Comparing Personal Orchestra to the multi-gigabyte samples in your previous products, isn't there a discernible difference in realism?
It depends. Maybe you'll notice if you're a stringed instrument purist, but that misses the point. Are we in this for complete orchestral replacement, or are we in it to inspire musicians in their knowledge and compositions? Garritan is not in the business of replacing an orchestra, though it's possible to do that to some extent with GPO. We've proven with new sampling technologies and clever programming techniques that we can provide very realistic orchestral and solo instrument mockups without requiring massive amounts of storage and charging musicians thousands of dollars. We wanted to start the pendulum swinging back the other way, to a philosophy of "less is more."
That must have been quite a surprise to the industry.
Everyone was expecting us to come out next with another super library that was even larger--a gargantuan terabyte library. Instead we went light, sweet, and small.
But wasn't the point of making ever-larger samples to capture more of the performance nuances of live acoustic players?
To a point, sampling can play back the sound of discrete notes, but it cannot alone provide performance nuance. Let's say you record every note on a Stradivari violin. Each time you play a note back you're stuck with that same exact note, which produces the dreaded "machine-gun" effect--every time you come to that note in a piece, it will sound exactly the same. It doesn't sound human; it's too perfect. Even if you sample the note multiple times, you still have the same problem with a finite number of variations.
So how do you get the expressivity in there?
GPO's programming nuances and performance tools allow you to vary a note each time it is played by introducing automatic changes in attack, timbre, pitch, or other variables. You get an infinite variety of note variations at a fraction of the storage size. You can play a single note and let intelligent, clever programming do the rest every time you come to that note. Approaching it this way, a composer can do so much more, like impart subtle nuances and perform on a virtual instrument like a real player does. I must add that much of the credit goes to Tom Hopkins, Jeff Hurchalla, and David Viens for making these innovations possible.
Do you find that traditional orchestral composers and performers fully embrace computers for music?
Yes; the early adopters of notation software were, of course, the education and orchestral communities. Notation programs have always been great for scoring and providing parts to live orchestral players. But until recently, they didn't integrate well with sampling technologies. So they didn't allow a composer to "play" what was on the score without diving into DAW [digital audio workstation] technology, which most composers didn't do for a long time.
What is the state of notation software today? Is it accurate enough that a composer can confidently hand a chart to a live orchestral performer?
Oh, yes. Notation programs have been quite accurate for scoring for some time now. Notation and DAW technology evolved along different paths, but now those paths are starting to converge. We're beginning to see sequencers and DAWs becoming more robust in their notation, and notation programs are doing the same by integrating internal sequencers and mixers. I think we'll even see notation programs adding multitrack audio recording soon.
How are notation and sample libraries converging?
Notation is the language of music, but sample libraries were originally developed as standalone [data for hardware samplers] and then later as plug-ins for sequencers. Until recently the notation community has been ignored. Garritan Personal Orchestra was the first sample library to make a commitment to notation users--it was specifically designed from the ground up to work with notation programs. Finale 2006's integration of the GPO library is a good example of that convergence.