In Praise of Hi Hat Gods (Kinda)
An intro to the concept of smartdumb. Topics include: hi hat gods, DaBaby convertible memes, "SUGE", reddit drum packs, round-robin sampling, Lowrey Berkshire Deluxe organs, ratcheting, etc.
One entirely reasonable question I get is: What is smartdumb?
What does it mean for music to begin with a question mark and end with an exclamation point? What's the difference between smartdumb and just dumb, aka dumbdumb? Is there such a thing as dumbsmart? Smartsmart? And so on.
For me, smartdumb is a feeling in two parts.
First, confusion: What the fuck is this?
Second, irreducible joy: Oh, wow.
It's the feeling I got when I first heard "Get Ur Freak On" by Missy Elliott at a "brain break" my freshman year of college. It's the same feeling I got when I heard "Galang" by M.I.A., or "Whip It" by Devo, or "Bang a Gong" by T-Rex, or "Sex Machine" by James Brown, or "Shari Vari" by A Number of Names.
These songs shortcircuit my brain and vibrate something deep and childlike inside me. The way my newborn son laughs with abandon when I make the same flatulent noise over and over again – whatever this feeling is, that's smartdumb.
There's something aggressively simple about smartdumb music. Most modern commercial music is simple and repetitive, but smartdumb seems to go one step further, bordering on the downright ignorant.
There’s a crucial sleight of hand though. Our brains can only handle so much repetition, and only certain kinds of repetition. We love locking into a dependable tempo. We like stuff that loops. But hearing *exactly* the same sound or motif repeat itself goes from boring to bothersome very quickly. A great locked groove loop functions in miniature like Eno's definition of ambient: constantly changing while also staying the same.
The artists and producers I admire most are the ones who seem to spend all their energy making you feel like they're not doing much at all. Their music may scan as shallow. But all this is in service of a very distilled, very childlike sense of abandon they want us to feel. Anything that would make you think they are remotely “good” at music gets in the way of that feeling.
A few years ago, Genius – f/k/a Rap Genius, f/k/a Rap Exegesis, the annotated song lyrics website built in the afterhours by a D.E. Shaw employee who didn’t know much about music but did know that a healthy percentage point or two of search traffic was for lyrics – launched a new video series called "Deconstructed". It's a behind-the-beat style video, where producers of well-known records walk us through their project files and show how the beat came together piece by piece.
In 2019, there was a Deconstructed for the song "SUGE" by the North Carolina rapper DaBaby. "SUGE" is a quintessential smartdumb record. I love everything about it: the lyric sheet ("She like how I smell - cologne"), the absurdly jolly 808 sound, the budget Jaws piano melody, even the producer tags. The video is one of the best music videos of the last 20 years. There's not much to the beat, and yet everything that is here works supports DaBaby’s vocal performance so deliberately. I like to imagine the splats of 808 as little cartoon thought bubbles that appear besides DaBaby's head.
The childlike joy of DaBaby's music was underscored by the contemporaneous comeup of Atlanta's Lil Baby, whose music was very serious, sentimental, and traditionally "lyrical" in contrast. (Serious, sentimental, and traditionally “lyrical” have never really been my thing.) I mourned DaBaby's recent homophobic remarks not just for the remarks themselves, but for the fact that it made him, and his approach to music, that much more dismissible.
So let’s talk about hi hats.
The "SUGE" Deconstructed video is one the biggest inside jokes among producers, up there with Unison MIDI Packs and CLA Vocals. At one moment, the record’s primary producer jetsonmade is talking about the original hi hat pattern for the record. The videographer zooms in on his FL piano roll – just straight eighth notes. By jetson's account, he passed the hi hat work over to co-producer Pooh Beatz, who "juiced them up." We hear and see the before and after: All Pooh has done, it seems, is change the pitch on a few hi hats. Jetsonmade felt differently though. "Pooh," he says, "he like a hi hat god, kinda."
There are multiple online merch shops that sell "He like a hi hat god, kinda" t-shirts now. There are Youtube reaction videos, like fellow Philadelphia producer BNYX's, that have nearly 200,000 views. The comments sections are filled with the kind of knowing exasperation and dismissiveness we’ve come to expect from people with no real skin in the game. But still the question remained: Does changing the pitch on two hi-hats really make you a hi-hat god? Kinda?
If you make commercial, rhythmic music, you're thinking a lot about hi hats. Hi hats are often the stripper pole of your beat – a sturdy metronome that the rest of the music slithers around. The instrument falls into the background, more felt than heard, and the relative tightness of the playing affects the way a track seems to bounce. Contrast the metrical looseness of Parliament or Rick James's drummers with the extreme tightness of something like Prince's LM-1 drum machine – two very different kinds of funky.
Modern computer-based drum production is almost always sample-based: A sequence of instructions that tell the computer to play back certain audio files at certain times. The benefits are: Really tight, consistent timekeeping! And the shorter transient times and lack of room and mechanical noises also means the drums can be that much louder without crowding out other instruments in the mix. It's why rap drums will almost always hit harder than rock drums, and it’s why you’ll rarely hear a commercial rock record now whose kicks, snares, and tom sounds aren’t bolstered with samples.
But as I mentioned, there's a certain point with repeated sounds where our brains go from bored to actively bothered. The "machine gun effect" is how people describe oppressively repetitive drum sounds, when you hear the *exact* same sound triggered in rapid succession with no variation. Hi-hats are most suspectible to machine-gunning, because they are the sound that's usually played most often and most successively. So this is the basic challenge for any producer in any genre to overcome: If I'm not actually playing a real hi-hat, how can I get my hi-hats to stay steady but not sound repetitive? How do I grease the pole?
The TR-808 had an inadvertent way around the machine-gun effect: the hi-hat circuit was synthesis, not sample playback. It used six tuned square-wave oscillators and a filter bank to create the hi-hat sound, so every hit was just different enough, and the natural limiting of the stereo output meant that the hi-hats changed in brightness when sounding alongside other sounds.
The TR-909 used a 6-bit sample of an actual cymbal recording, but also introduced three different variations of hi-hat articulations: closed, open, and choked (when open and closed were programmed to sound at the same time). The sample playback was also susceptible to variation because of all the particularities of the filter bank and frequency-dependent function generators that carry the sample into its final sound.
The LM-1 solved the hi-hat problem even more elegantly: It was sample-based, but the sample was constantly looping, so the start time of the sample changed every time the hi-hat was triggered. The LM-1 also had unusual quirks – the snare sample was 10ms late, which lent the machine its characteristic groove.
The sample-based Akai MPC eventually allowed for something called "round-robin sampling" – roughly, hitting the same finger pad would cycle through a grip of similar-sounding but slightly different samples with each successive hit (which in the case of hi-hats, replicates the acoustic experience kinda). New drum machines have parameter locking for more variation. And of course, running any drum machine through outboard amps, compressors, flangers, phasers, modulating delays, etc will usually help to create enough microvariation to keep the listener’s ear from being distracted by the repetition.
Weirdly: Digital audio workstation software, despite people knowing that this stuff has been an issue for decades, made machine-gunning a problem again. If you're using a computer for drum programming, your default sampler engine setup does not have these non-linearities built-in, and you have to go out of your way to set it up to do round-robin sampling or simple per-hit randomness or be really thoughtful with layering to create microvariations. The default FL Studio pattern window you see producers on Deconstructed assembling beats in is usually just a set of single-file sampler playback engines that play the file back perfectly, every time, always from the beginning. The NRA lobby is stronger than you think!
How a producer addresses the hi-hat machine-gun effect problem is, in my opinion, as much a part of their sound as anything else. The early days of trap production solved this with pitched hi-hats, exaggeratedly pitched snare rolls, and of course, ratcheting: the triggering of five or six or more hi-hat samples in the space usually used for one. In other words, a lot of producers solved the machine-gun effect problem by just leaning into it and making the hi-hat a kind of lead rhythmic instrument.
(Quick detour on ratcheting: It didn’t start with trap hi hats. You can hear 'ratcheting' in the sequences of Tangerine Dream and perhaps most iconically in the opening synth arpeggios of The Who's "Baba O'Riley' – via a ratcheting feature of Lowrey Berkshire Deluxe organ called "Marimba Repeat." It's only poetic coincidence that ratcheting hi-hats became one of the main production features of "ratchet music", which most people believe came as a mutation of the Louisiana slang "wretched." And it's even more bizarrely poetic that wood ratchets were used as police rattles in the 18th and 19th centuries – literally the sound of the police.)
Ratcheting was a very heavyhanded way of solving the hi-hat problem, that soon gave way to the kind of cool and understated hi-hat work you started hearing around the time of Metroboomin and 21 Savage's "Savage Mode" in 2016. The idea was to do just enough and no more: some slightest pitch variations, an occasional ratchet as an accent. Too much was the enemy — partly because the steady buzz of high frequencies would compete for clarity with the next class of rappers’ more deadened vocal deliveries. Too much velocity variation and you'll also lose the necessary low-level insistence; it just gets a little too jazzy.
Simple things like the start time of the hi-hat sample – how far behind the beat it sat, also began mattering much more. A lot of drum sample packs you would find on reddit at that time had what appeared to be sloppily prepared samples – there would be silence at the beginning of the file, so the sound didn't immediately trigger upon playback. But this was not sloppy at all; this was by design. And these little details, which were so baked into the production, was one way you could so immediately tell the difference between a "real" trap producer and the leagues of technically brilliant EDM producers who simply couldn't get their beats to bounce right.
(Lately you're just as likely to hear no hi-hat at all, like in this new Skrilla record (thanks Harrison) or the incredibly special 4batz record "act ii: date @ 8", which seems to use a soft white noise sound as a loose timekeeper and hi-hats as tight little accents. The straitjacket of the grid loosens ever so slightly. If there is no hi hat god, all is permitted…)
Which is my long-winded way of saying that maybe jetsonmade was right. (He is right.) Maybe Pooh is a hi hat god. (He kinda is.) What we don't see in the Deconstructed video are the ways the hi-hat, post Pooh, might sit behind the beat now. Or how the relative leveling of the other sounds in the beat affected the saturating and clipping effects of the FL Limiter, which in turn affects the timbre of the hi-hat. We aren't seeing how the extra shaker accent sample may have a ton of 30hz, which triggers compression and gives the beat an additional inadvertent bounce. The juice isn’t in that piano roll, but just because you can’t see it doesn’t mean it isn’t in there. That’s smartdumb. That’s (kinda) my point.
Jetsonmade and Pooh are true masters of smartdumb. Fools for this one, as it were. I also love "Whats Poppin", which they also produced, almost as much as I love "SUGE". It's incredibly difficult to make music this simple, and even more difficult to get the buy-in necessary from artists and their a&rs to get it out there and heard. But this is my favorite kind of music these days. What, wow, oh lord. The kind of music that turns you into a convertible.