I saw something very interesting on twitter, the other day. It was a video, posted by an accomplished cello player. It was interesting because of what it stood for.
The popular “Game of Thrones” series, made by HBO, has a theme tune that is very emotive, but the main cello part is not played with a real cello, using a real cello player to play it. Instead, it is a sampled instrument, from a sample library, which the composer has used to give the impression of a cello. It’s a reasonable facsimile, as sampled instrument parts go. You can tell that it wasn’t played by a cello player, but it’s not too bad, in my view. It works.
What the video was about was a complaint that the cello part, in the theme music, lacked any sort of passion or dynamics. The complaint is correct. The synthetic cello part is, indeed, flatter and less dynamic, compared to how a real, accomplished cello player would play it. To prove the point, the cello player who made the video offered his own remake of the theme tune, but playing the cello part himself, putting all the passion and verve into it that he thought the piece really ought to have had in the first place. To his credit, it is a good cover version. It is very good, in fact. Unfortunately, he misses a critical point.
I remember when sampled instruments were first introduced. I worked, for a while, for a company that pioneered sampled instruments. At the time, the musician’s union was up in arms, due to the perception that sampled instruments would be a cheap alternative to human orchestras. Their charge was that orchestral players would join the ranks of the unemployed. There was some merit to their fear. On the other hand, there were composers only too keen to have sampled instruments, because it meant they could use timbres, in their compositions that were way outside of their budget. They simply couldn’t afford to use human orchestras. To these composers, it was an opportunity to improve the quality of their recorded output, without having to fund it.
Of course, at the time, buying a sampling machine cost about as much as a small house, which meant you had to be quite keen to justify the expense. You could buy a lot of orchestra time, for that money. As with all technologies, the price eventually came down and it was possible to obtain sampled instruments very cheaply and sometimes even for free. The orchestra was seemingly finished. Who would need these players, with their comparatively inconvenient terms and conditions, when you could record with a sampled instrument at any time of the day or night? Maybe the musician’s union had a point.
Interestingly, now that sampling is over three decades old, heading for its fourth, we have learnt a few interesting things about sampled instruments. It comes down to this: there is always a cost associated with putting the passion into your composition.
As a composer, you have two choices, when it comes to making highly emotive, passionate music, played exquisitely. Your first option is to pay for the human orchestra and take advantage of musicians that have spent a lifetime learning to articulate their playing, to provide maximal audience impact. Junior orchestras won’t do. You need really accomplished players, who have played and practiced for a very long time. They also need to have worked out their instrument and equipment, so that they can produce the best sound reliably. This is the school of thought of the real cello player, who made the video remake. The first option, obviously, costs a lot. Not only do you have to pay the orchestra, you also have to hire a big recording studio to record them. It’s not a cheap way to realise your composition at all.
The second option is to stick with sampled instruments. What composers who use sampled instruments quickly discovered, however, is that you need a sample library that has what they call “deep sampling”, where a lot of different dynamics and articulations are recorded, so that the composer can use these subtleties of instrument performance, in their compositions. Those libraries cost a lot and use a lot of computing resources to run properly, so you have to spend quite a bit of money just to get to a point where you can produce passionate music with sampled instruments, displaying all the subtle articulations and using the dynamics captured by the sample library.
The next thing a composer discovers is that, in order to effectively use all those lovely articulations and subtleties, they have to play keyboards as well as a good cello player plays their cello, for example, so that the subtleties can be captured from their keyboard playing. That means they will have had to invest as much time learning to play the keyboard well as the cello player would have, throughout their lifetime.
Alternatively, composers must painstakingly craft each note in their composition, using a computer keyboard and mouse, carefully adjusting the performance parameters that operate the subtleties available in their expensive sample library, from within their DAW. This takes a lot of time and their time has a cost associated with it, too. They also need to have had quite a long and extensive (and expensive) musical education to think in terms of articulations and dynamics. There is a cost associated with learning how to think like an instrumentalist, playing their instrument in a passionate way. Because of the translation between piano keyboard, or mouse clicks and the strings and bow of a cello that is involved in creating the passionate performance, it is arguably even more difficult than learning to play the instrument in question, well.
To summaries, then, the cost of adding the passion to the music, whether you use real orchestras and human instrumentalists, or an expensive sample library, with which you have to enter all the gestural articulations, by hand, to get the performance to sound right, it’s expensive. There isn’t much cost difference between the two approaches, if you want to get the same quality of performance. The difference is in who gets paid.
The problem is that many soundtracks are produced on skimpy budgets. Sampled instruments are used to make the soundtrack deliverable at a lower price. The passion is what gets cut out. The composer, without the time and money to put all the subtleties and performance gestures in, is rushed and so produces a very static sounding work. It means he has more or less wasted his money, in buying a sample library that provides all the articulations he has no time and money to actually use. A cheaper one would have sufficed.
Consequently, what that cello player’s video was railing at was the wrong thing. His enemy wasn’t the composer who used a sampled instrument. His enemy was the series producer that provided a very low budget for the theme tune’s composition and recording. Had the budget been higher, it is highly likely that hiring a real orchestra, with passionate, human players, would have been considered. Alternatively, the composer may have opted to stick with sampled instruments, but would have had the resources to make the cello part sound indistinguishable from one played by a passionate, human player.
It’s a money problem. Sampling allows bland performances to be produced relatively cheaply. However, that’s not what the public wants. They want passionate, exceptional music. So far, there has been nothing invented that makes the production of beautifully played, emotive music, with detailed articulations, cheap to make. The reason is that, however you accomplish the goal, it’s going to take years of music education and diligent practice, along with a lot of expensive musical instruments or else very expensive sample libraries and high-end computing hardware. Technology is unlikely to provide a solution to this problem any time soon.
Meanwhile, as music consumers, we can either accept cheap, bland, unexciting theme tunes, or else we can demand the good stuff. Once we express a strong preference for well played music, which reaches us and moves us emotionally, the cost differential between using human players and sampled instruments begins to vanish. In this case, the human orchestra has a potential speed advantage. If the orchestra can capture a great take, without retakes, it’s much faster than placing all the articulations necessary in a DAW, by hand.
It is pretty clear that none of this was anticipated, when sampling machines were invented. Had it been, the advantages of the technology would have been less clear cut. Engineers would also have spent more time working on the capture of musical articulations and performance gestures in fast, convenient, subtle ways. History reveals that they didn’t.
In the mean time, passionate music costs time and money to make, however you choose to do it.