Subscribe to the weekly “audio edition” via iTunes
A few days ago, one of my kids came back from school with a decidedly underwhelming math quiz grade. So we sat down and tried to work backwards to figure out how this happened. And it didn’t take long to figure out that there were two things that led to this.

Problem number one, is that the kid wrote the formula down incorrectly in their notes. Problem number two is that because the practice problems they were assigned to do in the textbook did not have answers in the back of the book, they didn’t realize they were making the same mistake over and over – until they got their quiz back and saw all the red ink.

Needless to say, whether it’s learning a new math formula, directing choir rehearsal, running sectionals, rehearsing with your trio, or even practicing on your own, the more effectively and sooner you can pinpoint relevant errors, the more efficient your practice can be.

Which I know is super obvious, but…are we actually as good at detecting errors in our playing as we think we are? Or might we be missing more things than we realize?

As I was stewing about math textbooks, and why they only provide half of the answers to all of the practice problems (like all the evens or all the odds, but not both), I ended up stumbling across a whole set of studies on “error detection” in music, and thought it might be interesting to look at some of what researchers have found in this area…

Knowing what you want

You know when you show up for a first orchestra rehearsal feeling pretty good about yourself because you looked at your parts ahead of time, but then you discover in the most spectacular way possible, that the F natural in measure 72 was actually an F sharp?

Well, the first step in becoming a better error detector is knowing what “correct” sounds like – having some sort of internal standard of performance to compare things against.

What type of score study works best?

So in a 1996 study (Crowe), 30 undergraduate music students were given some time to do score study in one of a few different ways, and then asked to identify the pitch and rhythm errors contained in a selection of music.

And between score study alone, score study at the keyboard, and score study with an error-free recording, score study with a recording led to significantly better detection of pitch and rhythm errors than studying with the score alone.

An error detection hierarchy?

Of course, noticing note and rhythm mistakes is pretty basic, and it doesn’t take long before those things stick out like a sore thumb. What about error detection of more subtle things like tempo, phrasing, articulation, vibrato, tone, etc.?

Like, is there some sort of error-detection hierarchy? As in, do we have a tendency to notice certain categories of errors more readily than others?

Hmm…

Well…it’s not clear. One pair of researchers (Geringer & Madsen, 1989) suggested that we might be inclined to notice rhythm first, then pitch, tone quality, and loudness. But studies from 1989 (Doane) and 1992 (Hayslett) found that participants noticed pitch errors more easily than rhythm errors.

But then there are a couple studies from 1993 and 1997 (Byo) in which participants tended to be more accurate in detecting rhythm errors than pitch errors.

Wha…??????!!!!

It’s tricky

In all fairness, I think this is a really tough thing to study, because not all pitch and rhythm errors are equal. Their obviousness depends on a lot of different things. For instance, if I’m listening to a violinist play, and it’s something I know well, all things being equal, I’m probably going to notice pitch more than rhythm. But if I’m listening to a flutist or bassoonist or some instrument whose rep I don’t know very well, it’s very possible that rhythm might be the thing my ears go to first.

Which kind of gets at what might be the most interesting takeaway from these studies.

Specifically, that as the complexity of the music increases (i.e. the number of lines/instrumental parts, polyrhythms, etc.), the more difficult it becomes to identify errors (Byo, 1997; Crowe, 1996).

Umm…and how is that interesting, exactly? 

Well, I promise to come back to this in a second, with implications and actionable takeaways – but first, let’s take a quick look at a related question.

Namely, does the act of playing an instrument compromise our error-detection abilities?

Simulating choir rehearsal

Well, a team of researchers (Napoles, Babb, Bowers, Hankle, & Zrust, 2016) recruited 55 music ed students and assigned them to one of two groups.

One group played either the soprano or bass line on a piano while listening to recorded choral excerpts with rhythm and note errors (as if they were directing choir rehearsal), while the other group just listened to the excerpts.

When asked to identify the errors in the recordings, the playing group had a significantly harder time identifying the mistakes – some even thought it was a trick, as they couldn’t identify the mistakes even though they knew they were supposed to be there.

As one participant said, “it was much harder to hear the mistakes than I thought it would be; we should have more practice with this in our methods classes.”

Some caveats

Of course, if they knew the pieces better, or if they were more experienced, I could see how they would get better at error detection. But then again, the better one knows a piece, the more subtle the errors become, and hence more difficult to detect, so I could also see how the act of playing one’s instrument could potentially interfere with one’s ability to detect errors at any stage of learning.

So what’s the solution?

Takeaways

Recording, of course! =)

Which takes us back to the finding that as musical complexity increases, our ability to detect errors decreases.

Which suggests that even if you have a recording, and aren’t trying to catch every single error or imperfection in real-time, while you’re busy trying to play, it’s still quite likely that you’ll miss a number of things unless you make things easier for yourself.

And how would one do that exactly?

Well, music ed professor James Byo, makes a couple suggestions in the 1997 paper I briefly alluded to a moment ago.

Suggestion #1: Listen for just one thing

One way to amplify your error detection ability is to simply the task, and listen for just one type of error, in just one part. Like listening just for intonation in the soprano line only. And then perhaps listening for intonation and rhythm in the soprano line.

And gradually increasing the number of things you listen for, by either adding additional error types or instruments/parts to your radar screen.

This way, if you have a tendency to fixate on intonation, you can totally immerse yourself in a recording and make a note of all of the intonation issues. But then, when you listen back a second time, you can totally ignore intonation, and only listen for rhythm issues. Which might reveal that there were bigger rhythm issues lurking there than you realized!

Suggestion #2: Divide and conquer

It seems that the same sort of approach could apply to live listening too. Incidentally, this seems to be an argument for the existence of sectionals, where there is less complexity to sift through, making it easier to pinpoint and troubleshoot problems.

It also seems to explain why it can sometimes help to have individual members of your chamber music group stop playing and just listen to the rest of the group play. With more cognitive resources to devote to listening, without the distraction of playing, or your own expectations and intentions potentially distorting your perception, it might be easier to get a clearer, more accurate sense of what’s actually happening.

The tl;dr version

So, to sum up, to maximize your error detection ability…

1. Record

2. Listen back

3. Start simple, by either a) listening for just one category of errors, or b) listening for just one thing, in just one part.

4. Increase complexity gradually, either by listening for additional musical elements, or by analyzing multiple parts simultaneously.


References

Byo, J. L. (1993). The influence of textural and timbral factors on the ability of music majors to detect performance errors. Journal of Research in Music Education, 41(2), 156-167.

Byo, J. L. (1997). The effects of texture and number of parts on the ability of music majors to detect performance errors. Journal of Research in Music Education, 45(1), 51-66.

Crowe, D. R. (1996). Effects of Score Study Style on Beginning Conductors’ Error-Detection Abilities. Journal of Research in Music Education, 44(2), 160–171.

Doane, C. (1989). The development and testing of a program for the develop- ment of conductor aural error detection skills. CBDNA Journal, 6, 11-14.

Geringer, J., & Madsen, C. (1989). Pitch and tone quality discrimination and preference: Evidence for a hierarchical model of musical elements. Canadian Music Educator, 30, 29-38.

Hayslett, D. J. (1992). The effect of directed focus on the peripheral hearing of undergraduate instrumental music majors. Paper presented at the national convention of the Music Educators National Conference, New Orleans.

Napoles, J., Babb, S. L., Bowers, J., Hankle, S., & Zrust, A. (2016). The Effect of Piano Playing on Preservice Teachers’ Ability to Detect Errors in a Choral Score. Journal of Music Teacher Education, 26(2), 39–49.