I realize that some of you reading this post might be in the same "OH S#!T" situation that I found myself in when first attempting to do this, so I'll give you the quick answer first. A detailed description follows for those who are so inclined.
First of all, you must create Roles (or Subroles, if you prefer) for each one of your 5.1 tracks (L, R, C, LFE, Ls, Rs) and your 2.0 track.
Assign the appropriate role for each one of your tracks. Left for the Left track...
Right for the Right track...
Center for the Center track...
And so on. Then, in the timeline, you must change the Channel Configuration of tracks 1-6 (L, R, C, LFE, Ls, Rs) to Center. I know this sounds wrong and counter-intuitive, but you can read the explanation further down in this post.
Then, make sure you keep the Channel Configuration of your stereo track (track 7) set to Stereo.
Now you're ready to export. In the export settings, make sure you set "Roles as:" to "Multitrack QuickTime Movie" and configure your audio export settings as follows:
And that's it. You should be good to go.
And now, for those of you with a lot of patience, here's the long winded explanation as to how I got to those settings.
I recently worked on a TV commercial at ZapBoomBang studios in Houston, TX that was going to have an extremely tight deadline, and because we were editing in a hotel room in Mexico City I was going to have to travel lean and mean. No dual monitors, no Mac Pro, just a laptop. I was also going to have to manage and organize a very big amount of footage, given that this was a car commercial with star talent attached. I usually edit in Premiere Pro, but given the nature of this job I decided to edit on FCPX, due to it's excellent keywording abilities and because it RULES on a small, single display. The fact that I could use Auditions to experiment with different versions of shots also was a nice bonus, plus I am a fan of the magnetic timeline. I find it incredibly fast to work with once you get used to it's quirks.
Anyway, the offline edit went great, everything was done and approved on schedule and it was time to go back home to Houston, TX for the online process. That also went well, but then came the part of having to export masters for trafficking. When we export for HD we usually create our timelines as follows: 5.1 audio as separate, discrete channels on channels 1-6 and then a stereo track on track 7 with a stereo mix of the spot, for those of us who do not have surround systems at home. I've done this kind of export a thousand times on Premiere Pro, so how hard could it be on FCPX, right? Oh boy... Suffice it to say that because we were on a tight deadline I ended up exporting the masters from Premiere, but I was determined to find out IF FCPX could do this kind of export (which now a days is ELEMENTARY). I searched the message boards and I found a lot of confused people but I also found one very detailed post that made perfect sense and which detailed what frankly SHOULD be the way that FCPX exports this kind of file. You can see that post in detail at: http://www.interfacelab.net/blog/files/2c5acbd8fc822b71404c0e5342c02225-34.html
Unfortunately, when I exported following these directions and looked at the waveforms on Premiere Pro, what seemed to be happening was that channels 1-6 were getting mixed down to 1-2 and channels 3-6 were empty. Channel 7 (the stereo mix track) seemed to be fine. This is what that looked like:
So I started to experiment with the settings based off the understanding that our audio guys export each one of the 5.1 tracks essentially as mono tracks. So I decided to create individual Subroles for EACH TRACK OF THE 5.1 MIX (L, R, C, LFE, Ls, Rs).
Then, I assigned the appropriate Subrole to each track. Left to Left:
Right to Right...
Center to Center...
And so on. I made sure the stereo track (track 7) had the 2-0 role assigned to it. Then I changed the Channel Configuration of tracks 1-6 so that they matched their function in the 5.1 mix. Again, Left as Left...
Right as Right...
Center as Center...
And so on. I made sure the stereo track (track 7) was set to Stereo. I then exported the file with the following settings:
When I exported the file, I got this:
For a moment there, I thought I had achieved what I wanted, but when I compared the waveforms closely to the waveforms of the original audio tracks sent by our audio guys, channels 1-6 looked different. I took the file to Mundo Gomez, our audio engineer extraordinaire, so that he could properly compare it to the original tracks. They were indeed different, particularly the LFE track which got pretty much wiped. You can clearly see the difference in the audio peaks of the signals:
So I then figured I'd try to re-export the sequence, but this time, I would change the Channel Configuration of tracks 1-6 to all be Mono instead of their actual 5.1 configuration, just to compare.
I left the export settings the same. This is what I got:
And again there was a difference:
At this point, I was stumped. So I asked Salvador Garza, another one of our senior editors here at ZapBoomBang Studios, who also happens to be an Apple Certified Instructor, to see if he could offer some guidance or insight into this (I know, I should've asked sooner, but I'm stubborn, what can I say...) He immediately told me that it probably had something to do with the "Pan Law" and pointed me in the direction of Larry Jordan. https://larryjordan.com/articles/audio-mono-vs-stereo-levels/
Larry Jordan explains the "Pan Law":
"A recording and mixing principle that states that any signal of equal amplitude and phase that is played in both channels of a stereo system will increase in loudness up to 6.02 dBSPL, provided there is perfect response in the loudspeaker system and perfect acoustics in the room.” (Wikipedia)
In other words, “adding two identical signals [i.e. left and right channels] results in a 6 dB (6.02 dB actually) volume increase.” (Steinberg.net)
Because perfection is almost impossible to achieve, many audio consoles and audio software apply a -3 dB audio level change as a sound is panned from left or right (playing solely out of one speaker) to center (playing equally out both left and right speakers). In other words, the volume of a mono sound playing out a single speaker (panned fully left, for example) will be amplified +3 dB louder than the same sound panned center and playing out both speakers.
That seemed to make some kind of sense, but what was baffling was that FCPX had no way to override or disable this automatic level change. Panning the channel fully to any side didn't increase the volume either.
Then we realized that from all of our tests, the only channel from the 5.1 tracks that seemed to stay normal upon export was the Center channel WHEN WE LABELED IT AS THE CENTER CHANNEL.
So we figured we'd try changing the Channel Configuration of tracks 1-6 to Center, keeping track 7 as stereo.
Mind you, in the Export Settings, tracks 1-6 still have to be designated as Mono like this:
Once we exported, this is what we got:
And when we compared the audio tracks and the peak levels...
Confused? Yes, you should be. It really doesn't make any sense, but there you have it. That's our work around if you need to export SMPTE DTV audio from FCPX. Does anyone have an explanation for the software's behavior? I'd really love to hear it. Or maybe, just maybe someone from Apple will take the time to explain this whole process to us. Because clearly, from what I see in the message boards, there are a lot of people scratching their heads over this issue. I hope this post will help.