With DAW developers apparently not interested in integrating serious sample organisation tools into their otherwise-comprehensive production systems, it falls to the likes of Sononym to help us make sense of our increasingly vast libraries.
This handsome application uses machine learning to analyse any number of audio files across any number of ‘libraries’ (each one a folder hierarchy on your hard drive, to the top level of which a moderately sizeable database file is added - about 1GB for a 50GB folder), categorising each one as a loop or one-shot, tagging them by instrument type, and establishing five sonic parameters - Overall, Spectrum, Timbre, Pitch and Amplitude - for implementation of the remarkable ‘Similarity Search’ feature. This last does a surprisingly good job of hunting down and ranking samples according to how closely they match a source file (chosen from within Sononym, dragged in from your DAW or OS, or even recorded directly in) with regard to those ‘aspects’.
The algorithm isn’t perfect, of course, but generally speaking, it really does work, making it easy to discover variations on individual samples from across your no-doubt-chaotic folder structure. Text-based searching is in place, too, and the machine learning analysis also facilitates filtering by a range of properties including length, sample rate, loudness, pitch/frequency, brightness, harmonicity and noisiness. Just want to dig out one-shot basses with the words ‘fat’ and ‘dark’ in the name, pitched at G2, and hitting between -30 and -5dB RMS? Not a problem!
The most glaring issue with Sononym right now is that none of the metadata attached to each sample is at all user-editable. As well as that, we’d like to be able to carry out Similarity Searches across multiple libraries and drag more than one file out to the DAW/desktop at a time, and automatic library updating when new samples are added is an obvious must. We gather all of these are on the to-do list, though, making the future for this powerful new productivity tool look very bright indeed.
