The following tutorial’s creation was inspired by a conversation ScottSandwich and I had in the C3 Authoring Discord a few weeks ago. Scott was having an issue in which the “singer looks like someone’s shoving an invisible burger in his agape mouth during the first line after an instrumental part, instead of his mouth moving with the dry vox.” He remembered me remarking about having the same problem and fixing it by using Magma’s MIDI override function. I said I would probably whip up a tutorial on how to do that in the near future, and wouldn’t you know it? I’m updating an old release that had that same issue. So here is how I’m about to solve that problem, in as few words and screenshots as possible:

1. The first thing we want to do is clone our completely finished and finalized PART VOCALS and HARM1 tracks, then MUTE the clones. Is it necessary to clone HARM1 in addition to PART VOCALS? I don’t actually know, but it’s less effort to do it to both tracks than to test it and find out. ” src=”/wp-content/uploads/invision_emoticons/default_SA_smile.gif” />

   

   It will also be necessary to do this for the other vocal parts if there are backing harmonies. For this song, I indeed do have HARM2/3 singing the same part, even though they aren’t the focus. But in this case, HARM2/3 share the same dryvox, so I’m only going to do this for HARM2.

2. Next, we’re going to implant a snippet of both the dryvox and vocals from a different part of the song into PART VOCALS and HARM1. We’re going to plop those two pieces somewhere slightly before the spot where the lip sync is broken (which usually happens after a medium-to-long instrumental break). Try to use a small phrase. I have a conveniently sized “Ah” from HARM2/3 at measure 127 that I will be using. It looks like this:

   

   Here’s where we’re putting it. The angry face is where the vocals are broken, even though I recorded them 8,235 times.

   

   And here it is the result. PART VOCALS, HARM1, and HARM2 now have lyrics that aren’t really in the song. Don’t worry, those lyrics won’t stay there for long!

3. 

   …But what will stay is the lip sync we added. So you’re going to want to avoid using camera cuts that involve the mouths of the singers. In this situation, the guitarist/bassist/vocalist may be lip syncing the “Ah” we implanted. And that “Ah” may be mangled just as the beginning of the phrase “Maybe you will always be” was getting mangled.

   This is less of a nuisance when it’s just PART VOCALS and HARM1 singing without HARM2 and/or HARM3, because you can just avoid showing the vocalist. But if you’ve got the situation above as I do, you’ll probably want to focus on the drummer in the lead-up to the problem area or do some long or behind shots. Conveniently, I already had that spot covered with space_woosh ” src=”/wp-content/uploads/invision_emoticons/default_SA_biggrin.gif” />

   Anyway, I digress.

4. Now we render our dryvox and and export our MIDI as usual. Almost done! Now the critical part: we switch what’s muted and re-export the MIDI with the title override.mid. Like so:
   
5. Our project now has two very important MIDIs. The altered original and the new override.mid. The first MIDI is used by Magma when making the dryvox into lip sync data. For whatever reason, Magma needs that early dummy phrase so that it isn’t screwy when you want the singer to start lip syncing correctly. Once Magma generates the dryvox with the first MIDI, it uses the second MIDI, the correct one, to replace only the MIDI, removing the dummy “Ah” we added. It does not regenerate the lip sync. Both MIDIs should be in your project directory. So let’s crack open Magma and enable the override:

   

6. Now compile your custom and test it out! Your vocalist should perform as you expected him to before everybody blamed you for recording crappy dryvox.

Hope this helps somebody, because last year I was pulling my beautiful hair out. Cheers!