nolan wrote a screenplay the way you build a chorus
i went to see the odyssey and came out thinking about an illusion nolan has been using since the prestige. a stack of octaves slides upward under a volume curve that never moves, so the tones rise while the sound as a whole goes nowhere, and your ear climbs anyway. the useful part for a mix is that the rise lives in the handover between layers, which means you can build for a full minute and land exactly where you started with all your headroom still in your pocket.
i went to see the odyssey the other day. big film, loud film, i sat through the credits like a nerd.
and i came out thinking about a sound trick, which is not what most people come out of that cinema thinking about.
nolan has a thing for one particular audio illusion. he described it in an interview in 2017, when dunkirk came out: “a corkscrew effect. it’s always going up and up and up but it never goes outside of its range.” and then the line about that script, which stuck with me for years. “i wrote the script according to that principle.” not the score. the script. he’d explored it with david julyan on the prestige before that.
before anyone hits reply: i’m not telling you the odyssey uses one. i don’t know whether it does. a musician called jake lizzio has a breakdown arguing that one of the cues is built on the rhythm version of it, and that’s his reading, not göransson’s. nobody who made the thing has said anything either way. i sat in the dark wondering, which is a different thing from knowing.
what it actually is
roger shepard published the pitch version in 1964. the construction is a stack of tones an octave apart, sliding upward, underneath a volume curve that never moves. tones fade out as they climb off the top, new ones fade in underneath, and the loudest part of the sound stays put.
so the tones do rise. the sound as a whole goes nowhere. what moves is which tone happens to be sitting in the loud part, and that’s the thing your ear decided to follow.
jean-claude risset did it to tempo in 1986, two pages of comment in the journal of the acoustical society of america. same idea, one axis over: a rhythm that keeps accelerating and never arrives.
the part you can steal
the rise is in the handover.
which is a mixing instruction if you want it to be. most of us build a chorus by making it louder, and if level is the only thing doing the work then the limiter gets a vote on how big that chorus is. you spend the headroom on drop one and by the last chorus there’s nothing left in the account.
the shepard way is to bring the next thing in underneath before the last thing leaves. a synth an octave below the one that’s fading, a hat doubling as the pad thins out. the total sits still, the section still climbs.
and here’s the bit i like. it’s a loop. skip ahead, come back, measure it… nothing ever went anywhere. thats the useful part, because it means you can build for a full minute, land exactly where you started, and still have the whole push in your pocket for when you actually need it.
one caveat, because i’d rather say it than not. the one published listening study i could find on what this does to people had seventy three participants, all of it self-reported, and their tone was going down rather than up. so when you read that a shepard tone “hijacks your brain”, there is almost nothing underneath that sentence.
the video builds every demo from scratch, and then tries to break the illusion. that last part went better than i expected.
and the python that generates all of it is free, no email required: kernaudio.io/scripts/shepard-tone. run it, change the numbers, break it yourself.
more field notes
Sep 8, 2026
·sound science
the dialogue you mixed is not the dialogue they hear
a chain of reasonable decisions ends at the sofa with the voice buried. the film is about the chain. this issue is about your end of it: how to mix dialogue so it survives a stereo fold-down, a flat panel and an ordinary pair of ears, with the numbers i use, including one my own films failed on.
Sep 1, 2026
·sound science
a row of spikes nobody trained in
generated music leaves a mark, and it has nothing to do with taste. it is a row of evenly spaced spikes, put there by an operation that is also sitting inside half the plugins you own. which gives you a rule you can use on any analyzer: evenly spaced peaks came from a process, musically spaced peaks came from the music. plus a second tell the paper does not cover, and a note on why neither one is worth trying to EQ.
Aug 25, 2026
·sound science
nobody publishes where a phone speaker stops
two films, a few days apart, and the same thing happened twice. i went looking for the frequency a phone speaker gives up at so i could mix against it, and it isn't published anywhere. then i sat down to check the tuning arithmetic and found the vallotti preset in almost every synth is named after the wrong person. go and check a number, and it's either not published or it's filed under someone else's name.