New Models from AudioShake to Isolate Strings Instruments + More

New Models from AudioShake to Isolate Strings Instruments + More
It’s not every day we announce entirely new models. That’s because behind every updated and new model are countless months of work by our Research team to understand and identify the unique elements of a sound.
Today, we’re excited to release a model that can isolate strings–a stem type that proved a fun challenge for our teams. String instruments introduce a wide variety of timbres and sonic textures to a composition. And while they’re core to orchestral and modern arrangements alike, they’ve traditionally been difficult for source separation tools to isolate cleanly due to their complex harmonics and expressive dynamics.
Our new string separation model leverages advanced audio source separation and machine learning techniques to isolate string sections with greater clarity and accuracy. It works across orchestral, chamber, and hybrid recordings, distinguishing bowed strings from overlapping sources like synth pads or layered vocals. It also excels at separating solo string performances within dense mixes, delivering clean, high-quality string stems for use in remixes, mastering, and immersive audio applications.
Hear it for yourself:
[demo]
Song: Incarnations by Jacob Cummings
We’ve already seen this unlock new creative possibilities for users working on string-heavy projects—from music producers and remix artists to film and TV sound editors. Whether for reimagining a track or extracting elements for sync licensing, we’re excited to see how this enables new workflows.
New and Improved Drum Isolation Stems
In related model news, we’ve launched an updated drum separation model that significantly improves isolation and clarity across a wide range of percussive elements. From tight, punchy kicks and snares in hip-hop and electronic production to the ambient feel of live acoustic kits, this model better captures and distinguishes individual components within complex rhythmic textures.
How to Access AudioShake Stem Separation Models
These new models are now available across the AudioShake platform, including AudioShake Live, AudioShake Indie, and via our APIs. We can’t wait to hear what you make with them—from immersive listening experiences and re-releases to vocal feature swaps and remixes.