Since the inauguration of this column in the October 2012 issue, we’ve been discussing each section of the modern sampled orchestra in detail. For this final installment of the Virtual Orchestra series, let’s look at a few advanced techniques that can help our scores sound even better. Now that we’ve loaded up our DAW with a great collection of samples and articulations, nothing gets my heart racing more than a big empty arrange window and a deadline.
Guide the Way
Fig. 1. Spitfire Sable is a brand new string library focusing on small sections that is great for focused, expressive writing and for layering over larger sections: four violins I; and three violins II, violas, cellos, and contrabasses each.
When I’m sequencing a standalone orchestral piece that’s not serving as accompaniment to an existing rhythm track (as for a pop song), often I lay down a piano guide first. Although it takes some additional effort on the front end, this has ultimately proven to save me time and serve as a candle in the dark. Depending on the composition and style, it may be just the basic chord structure of the piece performed in a single pass or I may quickly overdub multiple passes on top of one another using duplicate piano patches. If I am hearing (or have written) certain lines or harmonies, I’ll stack and build them into a fairly detailed guide track. I’ll roughly edit the velocities, note-ons, and tempo map as necessary to reflect the compositional intention, slightly exaggerating the overall dynamics of the piece. The point is not to create an actual piano performance; rather to create an accurate guide with a neutral timbre that serves as a master reference for the overall timing, dynamics, and form of the score. (Sometimes I use a soft electric piano sound). I will then sequence the actual parts while listening to the piano guide. It’s not uncommon that I edit the guide track in more detail as I go.
It can be very difficult to start a musical passage from silence when there’s no surrounding context, as “somebody” has to go first. For example, it can be challenging to exactly tell the appropriate volume and dynamic contour of a cello line without hearing the violins or the rest of the orchestra. A guide piano that reflects the appropriate feel makes it much easier to find the musical context for the first few orchestral parts I lay down.
Sometimes I will actually mute existing tracks in a section I’ve already sequenced as I work on a new part, still playing to just the guide piano as a consistent reference that helps me find the proper velocity and/or volume for that particular instrument. This can help keep the timing in check, too; as well as letting me better hear the nuances of the part I’m working on. This takes some getting used to, as the transient attack of a piano can feel inappropriate for some musical passages; however, it’s a technique worth adding to your toolbox as a way to break the silence and get the ball rolling.
The Human Factor
Virtual orchestras generally benefit from some humanization. (It’s slightly ironic that the loftiest goal of a world-class 60-piece orchestra would be to play exactly together and be perfectly in tune with one another.) So, the deviations that occur naturally must be created by the composer and/or be built in to the sample library. Orchestral music doesn’t sound completely authentic without the imperfections.
One of the benefits of stacking individual horn players or separate layers of smaller string sections (see Figure 1) is that your performance will never identically match your previous one. Sometimes simply switching hands when playing a layered part will create interesting differences. Resist the urge to overly edit your sequences early on; allow those small inconsistencies in timing and volume to remain, at least for a while.
Fig. 2. Notice the two pitch randomization knobs in the lower right of Audiobro L.A. Scoring Strings Legato Sordino 2, which interjects small user-definable deviations in pitch from note to note. Any time a library offers round-robin samples, it’s usually preferable to use them, even though they may hit your computer a bit harder. It’s worth the resource usage. If there are any other randomization features in your DAW or library, experiment to hear how they might loosen up the performance. Other than round-robin samples, I generally stick with some subtle pitch randomization rather than modifying velocity or volume (see Figure 2).
Several libraries provide elaborate scripting for creating patterns and ostinatos using a combination of round-robins, sample-swapping, and tiny timing and pitch randomizations. These can be immensely useful for film, gaming, and commercial underscoring.
We’ve touched on this subject before: It’s going to take one or two good reverbs to pull all the disparate libraries together into a unified whole. For that reason, as much as I love the various room mic perspectives in the higher-end libraries, I tend to go easy on using them. I usually lean towards the shorter reverb tails—just enough to give some richness and depth to the samples and no more; as I know I’m going to have to add more reverb later when combining with other libraries.
I primarily use a room sound in a convolution plug-in, like AudioEase AltiVerb or Quantum Leap Spaces, to unify the various orchestral sample libraries into a single space. There are a number of great impulse responses available of various studios, rooms, and concert halls. This is usually followed either by a Lexicon hall to add some height or a plate reverb to add some backsplash, all depending on the nature of the composition.
Fig. 3. Vienna MIR Pro allows you to place instruments anywhere on a virtual soundstage. Venue, instrument position and rotation, and mic choice and placement are just a few of the many parameters. In addition to traditional reverbs, sophisticated room simulations like Vienna MIR Pro are seeing more use by composers (see Figure 3). These let you place each track on a virtual soundstage. Every room choice, instrument position and rotation, and mic choice and placement alters the sound, much like in a real acoustic environment.
Have you ever listened to a great soundtrack as a reference, and then listened to your orchestra mix, and wondered why yours sounded two-dimensional? As difficult as it is to get everything panned into its proper place, it’s still very easy to end up with a final mix where instruments sound either too hard-panned and weirdly disconnected, or jammed together with a muddy middle that sounds like a glorified keyboard ROMpler patch.
One last dirty little secret of the pros is image manipulation. When boxiness is the problem, try applying a judiscious amount of a stereo widener, such as Waves S1. Even better, a plugin with M/S control can help either fill in the hole or clear out the mud in the middle for a more spacious, three-dimensional sound.
There’s no getting around the fact that creating a realistic virtual orchestra is time-consuming and hard work, and even implementing helpful tips and tricks can add more to your workload. Like everything else, the more you do it the better you get. The hope is that you eventually develop a sophisticated, effective workflow that makes someone say, “Where did you get the budget to record that amazing orchestra?”