How is this different from a regular techno prompt?
Standard techno prompts usually sit at 125-150 BPM, most often 128-138, with a clean four-on-the-floor kick (a kick drum on every beat) and understated groove. Hard techno pushes the tempo higher, to roughly 145-160 BPM, and asks for a distorted kick instead of a clean one, plus rave stabs and industrial texture. If your prompt could just as easily describe a smooth club track, add more distortion and grit words until it clearly sounds harder.
What descriptors matter most for hard techno?
- Kick character: distorted, overdriven, torn, clipped, gritty. This is the single word that separates hard techno from techno.
- Tempo: state a BPM in the 145-160 range; leaving it out risks a mid-tempo techno result.
- Stabs: rave stabs, metallic stabs, acid stabs give the track its aggressive top layer.
- Percussion texture: industrial, metallic, clanking, mechanical push the mix away from polished club techno.
- Mood: pick one or two words such as dark, relentless, aggressive, hypnotic, and avoid mixing them with soft or uplifting terms.
Should the prompt include vocals?
Hard techno is almost always instrumental, so write "instrumental" explicitly in the prompt. Square-bracket section tags such as [Instrumental] are a common convention that models follow often but not always, so stating it in plain words as well helps. If you do want a vocal element, keep it to short chopped vocal stabs described by sound ("chopped robotic vocal stab") rather than sung lyrics, since full vocal lines pull the track toward a more song-like, less rave-focused result.
What do models get wrong on this genre?
Fast, distortion-heavy genres like hard techno, along with rock, metal and punk, often come out sounding softer or more polished than intended, closer to generic club music than a genuinely torn, aggressive sound. If the result feels too clean, push the distortion words harder ("heavily distorted", "clipped", "torn") rather than adding more mood adjectives, since texture words move the sound more reliably than feeling words for this genre. Also expect the model to sometimes soften fast hi-hat rolls into a steadier pattern; regenerating with the same prompt and a slightly different tempo can help.
Common mistakes
- Leaving out a tempo number, which lets the model default to a slower, more generic techno pace: state 145-160 BPM.
- Using only mood words like "dark" and "aggressive" without naming the distorted kick, which is the defining sound of the genre: name the kick distortion directly.
- Mixing in soft or ambient descriptors alongside hard techno terms, which confuses the model: keep to one consistent harshness level throughout.
- Asking for sung vocals on a genre that is normally instrumental: write "instrumental" or keep vocal elements to short chopped stabs.
- Adding studio jargon or dB numbers expecting precise mix control: these read as placebo in our measurements, so describe the sound in plain words instead.