When I say orchestral reverb, I’m usually not talking about real concert halls like you would hear in a recording of a classical piece. I’m referring to the larger-than-life movie score sound, which most of us are so accustomed to hearing that we rarely reflect on the fact that it’s not a natural sound at all. Movie scores usually have little to do with reality aside from that it’s real people playing real instruments (and sometimes not even that; if it’s a Hans Zimmer score, then all bets are off).
While classical performances are traditionally recorded with only a few strategically placed microphones that pick up the entire orchestra, most movie scores are recorded in pop/rock-fashion with the orchestra performing at a scoring stage with all instruments mic’d and recorded onto separate tracks. In other words, that massive reverb you hear in big-budget movies is added after the fact and comes from hardware units with scary price tags, not real concert halls.
The reason I’m mentioning this is that it’s easy to get caught up in the whole orchestra = realism thing, and it’s important to keep a couple of things in mind. First of all, the orchestral sound most people are used to hearing is not the classical sound, so don’t lose any sleep over the fact that your reverb can’t sound exactly like Wiener Konzerthaus or whatever. Secondly, sampled instruments are not real instruments and can not be expected to sound their best when placed in an environment suited for real instruments. Paradoxically, a fake orchestra often sounds even more fake with a highly realistic reverb. To help mask some of the glaring “MIDI-ness”, a reverb that plays more loose with reality is often called for.
Live orchestral music owes its special character to the fact that it’s played by large numbers of real and very skilled musicians, playing top quality acoustic instruments in a concert hall with exquisite acoustical properties. We’re trying to simulate this music by using no live musicians, no acoustic instruments, and no concert hall. It comes as no surprise that not even with cutting edge technology can we make a trained ear mistake it for the real thing. Orchestral music is just too complex and full of nuances to allow for easy faking.
As long as composers are stuck with using MIDI there will always be an unnatural, disjointed sound to the music we make. Having state-of-the-art sample libraries with legato sampling and ten million articulations, round-robin variations and velocity layers does indeed help believability. But given the cost of libraries like that it can’t be considered a viable solution for anyone but the small percentage of musicians who compose for a living. Everyone else — hobbyists, amateurs and semi-professionals alike — will have to look for other ways of making their music sound as good and convincing as possible.
While I definitely think you should aim for a certain amount of realism, be aware that realism in itself will not make your music any better unless your target audience is other composers with a similar obsession. Strive for realism where it counts. That is, composition, arrangement, performance, sample quality, in that specific order. If you have nailed those bits, choosing a reverb should not come down to what sounds most realistic, you should be looking for what makes your composition gel and sound like a pleasant whole. I’m not saying that the answer can’t be a realistic reverb, just that you shouldn’t automatically assume it is.
Next: Orchestral Positioning: Choosing a Reverb
Great article!
If only I had realized this a couple of weeks ago… Many hours were wasted trying to get a “hyper-realistic” sound stage ambience with virtual instruments. I went as far as wrote a small program to calculate per-instrument reflections based on their room position. Needless to say, it sounded like crap. Could have spent the time actually making music instead.
Thanks Otto.
I guess there is nothing like first hand experience huh? 😉
Just out of curiosity, what reverb were you using?
Unable to afford Vienna MIR or Altiverb, I first trialed VirtualSoundStage 2, which isn’t that bad actually, it just doesn’t have reverb tails. I then stumbled upon Voxengo Impulse Modeler (a graphical tool for rendering custom impulse responses), which seems cool but it didn’t run well under WINE (I’m on a Mac). Using roughly the same principle, though, I wrote a small piece of code which took a listener and a source position in a rectangular room and calculated the delay time, direction and gain reduction of each reflection up to a certain depth, adding “spikes” to an impulse response accordingly. The math is simple and looks pretty on graph paper, but… It wasn’t practical. Yes, the spatial positioning of different sections could be heard, but no amount of tweaking the frequency response and progressive diffusion of reflections could get it to actually sound good. And with the CPU usage constantly peaking due to each track having a separate convolution reverb instance, it just wasn’t worth it.
I switched to Apple’s complimentary Matrix Reverb plugin, which IMO is something of a hidden gem.
Really eye opening. I’m very excited to apply these principles to my current project. Thank you very much. I aspire to be professional instead of recording only when it’s dark out and the family is asleep, and I need all the help like this I can get.
Thank you very much for taking the time and effort to publish this series of articles. As you know, the music industry is not driven by those of us who deal in orchestral music or, as in my case, trying to emulate concert bands. Your advice is greatly appreciated. I’m looking forward to getting in the studio and applying some of the techniques you have outlined.