Justin Mariner and Peter Schubert, Schulich School of Music, McGill University
The philosophy of our new aural skills program at McGill is pretty well summed up in this New Yorker cartoon by Alex Gregory. We want our students to feel comfortable experimenting with musical materials. If this were cooking class, we would want them to taste the ingredients at every stage in the process, not simply follow a recipe. This means providing a “safe space” for improvisation with no risk of embarrassment.
We have just completed the first two semesters of the new curriculum at McGill’s Schulich School of Music. This program was spearheaded by Dean Sean Ferguson and Chair of the Academic Committee on Teaching and Programs Julie Cumming. The curriculum was designed by the authors, with assistance from Robert Hasegawa, Gabrielle Kern, and Meghan Goodchild. The following is a report on these semesters with a close focus on the types of improvisation we do, and the use of SmartMusic for students to practice improvisation, “play’n’sing,” and sight-singing at home. We will be introducing a similar approach this coming year in semesters three and four of aural skills, and in the two semesters of keyboard skills that students take concurrently with first-year aural skills.
Instead of using notation as a starting point, many of our class activities involve improvisation. Our use of improv is in line with recent writing about its benefits in teaching, including activation of multiple types of musical thought and engagement of students as creator-owners of course material (Schubert 2017, 175, 184). Exercises always involve a specific framework so that students must listen to themselves with some level of analytical awareness. A simple example from the beginning of the first semester is singing sequential transpositions of a given motive by ear, naming scale degrees or notes. Students also improvise endings of short melodies: in the first term these continuations involve singing back a two-bar basic idea presented by the teacher and adding a second original two-bar phrase. In the second semester, the students add a further repetition of the basic idea and then tack on a short modulation, either from the major to the dominant or from the minor to the relative major, as demonstrated in this video series on YouTube. Other improv activities involve two-part music, beginning with melody harmonization in parallel thirds and “horn fifths,” and advancing to improvisation over a ground bass in the second term. Students see a ground bass that has been “realized” with all possible triad notes placed above as dots. We call these “Lippius dots,” after the theorist who first proposed that finding a path through the possible chord tones was a way to compose against a bass line. This exercise is a fun way to practice a number of interrelated skills by ear: chord spelling, voice leading, and melodic intervals.
Improvisations are evaluated entirely on participation–students get full marks for “fooling around and trying things out” (quoted from the course outline); there are no right or wrong answers. When one student improvises in class, we try to keep everyone else involved as invested listeners: the other class members are often required to sing back whatever an individual has just improvised, using scale degree numbers and conducting to demonstrate accurate analysis of the tune (we call this “proto-dictation”). In proto-dictation, students first listen and memorize the tune, and then use singing and conducting as aids to analyzing it. We continue to include these important steps even when we progress to written dictation, with the writing itself added as a final step. Students must refrain from using scale degree numbers in their singing during written dictation, but other than that the class works together as a team during the memorizing and singing phases. In the first semester, proto-dictation completely replaces traditional dictation. In the second semester, we introduce written dictation, and students seem to find this transition easy to make. Their melodic analysis abilities are strong enough to tackle longer examples from repertoire (Handel minuets), and to hear the bass lines of these examples in their two-part context.
Our improvisation activities primarily take place in class. However, we have also devised ways for students to practice some of them independently, by using the SmartMusic application. SmartMusic is perhaps best known for its ability to assess how accurately a musician performs notated music, but it has many other possible uses that do not involve digital assessment. Michael Callahan has created a wide variety of SmartMusic-based exercises in his pioneering work teaching theory at the keyboard. His work has served as a model for our own design of various types of weekly assignments in our ear training courses.
SmartMusic displays music on a screen with a cursor advancing metronomically across the score. Students record as many takes as they need, and review them all before submitting their best work. When they submit, their recordings are automatically sent to a server where instructors can listen to them. In this application, SmartMusic is primarily a recording tool, but it also helps us supervise certain aspects of the performance: each assignment is set up to allow only submissions recorded at a prescribed tempo, and with the built-in metronome audible. Students may get comfortable with the task by practicing at a slower tempo, or even without recording at all, but by the end of the process they must have acquired enough proficiency to conjure up a melody on a fixed timeline. Making students practice with a metronome is one of our favorite features of SmartMusic, both for improvised and notated exercises.
Lippius-style ground bass improvisation, as described above, is an activity that lends itself well to practice in SmartMusic. Submissions provide nice examples of the diversity of approaches students can employ. Some students will think about the path they take and be aware of what chord factors they are singing, while others will wander about by ear. We are open to both the deliberative and instinctive type of musical thinking, and students tend to combine elements from each approach in varying degrees, depending on individual learning styles. As in the classroom version of this exercise, we reward all reasonable attempts to meet the assignment criteria. This recording of a student performing her homework is a good example of a more deliberative method. The assignment is to sing in second species (i.e., two notes to a chord). She has clearly planned her trajectory and basically sings the same thing twice. Since improv is meant to be a safe space for trying things out, there is no penalty for parallel fifths. In class the instructor might point them out or might just ignore them. This other student, by contrast, is obviously doing it more intuitively in her dotted version of third species that doesn’t always stick to chord tones. She arrives at a non-chord tone above the bass in m. 3 so she can continue the smooth melodic descent begun in m. 2. Sometimes you can hear her shrink back a little when she sings non-chord tones on downbeats. The second time through, having gotten into a very low register in m. 6, she naturally wants to go back up. She sings an ascending passing tone, but this unfortunately leads to a dissonance on bar 7; she hesitates, tries to adjust the pitch, but in order to keep up with the beat, simply goes on up to a chord tone. She ends with a cute flourish: an idiomatic appoggiatura.
While assessment in improvisation is based entirely on effort and attendance, play’n’sing is graded on achievement, both in class and when assigned as homework in SmartMusic. The advantages of play’n’sing have been extolled by us elsewhere (Schubert 2017, 181), so we will not go into it here. Starting in the second semester, students sight-read short two-voice examples in class. They may be required to sing the top voice while playing the bass line on keyboard, or vice versa. Students also prepare longer examples from repertoire as assignments in SmartMusic. In this excerpt you will hear a typical play’n’sing assignment from the second semester. The software is not currently capable of assessing two voices at once, so these assignments are evaluated by instructors listening to the submissions. We have also added a SmartMusic component to the summer version of our fourth semester course, even though revisions to that course will only be completed next winter. This example shows where our use of SmartMusic is headed.
Sight-singing is graded on accuracy and rhythmic continuity in every class. The ideal goal of any sight-singing program is for students to perform well on the first run-through, but obviously not all students can perform at this level. We know that aural skills abilities vary widely among students entering university. For some students who are having trouble, one remedy is repeated practice of the same material, but doing this type of practice in class is awkward and inefficient in several ways: it is repetitive, it is uncomfortable for students who need the most do-overs, and it is boring for students who could be moving more quickly. Engagement can easily suffer for all involved. SmartMusic is a boon here too, allowing us to move this problematic sort of work outside class time. Making the best use of in-class time and out-of-class time is the main tenet of the “basic flip” as described by Kris Shaffer. With its automatic assessment, SmartMusic also provides an additional level of assistance to students in these melody assignments.
Computer assessment is possible whenever the student is asked to perform a single-line melody exactly as it appears on screen, so we use this feature for our melody assignments. As Gretchen Foley explains (see this report on the 2013 FlipCamp Music Theory Conference), SmartMusic assessment offers several important benefits to students: it gives them instant feedback about the accuracy of their pitch and rhythm, and it can function as a regular weekly practice regimen that counts towards their final grade. This submission shows how a melody assignment works. As in our other types of assignments, the student records one or more takes with an audible metronome in a prescribed tempo. At the end of each take, the program provides an assessment score out of 100, and annotates the music to show where the student was correct or incorrect. Notes that are sung accurately appear in green, mistakes are shown in red, and omissions are shown in black. Any note in red contains additional information: standard pitch notation indicates where the performer was sharp or flat, and notes that were sung early or late appear to the left or right of where they should be. The annotated music gets sent to the instructor along with the recording.
SmartMusic doesn’t always get it right: notes that are not detected can lower a student’s grade, and sometimes even notes that sound right are marked in red and the student loses marks. These problems can arise because of voice quality, mic level setting, an air conditioner in the background, etc., and we felt it unfair to punish students for the program’s unreliability in “hearing” notes. One solution is for the teacher to listen and adjust the grade, as Gretchen Foley does. Our solution is to forgive a possible 20% to accommodate computer error, and establish a minimum grade of 80% to submit an assignment; for this the student would get 100%. In fact, during the past year we have seen many students get assessments over 90% because they are determined to defeat the machine.
It is important to stress that the software not only enforces practice, it also facilitates it. We have found that the computer assessment of musical accuracy is sophisticated enough to be quite helpful to students. Computer assessment does not replace an instructor or tutor who would provide methodology for improvement, but it does show students where they went wrong. This is a good starting point for further practice, since students don’t always accurately identify their own mistakes. We encourage our students to think of SmartMusic as “a robot that listens to them practice.” It is also fun for the teacher when recordings occasionally pick up comments made to a study partner or roommate at the beginning or end of a take (“I can’t believe I made that same mistake again, but I’ll send it in anyway, I got 85%”).
At first it may seem discouraging that SmartMusic is all about having students practice without us, but that’s exactly the point: listening to the recordings, being assured that they really are practicing at home, that they really are singing and playing with the metronome—we have found these positively thrilling experiences. Many apprehensions we had last fall were quickly dispelled. For one thing, our students required almost no tech support (the teachers were another story). Another worry we had was that many entering students, being unaccustomed to singing in public, would not perform the new elements of the curriculum comfortably. We believe that weekly SmartMusic assignments make it easier for students to adapt to singing in public (we had no students who refused to sight sing in class, although some said they did find it stressful). Students did encounter difficulties when we chose overly long melodic excerpts, since the software requires the performer to maintain intonation a cappella. Careful selection of excerpts is key to the successful use of SmartMusic.
The results of these prepared assignments are less hit-or-miss than what we would get on sight-singing tests in class, and so the learning process is much more pleasant for students and instructors alike. Computer assignments give students more opportunity to work things out on their own, leaving more class time for amusing and challenging activities.
In a questionnaire handed out at the end of the first term, we asked our students which activity was the most fun, which the most useful, and which they thought they were best at. Answers were, respectively: improvisation, sight singing, and SmartMusic.
This work is copyright ⓒ2017 Justin Mariner and Peter Schubert, and licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.