The following audio examples accompany the publications
T. Deppisch, S. Amengual Garí, P. Calamia, and J. Ahrens, “Direct and Residual Subspace Decomposition of Spatial Room Impulse Responses,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 31, pp. 927–942, 2023, doi:10.1109/TASLP.2023.3240657,
and
T. Deppisch, S. Amengual Garí, P. Calamia, J. Ahrens, “Perceptual Evaluation of Spatial Room Impulse Response Extrapolation by Direct and Residual Subspace Decomposition,” AES International Conference on Audio for Virtual and Augmented Reality (AVAR), 2022.
Direct and Residual Subspace Decomposition of SRIRs
These examples are binaural renderings of direct part SRIRs and residual SRIRs as obtained by the direct and residual subspace decomposition.
The first three examples correspond to the ones described in “Direct and Residual Subspace Decomposition of Spatial Room Impulse Responses”. Note that the SRIRs have been transformed to the spherical harmonics domain for binaural rendering. In the paper, this was only done in case of the third example.
Room | Original SRIR | Direct Part SRIR | Residual SRIR |
Conference Room | |||
Concert Hall | |||
Transition Between Office and Anechoic Chamber |
The next examples are the basis for the SRIR extrapolation and are described in “Perceptual Evaluation of Spatial Room Impulse Response Extrapolation by Direct and Residual Subspace Decomposition”. The examples comprise different source and receiver positions in a shoebox-shaped room.
Room | Original SRIR | Direct Part SRIR | Residual SRIR |
S1, R2 | |||
S1, R5 | |||
S2, R2 | |||
S2, R4 | |||
S3, R1 | |||
S3, R3 |
Extrapolation of SRIRs
These binaural renderings demonstrate the different conditions that were used in the listening experiment in “Perceptual Evaluation of Spatial Room Impulse Response Extrapolation by Direct and Residual Subspace Decomposition”. Note that in the listening experiment, the SRIRs were dynamically rendered with a sound field rotation according to head tracker data and the conditions were presented once in the front and once rotated 60° to the right.
Extrapolation E1, from R2 to R5, with source S1:
Condition | Drums | Speech |
Reference SRIR | ||
Static BRIR Plus Direct Sound (stat) | ||
Rotated SRIR Plus Direct Sound (rot) | ||
Rotated Residual Plus Direct Sound (res) | ||
Rotated Residual Plus Salient Reflections and Direct Sound (trans) |
Extrapolation E2, from R2 to R4, with source S2:
Condition | Drums | Speech |
Reference SRIR | ||
Static BRIR Plus Direct Sound (stat) | ||
Rotated SRIR Plus Direct Sound (rot) | ||
Rotated Residual Plus Direct Sound (res) | ||
Rotated Residual Plus Salient Reflections and Direct Sound (trans) |
Extrapolation E3, from R1 to R3, with source S3:
Condition | Drums | Speech |
Reference SRIR | ||
Static BRIR Plus Direct Sound (stat) | ||
Rotated SRIR Plus Direct Sound (rot) | ||
Rotated Residual Plus Direct Sound (res) | ||
Rotated Residual Plus Salient Reflections and Direct Sound (trans) |