Recent developments in generative artificial intelligence are spurring innovation in DAW software. A research paper from Georgia Tech, titled Composing with Generative Systems in the Digital Audio Workstation, proposed the term Generative Audio Workstation to describe this emerging class of DAWs. [1]
Three examples of notable GAWs are AIVA, WavTool, and Symphony V. AIVA provides parameter-based AI MIDI song generation within a DAW. WavTool offers a browser DAW equipped with a GPT-4 composition assistant and AI text-to-sample generator. Symphony V provides generative vocal synthesis, note editing, and mixing tools. [2]
Generative AI services have also become available through plugins that integrate with conventional DAWs. The utility of the AI ranges between each plugin; some offer mixing and mastering, others generate melodies and chords, and still others emulate effects using AI.[3] Early efforts from Google included Magenta Studio, DDSP, and Tone Transfer.
To reduce the strain on computer memory, some plugin companies developed thin client VSTs that use resources from a cloud server. For example, the audio-to-midi plugin Samplab offers a desktop application with user authentication and API calls that perform stem separation and MIDI transcription off of the computer's local device. This can improve load speeds or prevent applications from crashing.