Soapbox: Learning Tracks Part 2

Okay, so whilst I don’t really approve of using recordings to learn your music, I do live in the real world. I recognise that the world isn’t going to change its habits just because I have an opinion. So, today I am going send out a plea that if you insist on using learning tracks, you give some thought as to the type and quality you are going to use, and their likely effects.

First, can you check they’re accurate please? Don’t assume that just because you paid for them they’re going to be right. Do this before you send them out to your singers, and get them remade if they’re not accurate. If you leave it for some of your singers to notice any errors, that means that others will already have learned it wrong and you will waste a shocking amount of rehearsal time trying to fix those errors.

Second, can you think about the impression of the music the tracks will leave with the singers? There has been over the years a notion that the recordings are there so that the singers can ‘learn the notes and words’, but not the ‘interpretation’ – which is to be added in rehearsal. That is, they have been treated as a kind of automated ‘note-bashing’ system.

Now, the problem with asking people to learn musical content divorced from expression is that the qualities a recording has that marks it as leaving out the musical interpretation are precisely the qualities you don’t want to hear in performance. If you give out learning materials that sound like they’re sung by a robot, you are inviting your singers to learn to sing the music as if they too were robots. You only get one chance to make a first impression, so please make it a musical one.

I’m not sure which is worst, in this context, the learning materials that are actually computer-generated or those sung by human beings trying to sound ‘neutral’. The former have the virtue that they can at least render complex rhythms accurately, but I think they have also produced some of the most mechanical performances I have encountered. A human voice seems to elicit a more human response, even when it is trying to hide its humanity.

In some senses, the midi file output from notation software that renders the music as a essentially a series of beeps is preferable to the software that attempts to articulate the words, as it sends the singers back to the music. It becomes a learning aid rather than something to follow parrot-fashion, and the need to match up what is heard with reading the lyrics actually engages the singer in a more cognitively active learning process.

But if you want your singers to use a learning process that is essentially imitative, it follows that the more musical the materials you give them to imitate, the more musical the ultimate performance is going to be. I would suggest that it is better to have learning tracks that give a positive commitment to a musical vision, even if that vision is different from yours, than to start off with a bland and mechanical approach. If nothing else, the process of changing something good to something else good is going to be pleasanter than trying to add spice to a carefully-studied dull performance.