Don’t Sweat the Technique

I did not expect to hear a classic Eric B & Rakim track on The Crown, but at the end of season five, episode five, there was Dominic West as Prince Charles, dancing to this:

There is a lot going on here! The track opens with an upright bass and a drum kit playing a lightly swinging Latin rhythm. It sounds up close and full. After it plays twice, an entire funk band horn section riff enters, sounding like it’s being played on a small radio hanging from a nail on the wall. Then that same radio plays a riff from a single saxophone while Rakim’s voice, loud and clear, intones, “Don’t sweat the technique.” So that’s the first eight seconds. Next, a red hot distorted breakbeat enters, while the Latin bass/drums groove and the tinny radio sax riff continue looping. You would have no way of knowing that the breakbeat and the saxophone riff are sampled from the same recording, because they are mixed and processed so differently. This new groove plays four times, and then the saxophone riff switches to an angular and unpredictable funk riff using upper extensions of the Eb minor chord. All this before the first verse even starts. It’s a lot!

“Don’t Sweat The Technique” is built from six samples. Four of them are from “Give It Up” by Kool & The Gang. The horn section riff is at 0:38, the George “Funky” Brown drum break is at 1:42, and the two segments of Ronald Bell‘s tenor sax solo are at 2:07 and 2:23.

There are also two samples from “Queen of the Nile” by Young-Holt Unlimited. Both are from Eldee Young‘s bass solo, at 2:17 and 2:22.

Eric B speeds all of these samples up about six percent, raising their pitch a half step. Five of the samples respect the source material’s metrical structure. However, the main instrumental hook, that angular saxophone riff, is placed differently in “Don’t Sweat The Technique” than it is in its original context. Here’s my transcription of Bell’s solo from “Give It Up”, with the two sax samples marked.

And here’s how the samples appear in “Don’t Sweat The Technique”:

In “Don’t Sweat The Technique”, Sax 2 is displaced two beats earlier, so it starts at the beginning of the bar rather than halfway through it. The end of the sample flows seamlessly into its beginning, making the looped version of the phrase feel like an organic whole. This is an underappreciated aspect of the art of sampling. You aren’t just placing an idea in a new musical context; you’re repeating it, thereby connecting its ending with its beginning, creating a musical relationship that was not present in the original. Joseph Schloss helped me understand this in his excellent book, Making Beats: the Art of Sample-Based Hip-Hop:

[L]ooping automatically recasts any musical material it touches, insofar as the end of a phrase is repeatedly juxtaposed with its beginning in a way that was not intended by the original musician. After only a few repetitions, this juxtaposition… begins to take on an air of inevitability. It begins to gather a compositional weight that far exceeds its original significance (p. 137).

I don’t like the idea that sampling is somehow lazy, or that it’s stealing. It takes an enormous amount of trial and error to identify good loops. It’s even more challenging to combine multiple samples from different sources. Even if you can get their pitch and tempo to align (which was quite difficult back in 1992), that doesn’t mean that the melodies, harmonies, timbre and groove will sound good together.

Rap producers don’t just loop the samples, either, they usually do lots of other signal processing too. When I tried recreating “Don’t Sweat The Technique” using Ableton Live, I was able to locate and loop the samples without too much trouble, and I pieced the arrangement together quickly. However, my recreation sounded thin and weak compared to the real thing, and the drums in the Kool & the Gang loops fought hard against the ones in the Young-Holt Unlimited loops. I asked Twitter what Eric B had done to make the track sound so good. Bill Bainbridge had some good mix suggestions:

First guess:
Bass, Drums, Vocals: Indiv. Small Room Reverb + SubMix Buss Long Delay (2-3 Repeats max), Compression + Tiny distortion on Drums w/EQ
Horns: fairly simple Bandpass Filter setups, possibly with a little Distortion, Medium to Small Hall Reverb, SubMix Buss Delay.

— Billsy Bainbridge (@BillBainbridge5) December 9, 2022

David Abravanel had a good guess on what sampler Eric B used:

https://twitter.com/dabravanel/status/1601295903964286976

And Wreck Tech had suggestions for recreating that old-school sampler sound using Ableton Live:

That track was recorded at the hit factory so I don't think they had any super strange board or FX on it. Trying to get old school sound out of Ableton takes a few tricks:

— ẁ̨̨͡r̴̴̀ec̸̢̀̀k̵̴ ̷̡ t̢̀̕eć̡͏̢h̷̷̛͢͢ _ (Nat) (@wreck_tech) December 9, 2022

This last idea is an important one, but it’s technical, so let me break it down a little. The essential concept here is bit depth. Here’s a good explainer. The short version is this: computers (including samplers) encode sound by reading the voltage coming in from the audio source (the microphone, or instrument cable, or in Eric B’s case, the record player.) The computer encodes each voltage reading as a string of ones and zeros. It takes tens of thousands of these readings per second and stores them on disk. To play the sound back, the computer sends those same voltages to the speakers or headphones.

Each one or zero is called a bit. The more bits you use to encode each voltage reading, the more accurate the reading will be. If you only use one bit, you can only measure whether the voltage is on or off. If you use two bits, you can distinguish four possible voltage levels. If you use three bits, you can distinguish eight possible voltage levels. And if you use four bits, you can distinguish sixteen possible voltage levels. Most of the digital audio you hear has a bit depth of sixteen, which gives you 65,536 different possible voltage levels. This is a very fine gradation that permits extremely accurate encoding. However, back in 1992, computer memory was expensive, so samplers were limited to twelve-bit audio. That sounds intuitively like it would be almost as good as sixteen-bit audio, but twelve bits only gives you 4,096 different voltage levels. Twelve-bit audio is low-quality and grungy-sounding, with lots of audible high-frequency distortion caused by quantization error. Also, as my man Ben Casey points out, lower bit depth means a narrower dynamic range, and dynamic range compression makes drums sound hot.

The reason I think Eric B is so great is that he found a way to select and mix his twelve-bit samples so that their grunginess is a positive thing. “Don’t Sweat The Technique” doesn’t sound cheap and primitive. It sounds edgy, aggressive, and exciting. It sounds so good, in fact, that now DAWs come with bitcrusher plugins so you can intentionally degrade the quality of your digital audio to match that classic sound. You can also get plugins that will add vinyl crackle and surface noise to your pristine all-digital signal chain. It might seem perverse to use more advanced technology to simulate less advanced technology, but if it sounds good, it sounds good.

Digital samplers can simulate “real” instruments with varying degrees of fidelity. If you use expensive instrument libraries and you devote a lot of effort to tweaking the parameters, they can be pretty convincing. Most of the time, however, they just sound depressingly fake. The really smart electronic musicians, like Eric B, use digital instruments to sound like digital instruments, to do things that only digital instruments can do. Could you play “Don’t Sweat The Technique” with a band? Sure. Someone out there has probably tried it. Maybe it even sounded good. But there’s no way it sounded as good as those bitcrunched samples, with their specific mixing and EQ and compression and distortion and delay. Hip-hop deserves more respect.