# Binaural Target Speech Extraction on WSJ0 Source: [Binaural Target Speech Extraction on WSJ0]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:40:18.936Z Evaluation of Binaural Target Speech Extraction performance on the WSJ0 dataset, utilizing DNSMOS and objective metrics like SI-SDR and PESQ. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Binaural Target Speech Extraction | | Dataset (including any stated split) | WSJ0 | | Metric | DNSMOS Overall | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 6 of 6 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [DNSMOS Overall](): Higher is better; unit: SCALAR. - [DNSMOS Signal](): Higher is better; unit: SCALAR. - [DNSMOS Background](): Higher is better; unit: SCALAR. - [SI-SDR (dB)](): Higher is better; unit: SCALAR. - [WB PESQ](): Higher is better; unit: SCALAR. - [NB PESQ](): Higher is better; unit: SCALAR. - [STOI](): Higher is better; unit: SCALAR. - [Delta ILD (dB)](): Lower is better; unit: SCALAR. - [Delta ITD (mS)](): Lower is better; unit: TIME. ## Recorded results | Method | DNSMOS Overall | Result date | Report paper | | --- | --- | --- | --- | | BDE-BiTSE | 3.23 | Not recorded | 68fa7b4a1e209131cea87913 | | Bi-TSE-HRTF | 3.18 | Not recorded | 68fa7b4a1e209131cea87913 | | Bi-TSE-HRTF-RI | 3.11 | Not recorded | 68fa7b4a1e209131cea87913 | | BLCMV Beamformer | 3.044 | Not recorded | 68fa7b4a1e209131cea87913 | | Bi-TSE-DOA | 2.965 | Not recorded | 68fa7b4a1e209131cea87913 | | Unprocessed | 2.845 | Not recorded | 68fa7b4a1e209131cea87913 | ### BDE-BiTSE DNSMOS Overall: 3.23. Result ID: 699e43005698665ec400766c. [Reporting paper]() Recorded evaluation configuration: {"Type":"DNN-based"}. This record does not assert independent reproduction. ### Bi-TSE-HRTF DNSMOS Overall: 3.18. Result ID: 699e43005698665ec400766f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Variant":"Proposed","Input Clue":"HRTF","Type":"Complex-valued"}. This record does not assert independent reproduction. ### Bi-TSE-HRTF-RI DNSMOS Overall: 3.11. Result ID: 699e43005698665ec400766e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Features":"RI (Real-valued components)","Input Clue":"HRTF"}. This record does not assert independent reproduction. ### BLCMV Beamformer DNSMOS Overall: 3.044. Result ID: 699e43005698665ec400766b. [Reporting paper]() Recorded evaluation configuration: {"Type":"Beamforming"}. This record does not assert independent reproduction. ### Bi-TSE-DOA DNSMOS Overall: 2.965. Result ID: 699e43005698665ec400766d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Clue":"DOA (One-hot encoded)","Type":"Complex-valued"}. This record does not assert independent reproduction. ### Unprocessed DNSMOS Overall: 2.845. Result ID: 699e43005698665ec400766a. [Reporting paper]() Recorded evaluation configuration: {"Condition":"Raw Input"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.