Interface and first use

Here's a small task. Using the sampler below, get a soft piano sound you'd happily play under a slow verse. Count every decision you make before the first note you'd keep.

Maybe it took three decisions: pick a piano, play, try Tone. Maybe it took eight, because you weren't sure which sound was a piano, or whether the keyboard was listening, or what a control would do before you touched it. Every one of those moments is a small chance for someone to give up before they've heard what the instrument can do.

Will a first-time player reach a usable sound faster with three named choices, or with one knob that can do more?

Seconds to music

Some instruments are loved partly because they get you to music almost instantly.

The Suzuki Omnichord, from 1981, has rows of chord buttons and a touch plate you brush with a finger. Press a chord, stroke the plate, and you're strumming a harp-like arpeggio before you've read a word. Small battery-powered groove boxes with a speaker built in let you switch one on and have a sequencer playing a pattern within seconds, with no computer or cables. One portable synth colour-codes its four encoders to match the parameters drawn on its screen, so you know which knob moves which thing without a manual.

None of them is simple underneath. Each one makes the first minute simple, and lets the depth arrive later. That's the pattern to copy. The first minute decides whether there's a second one.

The first minute is a path

An interface isn't a screenshot. It's a path from "I want a sound" to "I'm playing it", with a few choices along the way and feedback after each one. Draw it and the problems become obvious.

The most effective change is usually removing a step, not explaining it. If the default sound is already good, the player skips "find a sound". If the keyboard plays the moment the page opens, they skip "figure out how to make it play". Help text explains a decision; a better default removes the need to make it.

Count the steps that are about music (choosing a darker tone, a shorter note) and the ones that are about getting ready (waiting for something to load, finding the output). Keep the first kind. Shrink the second kind, and when it can't be avoided, like a sample library loading, say clearly what's happening and what to do if it fails.

Name the choice the player is making

On the sampler's piano, Tone offers three words: Natural, Warm and Soft. Underneath, Warm rolls off the top above about 8 kHz and Soft above about 3.5 kHz. A player looking for a gentle verse sound doesn't need to know that. They know what "soft" means, and they can hear whether it's right.

A frequency knob gives more control and asks more of the player: you need to know what 3.5 kHz sounds like to aim for it. A named choice asks less and offers fewer destinations. Neither is better in general. For the first minute, a few well-named choices usually win. For a sound designer an hour in, the knob is what they need. Give the names on the front and the knob behind them.

The same goes for every label. "Space" and "Glue" on a mixing effect describe what you'll hear. "Max reduction" and "band attenuation" describe how it works. Lead with the first kind.

Put the important things first

A clear hierarchy tells the player where to start and what can wait. The play surface, the default sound, the output level and two or three expressive controls deserve the front. Configuration can live a click away.

Pocket, the ducking effect from controls and macros, puts three large knobs on its face and keeps its thirty-odd detailed settings behind an Expert disclosure. A new user has three decisions; an engineer who wants to tune the detector can still get there.

The trap is hiding something basic in the simplification. If everyone has to open the advanced page to turn the volume down or stop a stuck note, the hierarchy is wrong. And check your layout on a small screen: a design that looks clean on a big monitor can push the keyboard below the fold on a laptop, which turns step one into a scavenger hunt.

Pictures that tell the truth

Moving lines are persuasive. A player will trust a curve on screen more than their ears. So a picture has to be honest about what it is.

There are three common kinds. A measurement shows the actual sound passing through right now, like a spectrum analyser. A model draws what the settings should do, like an EQ curve calculated from its knobs. A decoration moves because something is happening, without meaning anything precise. All three are useful. The trouble starts when one looks like another: a model drawn like a meter makes the player believe the sound has been checked when it hasn't.

Maru's EQ draws its curve from the settings, which is perfect for seeing what a preset does and useless for seeing what your recording contains. Say which kind of picture it is, in the label or the caption, and players will use it correctly.

Same idea, different place: Seeing sound shows how every analyser hides some things to show others. An interface adds a second question: what will the player decide because of this picture?

Every hand gets in

People reach your instrument in different ways: a mouse, a trackpad on a train, a MIDI keyboard, only the computer keyboard, a screen reader. The first minute should work for all of them.

Try your path with the keyboard alone. Can you reach the sound picker, change a value, play a note and stop? Can you always see where you are? Can you type an exact value instead of dragging, and get back to the default in one move? Do your control names make sense read aloud, without the colours and icons around them? Each "no" is a player who quietly can't start.

Where it breaks

The two ways a first minute fails are opposites.

Too much help. Tooltips on every knob, a tour that runs before the first note, a paragraph explaining each section. If people get lost, more words rarely fix it. They're lost because there are too many decisions, and help text adds reading on top.

Too much simplification. Strip the screen to three controls and remove the one the task needed, and the player is stuck in a clean, empty room. Simple is good only when what's left still reaches a sound they want.

Hand it to someone

You can't see your own instrument fresh. You know where everything is. The quickest way to find the stall you've stopped noticing is to give the instrument to someone who's never seen it, with one goal ("make a sound for a slow verse") and no explanation. Watch the first time they hesitate. Afterwards, ask what they expected to happen at that moment. Their answer tells you which word, picture or step meant something different to them than to you.

Make it yours

Take an instrument you've made and give its first minute one improvement.

1. Do the task yourself with fresh eyes: open it and get to a sound you'd keep for a slow verse. Count the steps, and mark which ones were about music and which were about getting ready.
2. Remove or shorten one getting-ready step. Make the default playable straight away, move the keyboard up, or show a clear status while something loads. Now a new player hears music sooner.
3. Rename one control for the choice the player is making rather than the mechanism, or turn one technical knob into two or three named choices, with the knob still reachable behind them.
4. Save the new default or layout, and if you can, hand it to someone new with the same goal. If they reach the sound without asking you anything, the change worked.

To feel the difference a good first path makes, open the Sampler from the first lab, find the piano you'd use for that verse and save it as your starting sound.

Quick check

1. Testers keep asking what a knob called "Contour Amount" does. Should you add a tooltip? A tooltip helps, but first ask whether the label names the choice they're making. If the knob makes notes start with more snap, a name like "Snap" or "Attack bite" may answer the question before it's asked. Explanation is the fallback, not the first fix.
2. Your effect shows a beautiful animated frequency curve that moves with the knobs. A user says, "It shows my bass is fine." What's the risk? If the curve is a model of the settings, it says nothing about their bass. It shows what the EQ is set to do, not what the recording contains. Label it so nobody mistakes a model for a measurement.
3. You simplify your synth's front page to three macros, and first-minute times improve. A week later, players complain they can't make a pad. What happened? The simplification removed something the pad job needed, probably a long attack or release, or hid it too deep. Check that each common job is reachable from the front, with the details one click away.

Next steps

• Use it: Presets as product design. Choose starting sounds that help someone understand the controls on first play.
• Use it: Finding an audience. You've watched one new person use your instrument. Who is it actually for, and how would you know they care?