So I want to record some dialogue that will be spoken over music. I want to be able to raise the level of the music at some point for an emotional effect as well; without making any of this muddy.
 
Is sidechaining a multiband compressor using the voice track on the music track the right way to go?
the instruments will consist of orchestral instruments.
 
Also I'm looking at the LP64 multiband and wondering how I will sidechain it.

For something basic, a noise gate often works well and is more predictable. Just make sure the gate closes down only a few dB (whatever is necessary), and be judicious with attack and decay times. Tie the reduction amount to a control surface and then you can bring it up in real time, assuming that's how you want to handle it.
 
You may need to delay the audio compared to the gating signal so the gating happens just before the audio would normally trigger it.

Wait what? What I'm going for is I'd love to have that effect where the music swells and is almost equally as loud as the vocals for a finale of the song. While still remaining clean.
 
The music is already in a stereo wave file so I dont have access to tweak and change the instruments or elements of the song separately. So I though I'd just dip some frequencies when the voice comes on and carve out a space.
 
Can you please explain in further detail what the noise gate does to create this or a similar effect? I'm still learning.

ok,  you know the "ducking" effect you hear on radio when the DJ talks over a record?  You want to be using a very mild version of that.
 
How?   put your vocal on track 1, your pre-recorded music track on track 2
Put a compressor on track 2 (the music) with track 1 (the vocal) as the side chain input.
set the max reduction of the compressor to just a few db down from normal.  This gives the subtle reduction during the spoken parts,  adjust the attack and release times on the compressor to make slow smooth fades.
 
Automate the overall volume level on the Master so you can have your big swell at the end.
 
 
Edit because I'm stupid
 
 
Also edit to add,  You should NOT use a multi-band compressor.

If you're bringing up the music entirely for emotional effect, then simple volume automation is the way it's normally done. Long, smooth volume increases that start just before the end of the dialog. Just listen closely to any big-budget movie or TV show to hear it being done. It's rarely automatic via a dynamics processor, though, except in radio commercials that are knocked out by the dozen in a day.
 
When you do listen closely to movies, you'll notice that the music is NOT always brought up during every dialog silence. Doing so would draw the audience's attention to the music, but you probably want it to be subliminal. That's why it has to be done by hand, not with a sidechained gate or compressor.
 
If, on the other hand, you do want to duck the music to enhance the dialog's clarity, then that's another thing entirely. In that case, you want to duck only those frequencies that mask the dialog, rather than a broadband gate or compressor. Otherwise, it'll sound like a radio commercial. For that, you'd be on the right track with a multi-band compressor, but there are even more elegant and transparent methods.

I don't wish to argue with Dave,  but,  ducking with a multi-band compressor is good for integrating vocals into a song for exactly the reasons Dave says...  However, for narration or voice-over on top of music, you need full band gain reduction music to make the narration/voice-over stand out. Hence a standard compressor.
 
It'll only sound like a DJ talking over a record if you overdo it, it needs to be subtle.

Karyn
for narration or voice-over on top of music, you need full band gain reduction music to make the narration/voice-over stand out. Hence a standard compressor.
 

This is also why a gate can work well. Also a lot of gates have a hold time control so you can keep the level down during brief pauses.
 
Really there are a lot of ways to accomplish the basics of what you want to do, but some methods may yield better results than others under specific conditions of program material and timing.