If you ask musicians what they value most about making music, most of them will say—in some form or another—flow. Flow is that wonderful sense of being lost in your work, when "work" becomes joy. Time disappears, and so do distraction, anxiety, and just about everything else, yielding to a pure unity of creator and creation.
So wouldn’t it be strange if many of today’s musical instruments were designed to prevent or destroy flow? According to a recently convened group of audio experts, that seems to be the case. The group issued a report stating that most electronic musical instruments are complicated, confusing, and just plain frustrating to use—and when it comes to supporting flow, they compare poorly to instruments that have been around for centuries.
I was the facilitator for that group, which formed at this year’s Project Bar-B-Q, an annual conference designed to influence music hardware and software over the next five years. The conference offers a unique combination of presentations, brainstorming, socializing, and intense pressure, and it has yielded a surprising number of audio innovations.
Our workgroup was formed to analyze problems with the interface design of electronic musical instruments and come up with a potential solution. (We defined electronic musical instruments broadly to include equipment that often functions alongside or even like an instrument, such as mixers and effects.) The workgroup members were not simply armchair pundits; each member either plays electronic musical instruments professionally, makes or markets them, or both. (See the member names below).
As we began, we weren’t thinking in terms of flow; we just knew it was common for owners of electronic instruments to complain about them, often bitterly. I kicked off the session by asking for a list of the problems we had encountered working with our gear. Both the volume and nature of the responses were telling: Not only did the unhappy stories come in a flood, each situation was instantly familiar to every member in the group and was widespread among other musicians as well.
Here’s an example from my own experience that is typical of the stories we heard. (If you take medication for high blood pressure, you may want to skip ahead.) I wanted to record a sketch in my home studio featuring a drum loop and overdubbed electric guitar. I started by launching loop-editing software, only to find that it believed it was launching for the first time and needed to see the original CD for verification. As my enthusiasm began what would turn out to be a long downward slide, I found the installation CD in my thick binder of such CDs and inserted it in my Mac. I was then required to re-enter my name, company, and (very long) serial number. I had to look up the serial number in my PDA. Finally the software finished launching, and I was able to continue with making and recording a loop—although the next time I launched the software, I had to go through the whole exasperating sequence again.
Next, I needed to change the configuration of my digital mixer so I could monitor the prerecorded track while I added the overdubs—one of the most common tasks in any studio. I won’t name the brand of my mixer, but other owners will recognize the following torturous procedure, which is from a cheat sheet I keep handy:
Select Monitor>Source. Choose 1/2. To choose each channel to be monitored, press Param View, then select Shift-FX/Aux Send. Press F1–F4 On/Off. Choose Pre-Fader for channels to be monitored but not sent to Master. Choose Post-Fader for channels to be both monitored and sent to Master. (Turn F1–F4 off to scroll left-right, then back on.) Use the Value Dial to scroll the value in the upper right to 1/2. Use V1–V4 to set the level of the send going to 1/2 (that is, monitor send level). Set faders to Off on channels not to be sent to Master.
At this point I was ready to get a guitar sound. But there was a problem. In the guitar signal chain was a high quality outboard preamp/A-D converter that had suddenly developed a horrible buzz in its left channel. I checked the digital sync, which was fine. I wondered if some big something somewhere had switched on and turned my household AC supply to garbage. About ten minutes of troubleshooting followed. It turned out that everything (not just the preamp/A-D) just needed to be turned off and then back on again to clear up one of those grands mals that most digital gear is susceptible to from time to time.
Guitar playing commenced at last. But every time I erased unsuccessful takes, I had to remember to click the Tracks window of my recording software before hitting the Mac’s Delete key, because if I happened to hit the Delete key while the Mix window was active, it would delete a mixer channel, not the take I had just done.
The Bar-B-Q group heard (too) many such stories. At a certain point we just had to stop and see if the complaints they contained could be categorized. They could. Here’s how we chunked them:
The interface is too complicated: It’s hard to learn and hard to remember, and the learning that must take place is intellectual rather than musical.
There is a disconnect between gesture and feedback: Electronic instruments often impose a delay or require the performance of multiple actions, before any satisfying result occurs. For example, you must boot up or button-press through multiple menu levels.
What feedback there is tends to be not physical but informational, in the form of words and numbers on a screen, for example. Therefore, the user’s intellect is engaged rather than his or her emotions. But music is an emotional language, and emotions are experienced in the body, not the mind.
The instrument is unreliable, exhibiting unpredictable behavior or suddenly crashing.
The instrument fails to capture the full range of the player’s expression. Nuances may be lost—as with the quantization involved in MIDI recording or audio sampling—or some forms of expression may be ignored altogether. For example, tapping on the body of an acoustic instrument will usually produce a large range of percussive sounds, whereas tapping on a keyboard when playing a digital simulation of that acoustic instrument will likely produce nothing.
The interface is not aesthetically pleasing as a design. It either doesn’t look good or feel good to work with (or both); it’s not loveable. Performance on the instrument may also be uninteresting as performance, as in the case of digital DJs hunched over laptops (looking, as one participant pointed out, like they’re busy at a desk job).
Although this was a group of confirmed technophiles, all agreed that despite the enormous power offered by electronic instruments, with some notable exceptions we couldn’t say we actually liked them as much as we do many acoustic or electric instruments. And it was then, as we talked about exactly why that was, that we noticed the meta-problem that contained all the individual problems: Each one acts against flow. In other words, many electronic instruments tend to sabotage the act of music-making.
So far we had relied on an anecdotal description of flow, but as we realized that the concept was central to our discussion, we decided that we needed to be able to describe it more formally. And it turns out that flow does have a formal definition, thanks to the research of a leading creativity theorist, psychologist Mihaly Csikszentmihalyi (pronounced “chick-sent-me-HIGH”), former chairman of the psychology department at the University of Chicago and currently at the Drucker School of Management at Claremont Graduate University. Csikszentmihalyi has spent decades studying the experiences of painters, musicians, rock climbers, basketball players, and others who report experiencing flow. According to him, these are the characteristics of flow:
You are completely immersed in what you are doing.
Concentration is very high.
You know what you have to do moment by moment.
You have very quick and precise feedback as to how well you are doing.
Your abilities are stretched but not overwhelmed by the challenges.
You begin to forget all the things that bother you in everyday life, forget the self as an entity separate from what is going on. You feel you are a part of something greater and you are just moving along with the logic of the activity.
(I distilled these points from Csikszentmihalyi’s interview in What Is Enlightenment magazine and his book The Evolving Self, especially pp. 178–179.)
The Bar-B-Q group took its list of electronic musical instrument design problems and matched it with Csikszentmihalyi’s characteristics of flow:
It almost looks deliberate, doesn’t it?
At this stage the Bar-B-Q group was, frankly, pretty steamed. But at the same time we were excited. If we could now see so clearly what was wrong, maybe by a similar method we could find out what could be done right. We moved on to the next stage of the challenge: proposing a solution. We started by looking for universal characteristics of instruments that we felt support flow instead of fighting it.
The examples we focused on included acoustic, electric, and electronic instruments. These included the violin, the Gibson Les Paul guitar, Neve consoles, the Roland TR-series analog drum machines, Apple’s GarageBand software, and Ableton Live. The mood changed as various group members relived experiences with favorite gear. As before, we categorized our observations. Having already identified negatives helped us to identify positives by looking for opposites. Here are what we found to be the characteristics of a well-designed instrument:
Immediate, meaningful, physical feedback. Pluck a string, push a button, turn a knob: In each case, something happens, and it is probably at least intriguing if not pleasing. Crucially, the emotions are engaged more than the intellect, because the body is engaged, not the mind. Furthermore, "noodling" is encouraged—the instrument supports free exploration through intuition, as opposed to "formulate a goal and analyze a path."
Natural, ergonomic control. The instrument’s physical interface fits the scale and natural motions of the human body.
Aesthetic pleasure in seeing, holding, and using the instrument—for the user and ideally for an observer as well. The instrument is lovable.
Capture of a wide range of input data, including allowance for unanticipated uses. One group member pointed out that piano students at the Moscow Conservatory are required to produce 64 discernible levels of dynamics. That highlighted the challenge in making a sampled piano that would do justice to that ability, not to mention to such extra gestures as body taps and string scrapes.
Technology should serve simplicity. Although the underlying technology may be complex, the user experience should be as simple as possible. This does not mean "dumbing down"; it means efficient design.
Progressive complexity. The instrument presents itself as simple, and it reveals greater complexity at a rate that matches the player’s development. (As Csikszentmihalyi says, "Personal skills are well suited to given challenges.") For example, a guitar is basically six strings, a fretboard, and a resonating body. It’s obvious how to get sound out of it, and you can get better and better sound out of it as you learn more about it. But you will probably never have to haul out a manual or visit a users’ forum to remind yourself how to, say, bend a note.
Activity-oriented design. This is perhaps the most important item on the list: The instrument is designed from the get-go to support a particular activity, as opposed to having a lot of functions packed into it that then have to be controlled through a complicated interface. The group noted that the physical form of a traditional instrument embodies its function, and that it’s easy to overlook the efficient power of physical embodiment. When electronic technology separates function from its physical incarnation, the instrument maker must be aware of the huge design risk being taken—not that the risk shouldn’t be taken, but its true scope needs to be fully understood.
One instrument we discussed helpfully served to illustrate both success and failure in supporting flow: the classic Yamaha DX7. The DX7 was a great success as a performance instrument with instant-on, compelling sounds, one-button patch changes, uncluttered panel, and fairly high expressiveness. But the moment the user went into Edit mode, “God help you,” as one design engineer put it.
Some may recognize similarities between the Bar-B-Q group’s instrument-design preferences and the principles of user interface (UI) design, as established by pioneers such as Donald Norman, Jakob Nielsen, and Bruce Tognazzini (all three now at the Nielsen Norman Group). For example, UI design principles state that an interface should be as simple as possible, that the function of controls should be obvious from their appearance, that the UI should be consistent—that is, a control should not do one thing at one moment and a different thing later—and that design should be user-centered.
Most group members had little or no training in computer UI design. But as with Csikszentmihalyi’s work on flow, it’s striking—and I believe instructive—that there’s such a close correspondence.
Indeed, the group felt that its list of positive design characteristics carried the weight of fundamental principles, so that their importance would be nearly self-evident to anyone with significant experience using musical equipment. And yet, comparing the list to well-known electronic instruments, we were shocked at how often and how severely these principles are violated, in each case interfering with flow. Think about instruments you know that give delayed, confusing, or unsatisfying feedback, that have controls that are physically awkward to use, that just don’t look or feel good, that ignore important nuances of performance, that are too difficult to figure out and so force you to think all the time.
Such instruments appear to have been designed from the engineering-out rather than the activity-in perspective. From the user’s point of view, such priorities are upside down. If what musicians want most is to flow, surely instrument makers should consider the possibility that the flow experience is in fact what they are selling. Here’s another way of saying that: From the user’s point of view, the interface is not part of the product, it is the product.
One of the group members, Keyfax founder (and former Yes keyboardist) Julian Colbeck, also gave one of the conference’s feature presentations. In it he told a story that should probably be heard by every electronic musical instrument maker.
Among its other activities, Keyfax produces websites and video users’ guides. Recently the company was commissioned to create both a site and a guide for a new workstation keyboard (one of those instruments with the power to do pretty much everything involved in making music—most of which probably goes untapped by the average user). The workstation’s manufacturer anticipated initial sales in the low thousands. But Keyfax’s DVD-based guide and dedicated website turned out to make the instrument’s dense interface much easier to learn. These resources generated good word of mouth...and the workstation ended up selling almost ten times what had originally been expected. The manufacturer credited Keyfax’s support media with generating a large part of that windfall.
Imagine if the enhanced usability represented by these tools had been designed into the workstation from the beginning, as opposed to being retrofitted.
Even better, imagine if the maker of that workstation had been able to design it for flow.
We believe the way forward is through industry-wide adoption of flow-centered design principles. The next step is lobbying companies to study and endorse these principles or an improved version of them. No restrictive new standard need be imposed; each company would be free to interpret and apply the principles as it sees fit. But shared awareness of the goal should be a big help in getting there. And it will help customers make better decisions about what they want to spend their money on: deep thickets of features, most of which they may never actually find, or…
Jack Buser, Manager, Game Developer Relations, Dolby Labs
Kris Carter, President, Gibson Audio
Julian Colbeck, Founder, Keyfax New Media
Spencer Critchley, Composer/Producer
Dr. Patrick Gleeson, Composer
Alain Georges, CTO, MadWaves
Alistair Hirst, CEO, Omni Interactive Audio
Robert Hooper, Game Systems Programmer, Atomic Games
Geoff Kirk, Senior Sound Designer, Humongous Entertainment
Len Layton, Senior Vice President, North America, C-Media Electronics
Rob Rampley, Software Architect, Line 6
Scott Snyder, Product Manager, Audio Tools, Electronic Arts Canada
Guy Whitmore, Audio Lead, Microsoft Game Studios
Copyright © 2009 O'Reilly Media, Inc.