Choosing Sounds
Here is the least glamorous and most valuable idea in music production: the right sound needs almost no processing, and the wrong sound cannot be processed into the right one. A kick that already sits in the track needs no EQ. Spend twenty minutes choosing it and you save two hours mixing it — and you get a better result, because every corrective move you avoid is damage you didn't do. Most people skip this step entirely and then wonder why mixing feels like fighting.
Why Selection Beats Processing
Every processor is a compromise. EQ changes phase. Compression changes transients and dynamics. Saturation adds harmonics that weren't there. Used deliberately these are creative tools — used correctively, to drag a sound toward something it isn't, they are damage you're accepting in exchange for getting closer to what you wanted.
So the question is never "how do I make this snare brighter?" It's "is there a snare that's already bright?" There almost always is, it takes ninety seconds to find, and it will sound better than the one you spent twenty minutes EQing — because a bright snare was recorded bright, with all the harmonic content and transient shape that comes with it, rather than having its existing content amplified.
The tell
If you find yourself making a large corrective move — more than about 4 dB of EQ, or heavy compression just to control level — that's not a mixing problem announcing itself. That's the wrong sound announcing itself.
The compounding cost
Bad choices don't stay contained. A boomy kick forces you to carve the bass, which thins the bass, which makes you boost its mids, which crowds the vocal. One wrong sound generates a chain of compensations that all degrade the mix.
What "professional" often is
A great deal of what people hear as expensive production is simply better source material chosen more carefully, not better processing. The processing on many hit records is surprisingly light.
It's the fastest skill to improve
Unlike mixing, the feedback is instant and unambiguous. You audition, you compare, you notice. Ten sessions of deliberate auditioning will measurably change your records.
Always Choose in Context
If you take one practical habit from this page, take this: audition sounds with the rest of the track playing, never in solo. A sound in isolation tells you almost nothing about whether it belongs, and soloed auditioning is why so many people pick sounds that are individually impressive and collectively unusable.
The reason is masking. What matters is not how a snare sounds — it's what a snare sounds like in the space left over by everything else. A gorgeous, full, complex snare soloed can vanish completely in a busy mix, while a thinner one with a sharp 3 kHz crack cuts straight through. You cannot predict which is which by listening alone, and you don't have to: just loop eight bars of the track and swap sounds underneath it.
| Do this | Instead of | Because |
|---|---|---|
| Loop the busiest section | Auditioning against silence | The hardest place to fit is the only place worth testing |
| Level-match as you swap | Comparing at whatever volume each is | Louder wins every A/B for reasons that aren't quality |
| Give it 5 seconds, then move on | Deliberating over each one | Fit is immediate; if you're arguing yourself into it, it's not the one |
| Shortlist 3, then decide | Committing to the first acceptable one | You need a comparison to know what "good" means today |
| Check it in mono | Judging on width | Wide sounds flatter in solo and collapse in context |
A corollary worth internalising: a sound that needs the whole mix muted to be appreciated is the wrong sound. Nobody will ever hear it that way.
What to Listen For
"Does it sound good" is too vague to act on. These are the specific properties that determine whether a sound will work, and they're worth naming because once you can name them you can hear them.
Frequency footprint
Where does it live? A sound's usefulness is mostly about which region it occupies and whether that region is already taken. Two great sounds in the same range are worse than one great and one adequate in different ranges.
Transient
How sharp is the attack? This is what determines whether something cuts through or sits behind, and it's far harder to add convincingly than to remove. When in doubt, choose the sharper one — you can always soften it.
Length and tail
A long decay eats space and blurs the groove; a short one leaves room. In dense arrangements, shorter is almost always the right answer, which is why so many hit records use kicks that seem too short in solo.
Character vs. neutrality
Distinctive sounds are memorable and inflexible. Neutral sounds are forgettable and cooperative. A track can carry one or two genuinely characterful elements — make them the ones that matter and let the rest be useful.
Tuning
Kicks, 808s, toms and many percussion samples have a definite pitch. If it clashes with the key of the song it will sound subtly wrong forever and you will never work out why. Check the pitch and tune it to the root or the fifth.
Stereo width
Wide sounds are exciting in solo and dangerous in a mix — they collapse in mono and crowd the field. For anything foundational, choose the mono one. Save width for texture.
Not sure what key a sample is in? Drop it into the Song Analyzer, or play it against the Tuner. Most tonal one-shots are easier to pitch than people assume, and tuning your kick to the track is one of those small moves with a disproportionate payoff.
Layering — Building the Sound You Can't Find
Sometimes no single sample does the job, and the answer isn't to process one into submission — it's to take the part you need from each of two sounds. This is how most modern drum production works, and the principle is simple: each layer contributes one thing, and everything else it contributes gets removed.
| Element | Layer 1 provides | Layer 2 provides |
|---|---|---|
| Kick | The sub — weight below 100 Hz | The click — 2–5 kHz attack that survives on small speakers |
| Snare | The body — the 200 Hz "thump" you feel | The crack or noise tail — the 3–8 kHz that cuts |
| Bass | Clean sine sub, mono, no character | A distorted or mid-heavy layer that makes it audible on a phone |
| Lead | The core tone | A detuned or octave-up doubling for size |
Two failure modes to avoid. First: filter each layer to its job. If both kick layers carry full-range low end, you get the same summing mess the low-end rules warn about — high-pass the click layer hard and low-pass the sub layer, so they genuinely occupy different territory. Second: watch phase. Two kicks whose waveforms start in opposite polarity will partially cancel and the layered result will be thinner than either alone. If layering makes something smaller, flip the polarity on one of them and listen again — this counter-intuitive fix solves it surprisingly often.
Taming a Sample Library
The reason most people don't choose sounds carefully isn't laziness — it's that auditioning 4,000 kicks is genuinely miserable, so they use the first one that isn't bad. The fix is to make the library smaller, not to develop more patience.
Build a favourites folder
Every time a sound works, copy it into a small curated folder. Within a few months you'll have thirty kicks you trust instead of four thousand you don't. Start every session there and only go hunting if nothing fits.
Learn one library properly
Knowing a modest collection deeply beats owning an enormous one you browse randomly. Familiarity is what lets you think "I know the sound I want" and go straight to it.
Set a time limit
Give sound selection ten minutes and a decision at the end of it. Endless browsing isn't careful selection — it's procrastination that feels productive, and it's where sessions go to die.
Place a rough sound and move on
If nothing fits immediately, put in something adequate, keep writing, and come back. The arrangement often tells you what the sound needs to be — a decision that was impossible at bar one is obvious once the chorus exists.
Choose It, or Design It?
Selecting from a library is one route. Building the sound yourself is the other, and the trade-off is real: designing takes longer, but you get exactly what you want, nobody else has it, and — most usefully — you understand why it sounds like that, which makes every future choice better informed.
The honest guidance is to do both, on different elements. Use library sounds for the parts that need to be reliable and idiomatic — drums, most bass, anything that has to sound like the genre. Design the one or two elements that are supposed to be the identity of the track. A record needs a small number of things nobody else has, not a full set.
The Synth is built for exactly this: it wires itself up so you can start shaping in seconds, its eighteen presets are teardown material rather than finished answers, and any sound you build can be shared as a link. Pull a preset apart, change one module, and you'll learn more about why a sound works than any amount of reading — including this page.
Audition in context, never in solo. Judge footprint, transient and length rather than "good." Layer for a specific missing ingredient, not for size. And when a corrective move starts getting large, stop mixing and go change the sound — that's the fix.