Fellow sound artists often ask this question: how should I organize, catalog, and archive my recordings? I’ve worked on this a lot over the years, and have developed a scheme that works for me for everything I do.
tl;dr / In a nutshell
Stay Chronological: Store recordings in dated folders, with session description after the date. Use subfolders for “raw” “process” and “output.” Exceptions are for significant projects which get their own sound process flow. Why? Because time always happens.
Maintain a “working progress” sound library in addition to your primary library: Scan entire working folders into a continuous-progress database for search, retrieval, auditioning and metadata manipulation. Why? Because it gives you permission to be imperfect and yet still access your sounds.
Export finished edits to a “clean” library: When editing process is complete for a sound set, selected finished exports can be copied to “main” library and scanned into main database. Why? Because everything in this database is clean and ready to use without problems or complications.
As a working sound artist, I record new material quite frequently, and am constantly wanting to access this new material for my projects. However this activity does not always coincide with enough time and focus to complete the requisite editing and sound processing tasks to my satisfaction. Hence my sound library work takes place on an ad-hoc basis as time, needs, and focus arise and shift. The scheme outlined here conforms to a “bare minimum, continuous improvement” ethic. Some raw recordings never get a definitive “final mastered” version, some do. This process is for my personal library, which is constantly drawn on for projects. An organization may have more rigorous requirements for workflow.
My intentions with my system are:
- No questions about where things go. Have a consistent framework that I can use for anything without having to make decisions about organization. Any sound I record, I know where to put it.
- Everything always available. Ready and continuous access to the current state of all sounds in the system, with minimal upfront work required. I can always find anything quickly, at any time.
- Value can be added incrementally over time.
In more detail
At top level: Sounds explicitly part of a project (and only that project) get stored with the project. For instance, interviews recorded for a documentary film. Other sounds – anything that will live outside of one originating project – get stored in the general library.
For me, that path looks like:
Time always happens; a sound was always recorded at a time. So this becomes our key top-level organizing principal. I use a top-level folder per year, and within that, store sounds in folders prefixed with date, then session information such as location, subject, or both as is applicable to the sounds.
Some Example Names from my actual library. First line is folder path.
SoundLibrary/FieldRecordingsWorking/FieldRecordings-2018/ 2018-01 Berkeley Constitution Plaza Visit Ref Sounds.wav 2018-01 Grocery Outlet - Credit Card Scanner Beeps, Metal Sliding Double door.wav 2018-0205 SVN Door Creaks.wav 2018-0208 Ambeo Binaural HRTF Mic Tests, Sennheiser.wav 2018-0208 Ambience SFAI Rooftop City.wav
A few nuances: Sessions which span several days get the start date as their folder date. Sometimes I only use the month, I don’t always choose to be rigorous about the date format, though you can standardize if you like. For me, these names do not need to be machine-extractable so I don’t apply highly rigorous standards.
As you can see, I use a reverse date numbering system, which sorts chronologically in the file system, in the form YYYY-MMDD.
Inside that folder, make a subfolder each for:
- “Raw” raw recordings and other production-gathered info like descriptions, photos, video etc, in appropriate subfolders.
- “Working” edit sessions/project files, Pro Tools sessions, Reaper projects, interim versions such as iZotope RX files, etc. Any audio material that is neither raw nor finished. During editing, to the extent possible, media is referenced at its original location; it is not “moved” into project files. This preserves the original location of source recordings.
- “Output” edited/mastered/finished soundfile output. Write metadata to soundfiles describing the sound, it’s subject, the location, recordist, mics used, etc.
- “Documents” – for instance descriptions, contracts if any, I don’t use this much but occasionally it’s just right.
Workflow / Process
Folder names, file names, and folder hierarchy are how sounds are cataloged and retrieved in Soundminer (and other database systems) so it’s best to change them as little as possible. The workflow is designed with this in mind… once we put something in a place, we try not to change it. If we do need to change it, we make sure the changes propagate into the database so the sounds are still linked to their database records.
Logging during Production
I often take notes on my iPhone during recording. As of 2018 I use “Google Keep” app for this; photos and text can be captured to a document for the session, and retrieved later on desktop computer. Of course indexing these descriptions to the recorder’s file names is important. My notes typically include details which will be useful later in editing – for instance “at 2:03 lawnmower stops”, and I often write a description at this time that I can read as a tail-slate before hitting “stop” on the particular recording. Sometimes I write track markers into files, so I can later find certain sounds or features more quickly. Also – when appropriate, I clap my hands or make another sharp loud sound to indicate starts and stops of “clean” sound, i.e. where to trim heads and tails to remove handling noise, slates, or other extraneous unintended stuff from the body of the sound.
I record currently to Sound Devices 788T, 702T, Sony PCM-M10 and occasionally some other recorders. Only the 788T allows descriptions to be entered in the field using a USB keyboard, and I rarely use this… however, some newer recorders have smartphone apps which allow metadata entry. I suspect I would use these instead of my above method.
I set my recorders to record into daily folders when possible.
Back at the computer, I mount the field recorder’s drive and:
- Directly on the recorder’s media drive, I locate or create the daily folder for the new sounds I am importing. (or a dated session folder if I spanned midnight or had a multi-day session) and possibly add some descriptive terms to this folder.
- On my computer I create the appropriate Session folder, e.g. “…2018/2018-0205 SVN Door Creaks”
- Inside this field session folder I create 3 subfolders labeling them “Raw” “Edit” and “Output”
- I copy the new recordings folder into “Raw”
- After copying, on the field recorder media I move the source folder to a “TransferDone” folder, for temporary backup and to track that ingest has taken place for those sounds.
- I make any last tweaks to name of field session folder, thinking about searchable keywords.
- I gather any additional materials such as photos and descriptions written in the field, filing them into the “Raw” folder in appropriate subfolders.
The entire folder including raw source in scanned into my soundminer “continuous working” library. (Personally, mine is called “02_Garden_of_Sounds”)
Generally, I make a quick pass of labelling the sounds individually, either into the tail end of their filename, or into description fields.
I do a pass of tagging sounds with key metadata which I readily know. Microphone, recorder, recordist name (usually me), project (if the sounds are part of a project.)
I write this metadata back to the files.
Eventually I may write a section here about editorial, but I’d rather publish this sooner.