I’m really digging the Native Instruments software I’m using lately. I haven’t had time to really delve into the nuts and bolts of Reaktor 5, but am eager to. I’ve been using their synthetic drums soundpack (and have the bigger volumes 1 and 2 on the way for Kontakt), and have been relying a lot on Massive. I’ve been a fan of the B4 for a few years now.
Aside from the NI stuff: I’ve also found my new favorite electric piano soft-synth: Lounge Lizard. It’s great. Still not the only one I use (I like the MrRay73 MkII, and a good free one is LazySnake). Another soft-synth I use a lot is Arturia’s Moog Modular. For orchestral instruments and piano, I tend to use the Garritan orchestral library, though one day would like to get some of the other nice ones ($$) out there (I’m also curious about the Vienna samples coming with the Kontakt 3 update). Actually, I bought some dedicated piano libraries that were supposed to be top notch, but ended up still using the one that came bundled with Garritan.
Speaking of coputer-based music tools, a couple of people have asked where I get my loops from for the live set (some songs have tracks coming from a laptop). I don’t use loops! I put everything together note by note into Cubase 4, and that includes the arrangements that have drums. Each drum hit is usually put it one at a time (I play it on a MIDI keyboard and then go in and tweak). Some background stuff works well copying and pasting an entire section that I’ve already created, and then maybe altering that a bit, but any brief repeating pattern will always be newly played each time. My music isn’t generally repetitive, though. If something does repeat, I like it to be a little different each time.
Our minds pick up on the texture that comes out of the differences from instance to instance. Texture in general is really important in arranging, recording and mixing music, whether it be a product of distortion, dissonance, instruments being slightly out of tune and off beat from another (100 violins with the exact same material construction, playing perfectly in unison from the exact same position would just sound like one loud violin)… and the kind of texture I’m talking about in regards to not looping is more linear: subtle changes in timbre and dynamics from instance to instance.
Side note: A big part of music is remembering, if not consciously then subconsciously; which means taking in new information and comparing it to information you just took in, or took in much earlier in the song. This is one reason why some music has to grow on you: the relationship of occurrences (relationships between pitches, durations of sound and silence, dynamics, harmonic elements, melodies, timbres et al….) in the song or arrangement is not immediately apparent, but as you are increasingly able to remember the occurrences and compare them to one another, you are able to see the piece as a whole etc… immediately catchy tunes are usually comprised of very simple elements (most hooks are 2 – 4 notes), but because there’s very little to remember in order to compare and contrast the various elements of the song, it becomes boring quickly (hence the term “throw-away music”; the best thing that can happen for the longevity of that kind of song is for it to get some kind of sentimental value that triggers nostalgia later on).
Anyway, back to the idea of texture… Another example: if you have 4 backing vocals, you can group them and send that to a reverb (works in a pinch), but I like to send them individually with slightly different send levels. It’s very subtle, but is closer to how things work in real life. You can also use eq to give the impression of each voice coming from a different place in the room (even if it’s 1 foot different) and facing a slightly different part of the room. Recording the vocals at slightly different distances and angles from the mic helps for that as well, especially if you have a nice sounding room (even if the mic is in a cardioid pattern). These things are subtle, but accumulate into a more textured and believable whole.
(I’m really interested in what other people are doing to achieve these kinds of effects. If you have any techniques or ideas let me know!)
One analogy I like to use is a CGI forest. The ideal for a CGI-made forest would be thousands of unique trees, each with thousands of unique leaves. When the wind blows, each leaf would react to the wind individually. From a bird’s eye view of an actual forest, we don’t see each individual leaf, but we perceive an accumulative effect (texture) that we can’t really put a finger on, but we will miss when we see a crudely animated CGI forest. The extreme example would be if the animator made just one tree and copied and pasted that 999 times. There’s no way this will come across as realistic. The more detailed they can get the better, even though we can’t distinguish the individual details. That’s how I feel about mixing music. It’s a fabricated world or environment, but still needs to be believable. The only way to do that is through the accumulation of individually indistinguishable (and generally unquantifiable) details. A little extra distortion here, slight dynamic variation there, getting a little offbeat over here etc… the importance of not having everything perfectly on beat can’t be understated. If you have 5 drum sounds, and their attacks occur at the exact same time, some elements might get masked that would be better off being just slightly behind the beat. Like a hihat with a short decay. If it has a little bit of a sharp attack, it’ll stand out more if you put it a little ahead or behind the beat. So long as it grooves (and “groove” is often defined as being offbeat in a way that works), anything goes timing-wise (this is really a question of phrasing; phrasing is essentially the application of an individual performer’s interpretation of rhythm, pitch (such as sliding into notes) and dynamics; with an orchestra, it’s up to the conductor).