Table of Contents
- 1. Summary of the session
- 2. Sound organisation
- 3. Are search engines musical?
- 4. Categorising sound
- 5. Tagging your sounds
- 6. Sorting sound by various criteria
- 6.1. Linear organisation
- 6.2. Random organisation
- 6.3. Organisation via date
- 6.4. Organisation via a table
- 6.5. Organisation based on other properties
- 7. Sorting sound with sound, an introduction to audioGuide
1 Summary of the session
In this session we covered some of the ground concerning file organisation, meta data and ultimately restructuring sound with sound itself (using AudioGuide). Amidst the technical excitement, we also explored some of the consequences of organising sound, storing it and recalling it in interesting ways.
2 Sound organisation
- Edgar Varese – organised sound, a concept of shaping form and structure by using the nature of sounds themselves. A way of avoiding using melodic and harmonic functions ‘of the past’ in favour of some other innate properties of sound materials and their object-like nature.
- Listen to Bernard Hermann’s The Window from Hitchock’s Phsycho, 1960. This structure is incredibly simple but leads to complex perceptual reveries.
- The form is built up out of a series of simple descending steps and their inversion. A fantastically efficient use of material and idea.
3 Are search engines musical?
This perverse question leads to speculation of the importance and seriousness of search results and how these end up shaping our world or our behaviour. The results of search engines are interesting in that they give us potential to explore new sonic forms and find relationships between sounds that we may not have imagined.
Old and a little creaky, infrared was our own attempt at building a sound library archive. It offers a fantastic record, but the system that runs the archive needs maintenance, a perennial problem with archives of any sort, but particularly acute in the digital age where technologies exist in a state of flux.
FreeSound.org is a sample library populated by a large international usergroup. Anyone can upload and because of this sound quality is variable, description hap-hazard and standards far from ideal, but freesound does offer a very complete way of searching for sound, looking at the sound without having to audition it and finding things that are very difficult to record yourself.
LocusSonus sound map is another form of archive where the archive materials are actually live microphones placed around the world.
4 Categorising sound
Sound categorisation is a slippery subject because as we already know, sound has both an objective dimension that can be analysed (such as loudness, duration, spectra), and its subjective reality which is potentially different for each person hearing the sound. Combine this with the wide variety of naming conventions out there and you quickly realise we have to deal with the issue that a sound can be more than one thing at a time and therefore belong in multiple categories. Biologists have this problem too.
5 Tagging your sounds
It’s very useful and can really help your future self if you inject metadata into your own audio files. Not all file formats can handle meta data and the richest meta data is often held by compressed audio file formats such as the ID3 tags used by mp3 files. FLACs, AIFF, APPLELOSSLESS, CAF, BWAV all have their own fields for information and the fields aren’t the same.
The richness of a photograph in explaining microphone positions during recording is far clearer than a long message to yourself, but WAVs can’t keep a photo within themselves. There is nothing however preventing you from keeping a text file or .CSV file and photos with your source WAV recordings and this is no-doubt a very good habit to get into and we’re encouraging such an approach on this course.
The photo is useful for your instagram and blogging profile, a historical document and part of the memory-bank of your recording session. Text descriptions are useful in terms of explaining both the techniques you used, equipment used on the session and personell. They also make useful research documents for (auto)ethnographic research.
5.1 Broadcast metadata explained
As a fan of the command line for automating processes, I like to use it where possible to inject information into my audio files. This is possible with the command line version of mediaarea.net/BWFMetaEdit. You can install it on OSX using your brew install:
brew install bwfmetaedit
Explore bwfmetaedit’s potential by typing
You’ll see that at the top of all the help that gets spewed the following
Usage: "bwfmetaedit [--Options...] FileName1 [Filename2...]
From the commandline you can inject your name into every file and the names of your collaborators too:
bwfmetaedit --Originator="Martin Parker" *.wav
Note that the file originator is actually the person or entity responsible for the archive. In this case, that’s you.
BWF formats allow for you to add a huge amount of other information to the file such as the engineer, collaborators and so on. The different fields are complex, but worth exploring and deciding which ones you want to complete. The only way to learn what all the different fields are is to read up on this at the link above or within BWFMetaEdit application itself.
6 Sorting sound by various criteria
6.1 Linear organisation
- Advantage that it’s predictable goes from begininning to end in the order you specify, is learnable, repeatable
- Potential to be unsurprising, uninteresting after prolonged exposure
On the command line you can create a new structure with the following command. The
ls part is actually talking to the terminal rather than sox. It calls a list of all the .wav files in the directory.
sox *.wav alphabeticalSort.wav
6.2 Random organisation
- Advantage that it has the potential to be surprising and bring up things that you hadn’t considered before.
- Potential to aid invention and give variety that goes beyond your own habits as a designer
- disadvantages are that you don’t know what’s coming and there may be many surprises that you don’t like and you may have wasted your time.
On the command line with coreutils installed, you can use the shuffle function to randomly sort the files. The
ls part of the string below calls the terminal’s list function. When we have the list, we can do things to it, such as randomly sort the list and then rendering off a sound file. You’ll remember from the lecture that you should probably write the outputs to another directory if you’re going to do this more than once.
sox $(ls *.wav | gshuf) randomShuffle.wav
6.3 Organisation via date
6.3.1 Example order files by date ascending
6.3.2 Example order files by date descending
6.4 Organisation via a table
- You can create curves that iterate through lists. This is a very compositional approach where you can sweep through sound file lists on different journeys
6.5 Organisation based on other properties
It’s possible to find out how long a sound is by getting the duration of the sound file and sorting the list of files in order. The easiest way to do this on the command line is to sort by file size, this works perfectly well provided all sounds have the same samplerate and bit depth.
6.5.1 Example, organisation by file size ascending
-ls -1S *.wav
6.5.2 Example, organisation by file size descending
-ls -1Sr *.wav
7 Sorting sound with sound, an introduction to audioGuide