a better search for max devices

Device Randomizer by Hanako: Randomise Any Parameter in Live

· max4.live editorial

Most randomisation in Ableton Live happens at the note level. You scramble MIDI, you re-trigger a sampler, you nudge velocities. What you rarely get is a single control that reaches into a device's parameter set and shakes it. Device Randomizer by Hanako does exactly that: drop it in, point it at a device, choose which parameters are fair game, and randomise.

It sits in the Audio Effect category because that's where it has to live to be loaded on a track, not because it processes sound. Nothing passes through it. It's a controller for other devices, and that distinction matters for how you place it in a chain.

What version 2.0 actually added

The original idea "randomise all parameters, or pick the ones you want" is simple enough. Version 2.0, released two weeks after the initial upload in January 2010, is where the device gets genuinely interesting.

Two changes:

Hanako spells out the technique in the device description: load preset A, set it as min, Hot Swap to preset B, set it as max. The randomiser now interpolates randomly, parameter by parameter between two presets you already trust. That's a different proposition from pure chaos. You're not rolling dice on the whole parameter space; you're rolling dice inside a corridor you defined.

Three ways to use it in a session

1. Preset-to-preset exploration on a reverb or delay

Take a send with a reverb you use constantly. Find the two presets at either end of what the track needs a short, tight one and a long, washed one. Capture the first as min, Hot Swap the second in, capture as max. Deselect any parameter you don't want touched (dry/wet, if the send is already balanced; the input filter, if you've tuned it).

Now randomise between takes. You get variants that are all plausibly "the reverb for this track" but none of them identical. When one lands, note the settings and move on. This is the most reliable use of the device because the corridor does the curation for you.

2. Sound design starting points, with locks

Put a synth or a dense effect chain on a track and randomise everything except pitch-critical parameters. Hit it twenty times, bounce or screenshot the three that caught your ear. The selective-parameter list is what makes this survivable full randomisation of a complex device gives you silence or noise most of the time. Lock the oscillator tuning, the filter type, the output level, and let envelopes, LFO rates and modulation depths move.

A practical habit: randomise, then immediately narrow. If a result is close, set it as min, tweak by hand to the other edge of what you want, set that as max, and randomise again. You can converge on a sound in three or four rounds instead of dragging every knob.

3. Per-clip variation in a longer arrangement

Because settings persist with the Set, you can keep a configured randomiser on a percussion or texture track for the life of a project. Before duplicating a section, randomise the effect chain inside the min/max corridor. Each copy of the loop gets a slightly different treatment without you automating anything. It's a fast way to stop a 32-bar repeat from sounding like a copy-paste.

One caution for live use: the device randomises on demand, and the results are unpredictable by design. If you're performing, build your corridor narrow and rehearse the range. A min/max pair that sounds good on headphones at home can still produce an outlier at volume.

What 21,293 downloads and a 4/5 rating tell you

Twenty-one thousand downloads for a device from January 2010 is a lot of people reaching for one idea. The 35 comments are the more revealing number that's heavy discussion for a utility, and it usually means people hit questions, edge cases or compatibility issues and came back to talk about them.

The 4/5 rating fits that picture. It's a good score, and it's noticeably below the 5/5 carried by several comparable-popularity devices in the same category. My read: the concept is strong enough to keep pulling downloads for over a decade, while the execution has friction which is what you'd expect from a parameter-mapping utility built against Live 8.1.1 and not updated since.

That's the caveat to take seriously. The listed requirement is Live 8.1.1, and the last update is dated January 16, 2010. If you're on a current version of Live, treat it as something to test in a scratch project before you build a performance around it. The device has no stated licence, so there's nothing published telling you what you may do with it beyond using it.

Where it sits among similar-popularity Audio Effects

The devices clustered around Device Randomizer's download count are mostly doing something else entirely, which is a useful reminder that the Audio Effect category is a catch-all.

Crap Cassette by SESDevices (20,731 downloads, 5/5) emulates a worn C90 played back on a deck that has seen better days. It's a sound, not a tool. High School Crush by CaligulaCuddles (22,140 downloads, 5/5, updated to 1.1 in 2019) is a triple parallel bit crusher and downsampler with crossover wet/dry mixing again, a processor you put audio through.

The closer cousins are the utilities. Skin Editor by pechmann (22,008 downloads, 5/5) edits Live 9 skin colours from inside a device slot; it loads as an Audio Effect for the same reason Device Randomizer does, and we covered it in a separate deep-dive. Live Clip Chopper34 by Myralfur (19,476 downloads, 4/5, added March 2010) controls clip playback position from a monome grid another control surface dressed as an effect, from the same era, with the same 4/5 rating and an even heavier 45 comments.

The pattern across those two 2010 utilities is consistent: ambitious control devices from the early Max for Live years attract a lot of use and a lot of questions, and settle a notch below the perfect-score processors.

Who should skip it

If you do fall into the group it serves sound designers who want structured chance, people who get stuck optimising a single preset for an hour the preset-A-to-preset-B trick alone is worth the install. It's the one feature here that doesn't have an obvious equivalent elsewhere, and it reframes randomisation as interpolation between two decisions you already made.

More from the blog

All posts · Back to device search