A short post for devs who are not musicians.

Mixing is making a song or sound effect sound good at any volume on any playback device. There's a bunch of technique involved, but what matters is this: just because a song sounds OK to you on your headphones doesn't mean it will sound good on someone else's cell phone or home entertainment system. What do I mean by "sound good?" Simply put, the audio image is blurry and not as impactful as it could be.

Imagine if the colors on your computer screen changed, drastically, depending on the shape of your room and the dirty clothes on your floor. That's acoustics. The playback device and the room are a system, meaning they cannot be meaningfully isolated from each other.

Also, the balance between sounds in the frequency spectrum change unevenly as volume increases. Broadly speaking, sounds towards the top and bottom change in volume faster than sounds in the middle. At loud volumes they are louder than the mids. At soft volumes they are softer. If a particular component of the audio image - an instrument or part of a sound effect - has more extremes than mids, it moves into the foreground of the audio image as volume increases. This distorts the composer/sound designer's intention.

The mixer works in the closest he can get to an acoustically neutral space and tests the mix on a wide variety of devices at all volumes to ensure the accurate transmission of the composer/sound designer's conception. Done well, the mix is effective in the widest circumstances. 

And now you know what your audio guy is doing when he's mixing.

Vox and I had a brief chat recently and the topic of complicated music came up. As everyone knows, Vox is not a trained musician and admits he can barely play an instrument, yet he was a top 40 Billboard recording artist in the early 90's. Contrast that to a formally educated musician who works at Starbucks. How is this possible?

We got on this topic because I mentioned a time that an asset I had written for a company was rejected because it was too complicated. Vox recounted his experience with musicians who wrote songs that only other musicians would enjoy, and I explained - glibly - that was why I quit jazz. In my opinion the hardest skill for a musician to learn is how to hear music as it sounds to non-musicians. Vox wanted more: "If someone is skilled, how can he not know that? I know how to imitate writers, why is this tough for musicians?"

I was able to give a partial answer: there is in music a similar dichotomy as exists in persuasion, namely, between music that is dialectically correct and music that is rhetorically effective. It's a different set of skills. There's what you learn eight hours a day in a practice room, and there's what you learn in front of an audience.

Only half satisfied with that answer, I want to revisit my original assertion: "the hardest skill for a musician to learn is to hear how music sounds to non-musicians." Why is this so?

One reason might be is that music is perceptual. Most people cannot focus their attention on each separate element. To make music at a high level of skill, however, you need to be able to hear each element in isolation. Once you can, however, it's easy to forget that a non-musician can't. The next thing you know you're creating music for people that can only hear it the way you do.

As your ears become more refined you must not be lose track of the big picture. Mixing helps in this area, by guiding the listener's attention from foreground element to foreground element.

In music the meaning of a sound is different depending on what surrounds it. This differs from language where the meaning of a word has finite definitions from context to context. In music a single sound has infinite potential meanings - except for when it doesn't. There is no such thing as a musical dictionary. A minor chord in the key of E mean somber, unless it means groovy, upbeat, or something else. To compound this our sense of how context functions changes depending on what we've been exposed to. Musicians, by expanding their knowledge in pursuit of their craft, are exposed to more possibilities, and can't always remember how Green Day used to feel before being exposed to Ligeti. It becomes difficult to express yourself honestly in a simpler idiom.

https://www.youtube.com/watch?v=tnn4Y9FbEaQ

Sometimes musicians have bad educations. My relationship with music theory is akin to Vox's with economic theory. After much questioning, I concluded that most of what they teach undergrads is worse than useless. The fundamental problem is that harmony is placed at the center of the theory curriculum and it's the least important element of music. No one has created an adequate theory of rhythm, which is the basis of harmony, so it's not formally taught in school.

Western functional harmony is fundamentally a cycle of three types of chords: tonic, predominant, dominant - and back to tonic. To identify where the cycle starts you need to analyze the large scale rhythm. Most harmony textbooks acknowledge this, but then say "but it's beyond the purview of this book to analyze rhythm." It follows that a proper undergrad music theory curriculum would start with a theory of rhythm, with the aim of teaching students how to formally recognize the beginning of a rhythmic cycle, and only then move on to harmony. Instead, the vast majority of music curricula do not offer even a single course on rhythm.

There's very little out there in terms of formal rhythmic analysis. What little you find usually presents itself as an introduction to a field that has only begun to be explored - and this has been going on for decades. So it's possible for a music student to finish school thinking he understands harmony because he's learned everything there is to know ahout Neapolitan 6th, false cadences, and set theory. But he's never studied rhythm beyond a professor offhandedly telling him "well, you just kind of have to feel it." So he becomes one of Nicholas Taleb's "Intellectual-Yet-Idiots." He can speak for hours on a topic without realizing the foundation is faulty.

Contrast that to a naturally curious autodidact making Hip-hop beats at home. Without a teacher to distract him with a foundationless theory of harmony, he gets a ten year head start studying rhythm, and when he's ready to use more interesting chords, they are anchored in a solid rhythmic foundation.

There is also the aspect of finding the right symbols to connect with your audience. You can't learn that in a practice room. You learn that on a stage or whenever non-musicians candidly respond to your music. There's a genius that knows what symbols resonate with culture in a particular moment in time. It has something to do with freshness, timing, context, and other intangibles. It's a skill more like public speaking than writing. It's has more to do with fashion than architecture. All of the musical elements, rhythm, timbre, space, harmony - all these do is add up to a gesture. A formal music education can teach you how these elements add up to a gesture - they can form a coherent statement - but to know how they whether they resonate with an audience, you have to interact with one.

Formal music education isn't necessary to make music that people will pay for. All you need for that is curiosity and an audience. Creative longevity, however, requires wide exposure to many styles, and formal education can help with that. Many one hit wonders make a splash, but then run out of gas because they can't do anything else. You'll find that the truly great musicians have spent a lot of time in front of audiences and have also investigated many styles of music beyond the ones they perform.

There is craft in the piano reduction: the counterpoint works, there is a gradual rising and falling line that delineates registral space, a balanced melody, a sense of structure coming from the time signature change in the second half, a shifting of which voice carries the melody, appropriate sonorities, and so on.





It's easy to teach pitch and rhythm, but if that's all you know, you've been robbed. Music needs to capture your imagination. Ask, would you buy this?





Of course not! It sounds just like any other piano reduction.

We need color. The basic idea is an almost regal fanfare beset by eerie harmonies. Here's an early draft of the composition that I submitted to my team.





It's not bad, but did I sell it? I would say no. It sounds too polite, too nice. I needed more light coming through the windows and a sense of dark power underneath. Let me add some sustained strings in the upper register to represent the light, and a dark synth in the bass to represent the foreboding power. While I'm at it, I'll make sure the staccato strings are crisper and more defined - more aristocratic, if you will. I'll even add in a chromatic, arabesque harp line in the second half to signify a sort of dark aristocracy. Here you go:





The feeling is there. Time for a reality check. Did I evoke a world? Did I really do it? Prince, master of intangibles, says naw.





Whose fantasy is more vivid? Whose frame is stronger? Who better evokes aristocracy?

Prince, of course. Always trust the greats.

• Get off social media. Turn off the phone. Exit blogs and news sites. (Just keep email open.)

• Don't listen to music over and over again. Listen once, think of something to change.

• If stuck, try something at random. Anything at all. Learn by engaging, not thinking.

• If I'm having a hard time evaluating a proposed change, it could be because I'm putting too much effort into playing it while also evaluating it. Enter it into the DAW and evaluate.

• Don't write everything on paper. Just write what's helpful to see on paper. Put the rest into the DAW. You can can see, compared to what you will hear, what I wrote on paper was pretty minimal.

• Keep in mind counterpoint shortcuts. In my case that means all moves all valid except similar / parallel octaves and parallel 5th's.

• Trust experience. For example: I realized I was writing a lot of a parallel 5th's early on into the process, but realized that since the melody is based on 5th's, that's OK. Actually that's where much of the characteristic sound comes from.

• Tried all my old tricks first. I built the form around putting the melody in different registers, used pedal points, and employed tried-and-true textures like staccato strings over legato French horns.

• Used template from previous project. That's about 50 instruments that I didn't have to search for. For instruments I don't use very often, this saves me a lot of time.

• Hopefully I can this template for the entire project. I may even compose all my tracks in the same project to reuse the same mix settings from one track to the next.

• Certain thematic passages may be repeated in other tracks.

• Simplified mixing process: grouped 28 tracks into 8 submixes, five sections. mixed start to finish. Didn't mix from scratch but began with the levels I had set while composing. It helps that I set my levels intelligently from the get-go.

• Choir parts alternating between big unisons and fully voiced chords.

• The chord substitution in the 9th measure. I simply replaced Bb with B natural and it's a great sound, whatever the proper name may be for the resulting sonority.