Drum racks, samplers and granular sampling: building your own library.

Collecting samples is easy; turning that accumulation into a working library is the job almost nobody tackles in time. A second task usually follows: once material is catalogued, the next step is to stop depending on other people’s material and start creating sounds from what you have.

This guide has two halves. First, four tools for building new sounds from existing audio: drum racks, samplers, granular sampling and resampling. Then library work: how to create your own library, catalogue samples and find samples quickly while you are developing an idea.

What problem each tool solves

Each answers a different question:

• Do I need to trigger many sounds at once? Drum rack.
• Do I need to play a sound like an instrument? Sampler.
• Do I need texture where there was none? Granular sampling.
• Do I need to turn a process into new material? Resampling.

They do not compete: all four coexist in a normal project. The decision is not which is better, but which control you need now.

Drum racks and samplers: two ways to play samples

A drum rack is a container with pads. Each pad holds a file and can have its own effects chain, volume and pan settings, and output. Ableton documents it this way: Drum Rack groups chains, one per pad, with pad views, a chain list and note and velocity zones; macros above them can adjust several pads together.

That makes it the natural tool for percussion and anything organised around individual hits: a complete kit, a set of one-shots cut from a break, or a construction-kit sample drum set. The trick separating a useful rack from a slow one is internal organisation: keep the same pads in the same places across all your racks.

A sampler is different: it reads a file and plays it across the keyboard. You can tune it, apply envelopes, filters and modulation, and distribute multiple notes and velocities across zones—the defining feature of a multisample. In practice, a drum rack organises sounds, while a sampler performs one.

The simple rule: material that will play as a collection of hits goes in a drum rack; material that will play as an instrument—a voice moving up and down, a piano or a pad—goes in a sampler.

Granular sampling: texture from a fragment

Granular sampling cuts audio into grains, very short pieces, and overlaps them into a continuous stream you can manipulate by position, size, pitch and density. The official Granulator II page, Robert Henke’s Max for Live device, describes exactly this: a steady stream of short cross-sectional sections of the source sound, with size, behaviour, modulation and envelope controls.

Practical uses: turn two seconds of a bell into a pad lasting a complete bar; remove a break’s rhythmic character and retain only its colour; extract an atmosphere from a field recording no longer useful for anything else.

Two precautions worth taking from the start:

• You lose rhythm. Granular processing blurs transients. It is ideal for layers and textures, and the wrong choice when trying to preserve a groove.
• Less is more. Only a few active parameters produce movement; adjusting everything together creates the instantly recognisable “plugin” texture.

Resampling: re-recording what already played

Resampling records the project output, or part of it, as new audio. Ableton documents it as an Input Type option named Resampling: route the main output into a track, arm that track and record what plays. The recording track itself is excluded from the take, and files go into Samples/Recorded inside the project folder.

It has two uses, both worthwhile:

1. Commit to a decision. If an effects and automation chain sounds good, re-record it: you can then remove demanding devices and work with audio. The project runs more freely and the sound no longer depends on everything remaining enabled.
2. Create your own material. Loop a section, mute tracks, move filters and record half an hour, then trim it. That file is now yours, avoiding the source problem discussed in the record sampling guide and simplifying licences, because the starting material is your own mix.

A practical precaution: record with headroom. If the resampling track enters the red, the new file is clipped from the outset and cannot be recovered.

Building a library you can search

Creating your own library begins with one simple principle: separate what you buy from what you use. The classic mistake is piling downloaded packs into a huge folder and browsing through it every session.

A useful structure has four layers; the table summarises what each holds and why it exists:

• Original archive: untouched packs with their licences. Never rename or reorganise them; they are proof of purchase and your replacement copy.
• Working library: one shallow folder containing what you actually use, renamed and grouped by type: hits, loops, vocals and textures.
• Favourites: twenty or thirty files per category. This is the folder opened by people who own a thousand kicks but always use the same six.
• Projects: chops, resamples and stems live inside the folder of the track using them. Outside it, they are orphans.

One maintenance rule: each new pack first enters an inbox, and only material that actually gets used in a session moves into the working library. Everything else stays unobtrusively in the original archive.

Cataloguing samples: names, metadata and tags

Cataloguing samples means deciding what information is written into the file itself. The most underestimated tool is its name, because that is all you see in the file explorer, DAW browser and search results.

A useful name contains, in this order, instrument, tempo, key and a short descriptor. A kick does not need tempo; a loop does; a tonal fragment needs a key. Adding the original pack lets you identify its source without opening anything.

Complements to the name:

• File metadata. Some programs write BPM and key into the file; if your workflow reads them, you save time.
• Consistent tags. A small vocabulary of category, character, tempo and key, used the same way every time. Too many tags become useless because you cannot remember them when searching.
• Source record. Alongside the file or in a master document, record the pack and licence. This applies the same principle as the guide to organising the DJ library to samples.

Duplicates become a problem over time: the same kick under three names from three packs. Dedicate a session to removing them before the library becomes unmanageable.

Finding samples quickly

Finding samples quickly depends on a system that responds to how you search rather than an expensive program. Searching rarely means “browse folders until something appears”:

1. Decide what you want first. “Dry hit at 90” is a search; “something good” is not. The clarity of the question determines whether the system works.
2. Search names and tags. Use the system explorer for specific folders, the DAW browser for quick auditioning and a sample manager once the library becomes large.
3. Audition in context. An isolated sample is misleading; always try two or three in the track at its actual tempo.
4. Close the loop. If a sound works, rename it and save it to favourites immediately. Postponed tasks never get done.

After repeating these four routines for a few weeks, the library stops being storage and becomes a working tool: that is the sole justification for the time spent cataloguing.

Next step

With the library organised and your own samples working, the next step is bringing rights decisions up to the same level of order. Review legitimate sources for buying music and complete the cycle with organising the DJ library.