An Imperfect DAW
For those of you that aren’t aware, today, most music is made using a computer programme called a Digital Audio Workstation, or DAW. Tape is an increasingly rare medium and multi-track tape almost extinct. I worked on some of the very first DAWs, back in the late 1980s. What I find alarming is that some of the early design decisions, which owed their origin to limitations of the technology of the day (and not knowing any better), are still at the heart of these applications and it influences the quality of the music made.
The strangest hangover from those early days, to me (speaking as a musician), is the grid that virtually all DAWs quantise or adjust musical events to snap to and align with. The grid is evenly spaced, meaning every grid line is the same amount of time away from its neighbours. Musicians don’t keep the beat this way. Their timing varies and it varies deliberately. Part of the art of moving a song along or bringing it back is in how you vary the distance between beats.
In a typical DAW, there are ways to change the tempo in real time and to adjust the “swing” factor, which varies the distance between notes in a bar, but it’s a mechanical and hard way to do what comes naturally to a musician. Our brains evolved to find what is naturally beautiful and what is natural is actually imperfect. Often, in real musical performance, no two bars are the same, in how the beats fall. That being the case, why do DAWs impose a strictly rigid, perfectly geometrically-spaced grid by default? If you make music that adheres to that grid, it sounds ugly, dull and lifeless. Robotic. That might work in some aesthetics, but in general, it’s about as far away from making pleasing music as you can get and this is our default setting!
I have a friend, now retired, who worked on the border between musicology and psychology. A gifted concert pianist, he discovered that the beat placement and the dynamics of each note followed a structure that applied to each bar, but superimposed on this was a structure that applied to each phrase, each movement of the music and to the entire piece. Each structure overlaid the smaller one and modified it. His study was called Sentics. The structure roughly equated to the composer’s characteristic, signature pulse (essentially, how the composer kept time in his own head), blended with the “shape” of the emotion the composer wished to convey in the piece.
A performing musician will place the emphases and space the notes differently, when they are trying to make the music express anger, compared to when they are trying to express love or longing. They do this instinctually, but the differences are measurable. Composers almost cannot help but place notes and melodic intervals on the page the way they hear the time beating in their heads. What they write down reflects how it is heard in their musical imagination.
What this means is that the evenly spaced, rigid grid that all DAWs default to is an extremely poor choice. It destroys the linkage between what the composer hears in their head and how it is quantised in the DAW. It also denudes musical performance of any of its emotional content. Snapping musical events to an evenly spaced quantisation grid strips the beauty out of the composition and its performance.
One answer is to simply ignore the grid and to record a performance as it occurred. That works in some circumstances, but it’s not very flexible. If you wish to create musical parts in other ways, there is no grid to snap to at all. Also, if you want to vary the emotion, or rush or drag the beat in particular musical sections, for compositional impact, the only way you can do so is to re-record the performance, playing it that way (or take some judicious guesses with elastic time processing).
There are a couple of approaches that could exist in a DAW product, but which (mostly) don’t. The first idea is to simply analyse a real musical performance and create a grid from where the beats actually fell. In fact, the majority of DAWs do allow you to extract beat placement from audio and to create a beat map from that analysis, that you can choose as your grid. Capturing a few bars of a groove this way is something that can be accomplished fairly easily today, although it can be unwieldy as a workflow, to do so. It feels very manual and laborious.
Better, I think, is to have a library of different grids and the ability to superimpose variances according to the structure of the composition (at bar, phrase, movement and piece scales). If you want a section to be exciting, the grid could be modified according to how placing the beats and accents would have been played by musicians trying to make the section sound exciting. Adjusting both note placement and dynamics, whether recorded audio or as MIDI events, is now possible, thanks to elastic pitch and time signal processing.
The second way to get a grid is to have the composer record, through a haptic interface, how they hear the beat in their head (while reading their own score, for example). This is a pure grid recording and it can be the basis for the rest of the musical production to snap to.
To add a little more deliberate imperfection, modulating the note positions in time and their dynamics (and even pitch) with low amplitude brown noise (the same noise source used to simulate the sound of rolling thunder and render rocky coastlines, in CGI) can create subtle, small imperfections that might add to the beauty of the music.
Changing the structure of the piece, or the emotion of the music at some scale (again, bar, phrase, movement or the whole piece) could be done by simply recalculating the grid spacing and dynamics and applying the appropriate signal processing to make the recorded musical performances snap to the new, imperfect grid. A DAW product that had this facility would make the workflows needed to accomplish it very easy to do.
Another silly design decision that requires revision, in DAWs, is the insistence that all musical parts time-align precisely to the same grid. In the imperfect DAW, grids for individual parts could drift, allowing the insertion of hesitations, anticipations and part-by-part emotional excitations. Each instrumental part (or recorded track) can have a grid that relates to, but varies from, the master grid of the piece. That makes part-to-part time alignment an artistic choice, rather than a default imposition. Playing with time is a key dimension in creating pleasing music.
The benefits of the imperfect DAW are that it would tend to produce more humanistic music, regardless of the method of recording the musical parts (e.g. step-time note entry or electronic score, rather than recording live performance). It would also produce music that preserves its emotional impact, dynamics and the composer’s intentions. Imperfection controls also permit further degrees of creative freedom to the music producer.
Or you could record a real band and hope for the best, but that lacks a degree of post-production control and flexibility.
Creating an imperfect DAW would provide greater creative options, with rigid perfection being just one of the aesthetic choices available. With greater options, a larger aesthetic space could be covered and more pleasing music produced, with simpler, more straight-forward workflows (many DAW workflows fall below the standards expected of modern UX design). The added value would enable the DAW maker to protect their margins and to differentiate their product in what is becoming a commodity market. Being the first to offer a new approach to music making is always exciting and attractive to the music production community.
With a distinct move away from pure electronica and dance music, toward more organic, acoustic forms of music making, in recent times, being able to produce and manipulate the finish and polish of musical works, without destroying their organic feel and charm, would be highly desirable to many musicians. Technically, there are no significant impediments to realise this functionality in a DAW product, as the elastic audio processing techniques already exist. The risk is that musicians and music producers find the new workflows counter-intuitive or the musical results unattractive.
The biggest benefits of the proposed product extension are that musicians will be able to explore a larger gamut of creative options and music consumers may, as a result, encounter new musical works that couldn’t have been produced in any other way.
If you are interested in bringing this possible future product design into existence, please contact me. Execution is relatively straight forward. The creative possibilities are greatly extended. The first DAW to incorporate these features would enjoy considerable competitive advantage.
References
https://qz.com/1044781/this-music-production-tool-is-the-reason-why-all-new-music-sounds-the-same/
About the author
Michael Topic is a freelance Product Manager and musician, with over thirty years experience delivering products that didn’t exist before. He welcomes contract enquiries to define new, competitive products, design them and deliver them. His speciality is software-based products.
Disclaimer
The hypothetical design ideas discussed are intended to demonstrate possibilities in product design. They are not intended to imply anything whatsoever about past or present employers, nor to infringe on any intellectual property.
About the “Possible Future Designs” Series
This occasional series of articles examines the many ways in which product management discipline can be applied to a variety of markets, to propose and examine innovative, new product opportunities.
Organisations wishing to pursue any of the ideas discussed are encouraged to contact the author to discuss potential research and development collaboration.
