This project simulates a 5.1 virtual surround sound field through binaural audio processing, allowing immersive playback over headphones. Using psychoacoustic cues like ITD (interaural time differences) and ILD (interaural level differences), the system creates spatial perception beyond typical stereo width.
To prototype a virtual surround experience using a stereo output format suitable for headphone listening, simulating surround speaker positions (L, R, C, LS, RS) with dry mono sources.
- A 5.1 multitrack input with mono stems representing five surround positions (L, R, C, LS, RS).
- Applied binaural spatialization using measured HRIR data from the SOFA-format dataset:
D2_HRIR_SOFA/D2_48K_24bit_256tap_FIR_SOFA.sofa - Used the HRIRs to convolve each virtual speaker source for accurate headphone-based spatial positioning.
- Combined the spatialized signals into a final stereo (binaural-style) mix for immersive playback.
- SOFA HRIR convolution for direction-dependent filtering
- Interaural time and level difference simulation
- Pan law adjustments and center imaging
- Spatial summation for a natural-sounding surround image
- HRIR Source:
D2_48K_24bit_256tap_FIR_SOFA.sofa - Format: SOFA (Spatially Oriented Format for Acoustics)
- Resolution: 48 kHz, 24-bit, 256-tap FIR filters
- Used with np.convolve
- A stereo (binaural-style) mix that mimics a surround speaker setup through headphones
🧠 Designed for headphone playback.
Experience the best results using neutral headphones in a quiet listening environment.
VSS_clear_all_outputs.ipynb: Cleaned notebook with full signal flow, spatial routing logic, and audio processing code.
⚠️ Note: Audio files are excluded due to GitHub's size limits. Please contact me for access or recreate using your own 5.1 sources with the code and mix instructions in the notebook.