Effect of feature variations on MAEs and RMSEs in the ITD predictions of Model 3. The errors were calculated by evaluating the predictions of Model 3 using inter- and intra-individual variations when extracting the selected anthropometric and HA features from participant 18 (female) multiple times, see Table 3, in comparison to measurement-based ITD estimations. Deviations with respect to the mean MAEs and RMSEs, cf. Table 4, are indicated by Δμ, with percentage deviations provided in brackets.
|Dataset||μ||σ||Δμ (Δμ in %)||μ||σ||Δμ (Δμ in %)|
|HRTF||16.7/21.8||1.8/2.6||−1.3 (−7.2)/−0.6 (−2.7)||16.1/21||2.4/4||−1.9 (−10.3)/−1.4 (−6.3)|
|HARTF (front)||26.8/34.1||7.1/7.7||8.5 (46.7)/11.5 (50.6)||24.5/31.3||0.8/1.4||6.2 (34)/8.6 (38.1)|
|HARTF (rear)||22.2/27.7||6/7.7||4.1 (22.9)/5.2 (22.8)||19.6/24.6||0.4/0.6||1.6 (8.7)/2 (9.1)|
Note: MAEs and RMSEs are provided in μs. μ and σ represent mean and standard deviation, respectively.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.