Noise, Noise, Noise

Noise is a the foe of any information transmission, yet the total elimination of noise is–counter-intuitively–bad.


Nature is, from our human perspective, inherently noisy, and this noise is inherent to its beauty. When we try to over-fit a simple order we get something that is, yes, lined up neatly, but also dead.

Nature is not truly noisy, but the mathematics that underlies its order is more complicated than we can easily model (it is only in the past century that we’ve developed the tools necessary to model tomorrow’s weather with any degree of accuracy, for example.) Whether we are drawing trees or clouds, controlled randomness works far better than regular repetition. (If you want to get technical, the math involved tends to involve fractals.)

In photography, dithering is the intentional application of noise to randomize quantitization errors. According to Wikipedia:

Dither is routinely used in processing of both digital audio and video data, and is often one of the last stages of mastering audio to a CD.

A common use of dither is converting a greyscale image to black and white, such that the density of black dots in the new image approximates the average grey level in the original.

…[O]ne of the earliest [applications] of dither came in World War II. Airplane bombers used mechanical computers to perform navigation and bomb trajectory calculations. Curiously, these computers (boxes filled with hundreds of gears and cogs) performed more accurately when flying on board the aircraft, and less well on ground. Engineers realized that the vibration from the aircraft reduced the error from sticky moving parts. Instead of moving in short jerks, they moved more continuously. Small vibrating motors were built into the computers, and their vibration was called dither from the Middle English verb “didderen,” meaning “to tremble.” Today, when you tap a mechanical meter to increase its accuracy, you are applying dither, and modern dictionaries define dither as a highly nervous, confused, or agitated state. In minute quantities, dither successfully makes a digitization system a little more analog in the good sense of the word.

— Ken Pohlmann, Principles of Digital Audio[1]

The term dither was published in books on analog computation and hydraulically controlled guns shortly after World War II.[2][3] 

This mechanical dithering is also why many people play “white noise” sounds while studying or falling asleep (my eldest used to fall asleep to the sound of the shower). As the article notes:

Quantization yields error. If that error is correlated to the signal, the result is potentially cyclical or predictable. In some fields, especially where the receptor is sensitive to such artifacts, cyclical errors yield undesirable artifacts. In these fields introducing dither converts the error to random noise. The field of audio is a primary example of this. The human ear functions much like a Fourier transform, wherein it hears individual frequencies.[8][9] The ear is therefore very sensitive to distortion, or additional frequency content, but far less sensitive to additional random noise at all frequencies such as found in a dithered signal.[10]

In digital audio, like CDs, dithering reduces the distortion caused by data compression (a necessary part of the process of producing CDs). The same process works in digital photography, as demonstrated in these photos:

from Wikipedia

In art, dithering allows artists to create a wide range of colors out of a limited palette (this is, of course, how the colors in newspaper comics are created).

There are many different ways to dither, and subsequently many different dithering algorithms, all created with the intention of using random noise to increase the quality of images/sound/readings, etc.

Too much noise is of course a problem, but as photographer Frederik Mork notes:

Don’t forget that noise can also be a creative tool. Especially when paired with black-and-white, high ISO noise can sometimes add a lot of atmosphere to the image. Some images need noise.

Symmetrical faces are supposed to be more attractive than less-symmetrical faces, but this only applies up to a point. Even very beautiful faces are not truly symmetrical, and increasing their symmetricality (by mirroring the left side over the right, or the right side over the left) does not improve them:

Nicole Kidman

I don’t know where this image came from originally, but I found it in a YouTube video.  There is a literature on this subject, probably primarily related to plastic surgery. The image on the left is a Nicole as she normally appears; the image in the center is Nicole with the right side of her face mirrored, an the image on the right is the left side of her face. Note that the original contains several asymmetries: her hair, her eyebrows, and even the way her necklace lies on her chest.

This post was inspired by a conversation about what it meant to be a “good” musician, and whether one can be a reasonable judge of music outside of one’s own musical tastes. If I love rap but hate techno, can I still recognize which techno songs are considered “good” by some standard other than “techno fans like it”? Is all art subjective, or are some pieces actually better or worse than others?

I am a simple creature, and I do not really understand the depths of what goes into creating music. I can squeak out a few notes on a recorder and Twinkle Twinkle Little Star on the piano, but deeper theories behind things like “harmonic intervals” and “chords” are beyond me. Music to me is an immersive sensory experience (more so since quarantine has induced a state of semi-zen that has quieted the normally very loud meta and meta-meta narrative in my head). Whether something is good or bad I cannot say in any quantitative sense, but to paraphrase the words of Justice Stewart, I know it when I feel it.

I had very little exposure to popular music as a kid (I listened to a lot of “Christian rock”), and only began listening to music seriously and sorting out what I liked and didn’t in grad school, so whatever arguments you have about people forming their musical taste in highschool don’t apply.

I tend to like music that has a lot of distortion–noise, if you will. (I like music that reflect what I feel, and my feelings look like a Jackson Pollock.)

You might have noticed that my current favorite musician is Gary Numan (the guy who sang “Cars” back in 1979.) I don’t like most of Numan’s early work (like “Cars”); the sound is too simple. It took a couple of decades for Numan to develop the layers of complexity and necessary to create the sort of musical soundscape I enjoy. Compare, for example, his 1979 performance of “Are Friends Electric” to his 2013 performance of the same song. They are the same song, but with slightly different arrangements; I don’t know the technical words to describe them, but I find the 1979 version merely acceptable, while the 2013 hit me like a brick. (That’s a good thing, in this context.)

Or, heck, let’s compare Cars 1979, with Cars 2009 (with NIN). Okay, so the first thing that stands out to me is that 2009 Gary Numan has gotten laid and is no longer afraid to move around the stage, unlike 1979 Numan. Second, Trent Reznor on the tambourine is hilarious. Third, there is WAY more noise in the second recording–especially since it is live and there is a screaming audience–and this does not detract from the experience: I vastly prefer the 2009 version.

One of the things I love about Numan’s music is that you can do this; you can listen to the same piece from different eras and see how his style has changed and evolved.

To really appreciate his new new style, though, I think it’s best to listen to new compositions, rather than covers of his older work, like I am Dust, Ghost Nation, or Crazier (with Rico).

Music, as I understand it, is built from a disturbed mathematical progression of sounds. A sequence of sounds in a regular pattern builds up our expectation of what will come next, and the violation of this sequence creates surprise, which–when done properly–our brains enjoy. If the pattern simply repeated over and over, it would become boring.

I recently enjoyed a documentary on Netflix about ZZ Top, (who knows, maybe they’ll become my favorite band someday). At one point in the documentary the band described the difficulties of their first recording session. They had their song, had their band, had their instruments, but the guy doing the recording just couldn’t capture the right sound. He had microphones all around the recording studio, but just couldn’t get what he wanted. So he proposed that the band loosen the strings on their guitars (or maybe it was just one guitar, forgive me, it’s been a while) to create a slightly out of tune sound. The band’s manager was having nothing of it: the instruments needed to be in tune. Finally the recording studio guy proposed that the manager run out to get them some barbecue, because it was getting on toward dinner time, and conveniently directed the manager to a restaurant across the county line, a good half hour away. With the manager gone for the next hour and fifteen minutes, they untuned the guitars and finally got the sound they were looking for.

(Here’s ZZ Top’s Sharp Dressed Man.)

Is this a “better” sound? It’s a different sound. It’s not the standard sound, and if you’re looking for the standard sound that guitars are supposed to make, this isn’t it. Is it wrong, on technical grounds? Or is it right because it was the sound the band wanted to make?

This whole post was inspired by the claim that Kurt Cobain was, technically, not a good musician. I found this accusation absolutely flabbergasting. Of course, I regard pretty much anyone who can crank out a tune on a guitar as “good”, and anyone who can top the charts as “excellent” by default. If we want to differentiate between top stars, well, I guess we can, but in the immortal words of @dog_rates, they’re all good dogs, Brent.

I’m not a huge Nirvana fan–like I said, I never listened to them as a kid (I didn’t have MTV and couldn’t have told you the difference between Pearl Jam and Oyster Jelly), but I like grunge and Nirvana is part of that.

After some conversation, I realized that my interlocutor and I were using different definitions of “technically,” which is always a sign that you should stop arguing about dumb stuff on the internet and go take a walk, except when you’re quarantined. I meant “technically” as in “actually,” while he meant it in the sense of “the proper way of doing something; by the manual.” His argument was that Kurt Cobain did not play the guitar properly according to the manual for how to play a guitar. Kurt was dropping notes, or not pushing down the strings properly, or otherwise playing lazily and not doing it right. Someone who has taken lessons on “how to play a guitar” will not be able to play like Kurt because he was being sloppy and not playing properly. They will have to learn how to be sloppy.

Naturally, I found this argument baffling. I have no idea how to play the guitar, but my standard for whether a piece of music is good or not is based on how it sounds, not how it is produced. I am listening to Nirvana now and I don’t hear–to my ears–any flaws.

To say that there is a “proper” way to play a guitar that is a standard benchmark against which musicians are judged sounds like some sort of  prescriptivist nonsense. It’s like saying that there is a proper way to paint, and that “impressionism” isn’t it:

Monet’s Impression, Sunrise

An outraged critic, Louis Leroy, coined the label “Impressionist.” He looked at Monet’s Impression Sunrise, the artist’s sensory response to a harbor at dawn, painted with sketchy brushstrokes. “Impression!” the journalist snorted. “Wallpaper in its embryonic state is more finished!” Within a year, the name Impressionism was an accepted term in the art world.

If the name was accepted, the art itself was not. “Try to make Monsieur Pissarro understand that trees are not violet; that the sky is not the color of fresh butter…and that no sensible human being could countenance such aberrations…try to explain to Monsieur Renoir that a woman’s torso is not a mass of decomposing flesh with those purplish-green stains,” wrote art critic Albert Wolff after the second Impressionist exhibition.

Although some people appreciated the new paintings, many did not. The critics and the public agreed the Impressionists couldn’t draw and their colors were considered vulgar. Their compositions were strange. Their short, slapdash brushstrokes made their paintings practically illegible. Why didn’t these artists take the time to finish their canvases, viewers wondered?

Indeed, Impressionism broke every rule of the French Academy of Fine Arts, the conservative school that had dominated art training and taste since 1648.

The “proper” way to play a guitar is however sounds good, and if it sounds better with dropped notes and imperfectly depressed strings, then that is the proper way to play. This is grunge, and grunge is intentionally high on the distortion.

Whether it sounds good or not is, of course, a matter of opinion, but Smells Like Teen Spirit has over a billion views on YouTube, so I think it’s fair to say that there are a lot of people out there who think this is a very good song. From their perspective, Kurt Cobain is a very talented musician.

There’s a saying that you have to learn the rules in order to know when to break the rules. It applies primarily in art and literature, but I suppose it applies to the rest of life, too. When you are learning to write, you learn the rules of grammar and punctuation. If you become a poet, you know when and how to throw all of that out the window. If you want to be an artist, you need to know how to paint; later you can throw together whatever colors you want. If you want to play music, then you need to learn to play properly–harmonic intervals and all that, I suppose–but when you actually play music, you need to know when to break the rules and how subtle differences in the way the notes are played result in massive differences in the music. Compare, for example, NIN’s Hurt to Johnny Cash’s.

This is a song that gets its power from our subverted expectations; we expect an increase in tempo that never really comes, creating a tension that stretches out across the song, finally breaking at 4:33 (in Trent’s version).

In these two songs, I think Trent’s version is “better” in the technical skills sense, but Cash’s version is better in the absolute punch in the guts sense. This is simply because of who Trent and Cash are; they each bring their own sense of self to the song: Trent the sense of a bitter youth; Cash the sense of an old man composing his epitaph.

Let us end with some Alice in Chains:

I hope you have enjoyed the songs.

6 thoughts on “Noise, Noise, Noise

  1. The thing with aesthetic preferences is that they can be graphed and when graphed there is a rough pattern that emerges.

    Everyone can generally agree on what is good. But not what is best.

    Liked by 1 person

  2. Have you read N N Taleb’s Incerto series?

    A quote.

    >It is remarkable how Maxwell’s neat mathematical derivations and the dangers of tight control can be generalized across domains and help debunk pseudo-stabilization and hidden long-term fragility.1 In the markets, fixing prices, or, equivalently, eliminating speculators, the so-called “noise traders”—and the moderate volatility that they bring—provide an illusion of stability, with periods of calm punctuated with large jumps. Because players are unused to volatility, the slightest price variation will then be attributed to insider information, or to changes in the state of the system, and will cause panics. When a currency never varies, a slight, very slight move makes people believe that the world is ending. Injecting some confusion stabilizes the system.

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s