One of my big goals for the year is learning orchestration. I am at the start of a year-long class and currently focused on strings.
The class teaches how to write for live orchestra and also how to write for MIDI-generated, sampled orchestra. For those of you that are not sure what a sampled orchestra is, here is a simple explanation. There are companies that build libraries of orchestra sounds. For example, they take a violinist into the studio and have her record every note in the violin range with different articulations and dynamics. A violin has numerous articulations so you can only imagine how long that takes. Then they record multiple violins together doing the same thing. There are other things they do too but I will not go into them here. Suffice it to say that just recording violins for a sample library is a huge endeavor.
Of course, there are a lot of instruments in the orchestra spread out among strings, woodwinds, brass and percussion. By the time it is all said and done, there are millions of samples that take up hundreds of GB of hard drive space.
Once all the samples are recorded, those sounds can all be combined together in music. MIDI is the format of a file that contains a song of notes, dynamic levels, tempo, etc. In software known as a DAW, samples can be assigned to every note and a song comes to life.
There have been many times over the years where I have laughed at the idea of doing orchestration that way. A live orchestra is unquestionably better in many ways. In spite of the enormous numbers of samples in a library, it is still not possible to capture all the nuance of a great violinist.
On the other hand, over the past ten years, technology has changed to the point where sampled orchestration is no longer just used on budget projects. In fact, the biggest movies hitting the box office are being produced with sampled orchestration. Look at it this way: if a movie with a budget of hundreds of millions of dollars uses sampled orchestra, it is sort of silly for me to laugh at it.
The other thing that changed my mind for good on sampled orchestra was my production of the Danny Craig project last year. All of the big songs on that project were built with sampled orchestra, masterfully done by Ben Botkin. I have to admit that those tracks are better than any tracks I have worked with even though all of my experience up until now has been with live orchestra. (You can hear them here.)
So why would sample orchestra be better? The most obvious answer is that you have more possibilities. If you want 60 string players, you get 60 string players. If you want dozens of unique percussion instruments, you have them. Granted, you can do that in real life too but unless you have unlimited money, the financial side of things will limit your orchestra. With samples, the only thing holding you back is the power of your computers.
The other big advantage of a sampled orchestra is the precise control you have over everything. Every single factor in the music can be tweaked easily. There are never mistakes that can’t be fixed. Never does a song end up being too slow or too fast because a simple adjustment modifies the tempo by a few beats per minute.
So, while the music I am writing would work for live orchestra or sampled orchestra, I am currently working with sampled orchestra. And what I am finding about working with sampled orchestra is that the technology is almost as complex as the music. In fact, writing the music has been easy compared to the technology issues.
For those of you that know this world (or care to know more), here is my setup. I have a 27″ iMac running Logic Pro X. The machine is fast and I upgraded the RAM to 32 GB. The sample library I am using is EastWest Orchestra Diamond. I know that eventually I may need more libraries but that one is enough for now. I have the library on a SSD (solid state drive).
When you consider that you might use 4-8 articulations for each instrument in an orchestra, you end up with a lot of data in memory at any time. In fact, a fairly basic template of instruments with a variety of articulations may use more than 10 GB or more of RAM. The CPU power needed to run that kind of thing is fairly extensive as well. A lot of my time has been spent on trying to get my already fast iMac to run faster and more efficiently. I don’t want to have to buy the absurdly-priced Mac Pro but sometimes, I wonder if that is inevitable.
Here is my workflow at the moment: I write in Finale and then either play each instrument into Logic Pro through my keyboard or just export from Finale as MIDI and import the file into Logic. There are many people that can just play lines in to the DAW without writing them out first. I can’t do that because invariably, that leads to me leaving out necessary notes in the chords or doubling others. I suppose I may get there eventually but at the moment, I can’t even imagine it.
I guess the main thing I have to say about sampled orchestration is this: it is just enormously complex, especially if you are trying to do it well. I do think it is a skill worth learning but I am learning slowly.