(cue hyperaddictive AC/DC song) When Jered, Carrie, and Gad emerge from the secret tunnels, they find themselves in the midst of an approaching storm, which of course works out very well for helping to cover the sound of their movement. The sound of rain striking the ground is a different layer entirely which I address below, but for now let's focus on the sounds emanating from the heavens.
One ambient layer sits pretty low in the mix, this is just general storm noise, occasional rolling thunder, intended to sound somewhat distant. The thunder in this background layer does not correspond to anything visual; as in a real storm, not all lightning is visible, though we may hear its retort quite well.
For the actual lightning flashes on screen, I had a few different thunder crashes at my disposal, some more intense than others. These were timed to be right around two seconds from when the flash is seen. Hence, if this WikiHow article is correct, then the strikes are occurring less than a kilometer away. Those who watch BloodSpell with a stopwatch in hand (I know I do!) will observe that over time, through the end of Episode 10 and on into Episode 11, the gap between flash and thunder gets smaller and smaller as the storm inches closer. It's little touches like these, which admittedly won't be observed by all but the most obsessive-compulsive viewers, that nonetheless help this fictional storm graduate from inert background noise to almost living thing.
Another thing fellow Monkites might notice, the sound of the magic when Jered and Carrie kiss is slightly different than the casting sound we've heard previously. The core sound is still present, but there are some new layers. I wanted the "smooch-casting" to have a slightly different flavor than the magic which actually involves spilled blood, and found the perfect multisound in Absynth to beef it up yet still keep the magic sounding "musical." I think it came across well, what do you think?
Rain, Rain, Go Away!
Complex soundscapes can be fun to build, but they have their problems. If you can't hear precision footfalls over the sound of rainfall, that's not really a big deal, it plays to the realism aspect. But this isn't just about realism, it's cinema; and if the dialogue is having trouble competing then you've got a more serious issue to deal with.
Such was the case for our friends in the rain. It was very difficult to make out what was being said, and that of course will not do at all.
75% of addressing a situation like this is gaining an understanding of WHY the vocals are having trouble being heard. The gut reaction is to just bump up the fader to make them louder, but that doesn't help as much as you'd think, and risks throwing the sound balance of the episode off its rails.
The vocals were competing with background thunder, rainfall, and torchlight and braziers in some spots. There was no significant reverb on the vocals, so I knew that wasn't clouding things up. The fire sources I could rule out right away, because they were very low in the mix (almost subconscious). So that left the thunder and rainfall as the two possible culprits I wanted to examine.
It is important to understand what qualities of human speech give it its clarity. It has very little to do with how loud someone is talking (ever notice how Trent Reznor can whisper and it comes right through the mix?). Consonants and breathiness are the clarifying aspects of speech, and these occur at the higher end of the frequency range. This sibilance range, it turns out, was right in the same pocket where our nice crisp-sounding rainfall was doing its thing. So all that was required was pulling back some of these frequencies in the rainfall layers (layered by the same principle we used with crowd noise in Episode 8) to allow the vocals to pass through with less competition. No adjustment of EQ was made to the vocal tracks themselves. (Remember: EQ, and any other effects, are a distortion of signal, so if you have to apply them, apply them first to the less important sound).
At this point, problem solved, I became a little agitated, because I was really liking how that rain was sounding before I'd muffled some of its finer frequencies. So with a little fidgeting, I came up with a way to turn the EQ adjustment on and off as needed via Cubase automation, timing those switches with camera changes so they feel like the natural result of setting up the camera in a new spot. Oh man, I love machinima.