I didn't know anything about IR codes a week ago incase the lingo Im using is wrong)

208, 896, 208, 1724, 208, 758, 208, 2828, 208, 1310, 208, 1310, 208, 1172, 208, 2690, 208, 13800, 208, 896, 208, 2690, 208, 758, 208, 758, 208, 758, 208, 896, 208, 758, 208, 758, 208, 80400

Which is pretty close to the codes from irb.tk under comcast xmp-1 raw

210, 896, 210, 1712, 210, 760, 210, 2800, 210, 1304, 210, 1304, 210, 1168, 210, 2664, 210, 13800, 210, 896, 210, 2664, 210, 760, 210, 760, 210, 760, 210, 896 , 210, 760, 210, 760, 210, 80400

Or the codes I pulled (saved here) from AnalasysIR that I averaged out from 10 tries

224, 893, 224, 1698,  227, 738, 226, 2784, 226,  1280,  226, 1280, 226, 1151, 226,  2658, 225, 13779, 224, 892,  225,  2788, 225, 736, 225, 764, 225,  736, 224, 736,  226,  731 ,223, 735,  225, 80351

Basically I went this route to learn a little bit of programming even if it is a bit terrible and I'm aware there are probably 100 easier ways to go about it. Plus it seems to save memory and its very simple to tweak all the values with a few integer changes.

Only thing I'm really worried about is the codes I pulled myself seem to be off. I'm not sure if its the receiver/transmitter or both. I'm using two arduino unos at the moment. One to send with the code above and one to receive hooked up to analasysir. Also I used my own xr2 remote to generate the list of codes on the spreadsheets.