(video of Michael McDonald and Jeff Porcaro on Soul Train lip synching "I Keep Forgetting")
Man vs. Machine: Ever Closer
Musicians have argued for years whether accurate meter as played by a drum machine or with a click track could swing. I've always felt that a good machine-generated rhythm was pretty cool. I've programmed some rather funky stuff on a drum machine -- stuff that actually cooked. And what about sampled beats? They seem to be doing a good job in a lot of situations. Well, somebody finally decided to look into it, and it turns out people do prefer rhythms that breathe … human rhythms.
Here’s the story.
Researchers played music for groups of subjects and then asked them how much they ‘liked’ the rhythm. In one group, the drum part was played by a drummer (Jeff Porcaro unbeknowst to the participants): in another test group, the same rhythm was played by a drum machine. Although subjects could not quantify their answers, they somewhat preferred the 'real' drummer.
A lot of people will say, "Of course, what did you expect?" Thing is, the subjects liked the drum machine just fine, and their preference for ‘real’ drums was not that dramatic.
Another aspect of the study was to allow the music to 'breathe'. The researchers programmed their drum machine to throw in subtle time variations. The subjects weren't swayed much by this wrinkle, still preferring the live stuff. But they also preferred the breathing drum machine to the non-breathing variety.
So can we close the book on the man versus machine argument? Definitely not. There are music styles that can only be played by real musicians, and there are styles that really are best handled by a drum machine.
Now I will always opt for a real drummer, but I see no problem with a mechanized substitute if it makes sense. I'm pretty sure I'd not want to see a jazz band with a drum machine, but in other situations a machine might be just the thing. I guess as long as we do right by the music and the audience, it doesn't matter what the solution is, just as long as it works.
For a summary of the experiment, visit: