Part 4: Music Editing
When recording people individually to create a section it becomes apparent just how much great musicians and singers do practically instantaneously and often automatically and subconsciously when playing or singing together. From phrasing, to intonation to breathing, when a group performs together they practically become a single organism – it’s really extraordinary. It’s magic. But when recording people independently and putting them all together none of these things can happen, and you’ll need to address them yourself.
Phrasing
As discussed previously, indicating phrasing clearly in your notation is important. But once you get your tracks and put them together you may notice that despite your best intentions different musicians still interpreted them slightly differently. Some players or singers may be right on the beat, others a little behind or ahead. When building virtual sections or choirs, you’ll need to spend some time cutting the tracks to better match the phrasing of each musician so they sound together as they would if they were playing together.
For choirs another consideration is breaths. When singing long notes, choirs will stagger their breathes, meaning different singers will breathe at different times so the overall effect is a continuous note. But when recording individuals remotely they can’t do that. I discussed this a bit earlier in the blog, but if you get too many singers taking a breath at the same time you may need to move where some breaths happen by shortening the A side of a note before the breath and then lengthening the B side after the breath to manually stagger the breaths.
Another note about breaths, when the choir all breathe together per phrasing, using multiple individual tracks stacked together can really amplify those breaths. As I discussed in Part 1, I’m not a fan of eliminating breaths, but you’ll likely need to reduce them in each track so the accumulated volume is reasonable and not too loud.
Intonation
Intonation can be a big issue. What I learned is that often even if an individual recording sounded great with my mock-up, as I started stacking strings or singers together intonation became a little wonky. For most instruments pitch is affected by the performance. Strings players may slide into a pitch (sometimes obviously, sometimes ever so slightly), especially when leaping up or down the bridge. Singers may swoop into a pitch, again especially during interval leaps. Individuals hear pitch ever so slightly differently and some may tend to be a few cents sharp of flat, or more commonly vary a tiny bit while playing sustained notes. This is normal and part of the magic of live performance. When they are together they listen to each other and constantly adjust and become one.
But that’s not the case when they record individually – how could it be? So you may need to spend time doing some pitch correction. I mostly use Auto-Tune Pro in it’s advanced graphic mode. This creates a visual representation of the pitch in a piano roll and let’s me see exactly what’s happening. This allows me to surgically tweak just the parts of a note that need tweaking rather than making global pitch adjustments, which can sound processed and unrealistic. I don’t want to lose that natural swoop or slide, or the shape of the note. I want to make sure the pitch center matches from one track to the other. Sometimes that means only affecting a part of a note, other times it may mean ever so slightly raising or lowering an entire note or even a phrase. Basically I have to make those minute adjustments the singers would make on their own.
In the above screenshot you can see 3 notes from one of the choir singers that I slightly tweaked. You can see I didn’t flatten the swoops at the beginnings of the notes, or try to flatten the pitch or vibrato – you can see the vibrato increases at the end of the note. But I did slightly adjust this to center the pitch. I did this type of work over and over to improve overall intonation. You’ll notice in this case (top left) that for input type I selected Alto/Tenor since this was an Alto. Though I found picking the matching input type isn’t always the best choice depending on the timbre of the voice. So don’t be afraid to try the Soprano setting on a Bass or vice-versa. Most often the best choice will match, but if you’re not happy with how it sounds try a different setting. For example for women with deep voices the Low Male setting sometimes worked better than the Alto/Tenor setting even though what they were singing wasn’t that low. Sometimes the Instrument setting isn’t the best for certain instruments. Don’t be afraid to experiment. Whatever you’re doing, if the result doesn’t sound natural you’re probably doing too much.
When making this video I worked with a mix of professionals and amateurs so on occasion I found a mistake, a wrong note or a level of pitchiness that I couldn’t correct without it sounding mechanical. I found that creating a 2-3 second fade out before the problem moment, muting the bad moment and then having a 2-3 second fade in after the moment in questions worked well. It masked that an individual dropped out momentarily. When listening to the entire section it was completely imperceptible.
If you have the time, it might be worth recording one singer for each part and working with them to nail it. Then you can send that recording along with your mix-minus (or incorporated into your mix-minus) for the rest of the singers to sing with. Same thing for any other section, especially strings. That way the rest of the players/singers are working to match the lead performance rather than your mockup and you may get better results.