Engine-Swap on Two Spatial Audio Plugins Will Be Easy, Right? - Lessons Learned

Learn from real-world challenges in spatial audio plugin development through this conference talk that explores the complexities of swapping convolution engines between two JUCE-based audio plugins. Discover how what initially seemed like a straightforward task of replacing the core of an existing spatial audio plugin (SPARTA 6DoFConv) with a more efficient convolution engine revealed numerous unexpected obstacles and valuable learning opportunities. Explore strategies for understanding and navigating complex, sparsely commented audio DSP codebases, including practical approaches to embedding new convolution engines while managing trade-offs between efficiency and integration complexity. Examine the unforeseen issues that emerged during development, such as handling drawbacks of highly optimized algorithms and implementing crossfade systems for seamless audio transitions. Gain insights into working with multichannel convolution engines specifically designed for Ambisonics reverberation and six-degrees-of-freedom navigation in extended reality applications. Understand best practices for customizing and improving existing plugin code, particularly valuable for developers beginning to work with others' audio software implementations. The presentation covers addressing challenges with optimized algorithms while accepting necessary trade-offs, general lessons learned when working with existing plugin codebases, and practical knowledge of different convolution approaches for spatial audio applications in VR environments.