-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathDASR.html
440 lines (440 loc) · 24.9 KB
/
DASR.html
1
2
3
<!DOCTYPE html><html> <head> <meta http-equiv="content-type" content="text/html; charset=utf-8"> <meta name="viewport" content="width=device-width, initial-scale=1.0, minimum-scale=1.0"> <meta name="description" content="E2E Location Guided ASR - Audio Demo"> <meta name="author" content="Aswin Subramanian"> <title>Directional ASR</title> <!-- Bootflat CSS --> <link rel="stylesheet" href="css/bootflat.css"> <!-- Bootstrap Core CSS --> <link href="css/bootstrap.min.css" rel="stylesheet"> <!-- Custom CSS --> <link href="css/index.css" rel="stylesheet"> <!-- Custom Fonts --> <link href="font-awesome-4.7.0/css/font-awesome.min.css" rel="stylesheet" type="text/css"> <link href="http://fonts.googleapis.com/css?family=Lato:300,400,700,300italic,400italic,700italic" rel="stylesheet" type="text/css"> </head> <body> <div id="container"> <div id="paper_title"> <h2> <meta charset="utf-8"> DIRECTIONAL ASR: A NEW PARADIGM FOR E2E MULTI-SPEAKER SPEECH RECOGNITION WITH SOURCE LOCALIZATION</h2> </div> <div id="paper_authors"> <h4> <b>Authors:</b> <em>Aswin Shanmugam Subramanian, Chao Weng, Shinji Watanabe, Meng Yu, Yong Xu, Shi-Xiong Zhang, and Dong Yu</em></h4> </div> <em> </em> <div id="demos"><em> </em> <h3 style="text-align: center;">Audio Demos</h3> <br> <table class="audio_demos"> <thead> <tr> <th style="text-align:center;"><em>Method</em></th> <th style="text-align: center"><em>Sample 1</em></th> <th style="text-align: center"><em>Sample 2</em></th> <th style="text-align: center"><em>Sample 3</em></th> </tr> </thead> <tbody> <tr style="border-top:1px solid black"> <td style="padding:0 35px 0 35px;"> <b>Input Mixture (Ch-1)</b> </td> <td style="padding: 0px 20px; width: 380px;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/40_RIR-40_spk1-307_spk2-245_444c0212_445o030t.CH0.wav" type="audio/wav"> </audio> </td> <td style="padding:0 20px 0 20px; width: 380px;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/595_RIR-595_spk1-124_spk2-151_440c0213_445o030k.CH0.wav" type="audio/wav"> </audio> </td> <td style="padding:0 20px 0 20px; width: 380px;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/520_RIR-520_spk1-324_spk2-77_440o030k_441c020l.CH0.wav" type="audio/wav"> </audio> </td> </tr> <tr style="border-top:1px solid black"> <td rowspan="2" colspan="1"> <b>Ground Truth (Dry)</b> </td> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/40_RIR-40_spk1-307_spk2-245_444c0212_445o030t.S0.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify"><b>REF1:</b> realized capital gains increased forty two percent to nine hundred nine million dollars from six hundred forty point nine million dollars</p> <p class="MsoNormal" style="text-align: justify"><b>Ref Location 1:</b> 307 degrees</p> </td> <td style="text-align: center;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/595_RIR-595_spk1-124_spk2-151_440c0213_445o030k.S0.wav" type="audio/wav"> </audio><br> <div style="text-align: justify; margin-left: 21px;"><b>REF1:</b> he said such products would be marketed by other companies with experience in that business<o:p></o:p> </div> <p class="MsoNormal" style="text-align: justify"><b>Ref Location 1:</b> 124 degrees</p> <br> </td> <td style="text-align: center;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/520_RIR-520_spk1-324_spk2-77_440o030k_441c020l.S0.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>REF1: </strong>coniston partners of new york said it has a six point eight percent stake in gillette and may seek to acquire the company or gain seats on its board .period<o:p></o:p></p> <p class="MsoNormal" style="text-align: justify"><b>Ref Location 1:</b> 324 degrees</p> </td> </tr> <tr> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/40_RIR-40_spk1-307_spk2-245_444c0212_445o030t.S1.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><b>REF2:</b> but even before the stock market's crash ,comma fed chairman greenspan saw no signs of high inflation .period</p> <p class="MsoNormal" style="text-align: justify"><b>Ref Location 2: </b>245 degrees<br> <o:p></o:p></p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/595_RIR-595_spk1-124_spk2-151_440c0213_445o030k.S1.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>REF2:</b> the diversified utility provides gas and electric services in pennsylvania .period<o:p></o:p></p> <p class="MsoNormal" style="text-align: justify"><b>Ref Location 2: </b>151 degrees</p> </td> <td style="text-align: center;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/520_RIR-520_spk1-324_spk2-77_440o030k_441c020l.S1.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>REF2: </strong>those identified as beneficial owners hold at least ten percent of a company's equity securities </p> <p class="MsoNormal" style="text-align: justify"><b>Ref Location 2:</b> 77 degrees</p> <p class="MsoNormal" style="text-align: justify"></p> </td> </tr> <tr style="border-top:1px solid black"> <td rowspan="2" colspan="1"> <b><em>Oracle</em> Binary Mask</b> </td> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-2-IBM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> realized capital gains increased forty two percent to nine hundred nine million dollars from six hundred forty point nine million dollars</p> </td> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-1-IBM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> he said such products would be marketed by other companies with experience in that business</p> </td> <td style="text-align: center;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-1-IBM.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP1: </strong>coniston partners of new york said it has a six point eight percent stake in gillette and may seek to acquire the company or gain seats on its board .period</p> </td> </tr> <tr> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-1-IBM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"> <b>HYP2:</b> but even before the stock market's crash ,comma <span style="color: red;">said</span> chairman greenspan saw no signs of high inflation .period</p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-2-IBM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"> <b>HYP2:</b> the diversified utility provides gas and electric services <span style="color: red;">and</span> pennsylvania .period </p> </td> <td style="text-align: center;"><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-2-IBM.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP2: </strong>those identified as beneficial owners hold at least ten percent of a company's equity securities</p> </td> </tr> <tr style="border-top:1px solid black"> <td rowspan="2" colspan="1"> <em><b>Oracle </b></em><b>DOA</b><em><b> Mask</b></em> </td> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-2-DOAM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> realized capital gains increased forty two percent to nine hundred nine million dollars from six hundred forty point nine million dollars</p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-1-DOAM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> he said such products would be marketed by other companies with experience in that business</p> </td> <td style="text-align: center;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-2-DOAM.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP1: </strong>coniston partners of new york said it has a six point eight percent stake in <span style="color: red;">july</span> and may seek to acquire the company or gain seats on its board .period</p> </td> </tr> <tr> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-1-DOAM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP2:</b> but even before the stock market's crash ,comma <span style="color: red;">said</span> chairman greenspan saw no signs of high inflation .period </p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-2-DOAM.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"> <b>HYP2:</b> the diversified utility provides gas and electric services in pennsylvania .period </p> </td> <td style="text-align: center;"><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-1-DOAM.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP2: </strong>those identified as beneficial owners hold at least ten percent of <span style="color: red;">the</span> company's equity securities</p> </td> </tr> <tr style="border-top:1px solid black"> <td rowspan="2" colspan="1"><strong>MIMO-Speech</strong></td> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-2-MIMO.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> realized capital gains increased forty two percent to nine hundred nine million dollars from six hundred forty point nine million dollars</p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-2-MIMO.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> <span style="color: red;">they</span>
said such products would be marketed <span style="color: red;">to</span> other companies with experience in that business</p> </td> <td style="text-align: center;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-1-MIMO.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP1: </strong>coniston partners of new york said it has a six point eight percent stake in <span style="color: red;">july</span> and may seek to acquire the company or gain seats on its board .period</p> </td> </tr> <tr> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-1-MIMO.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP2:</b> but even before the stock market's crash ,comma <span style="color: red;">said</span> chairman greenspan <span style="color: red;">stalled</span> no signs of high inflation .period</p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-1-MIMO.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"> <b>HYP2:</b> <span style="color: red;">in</span> the <span style="color: red;">new
york office martin white states</span> and electric services in <span style="color: red;">the northeast and the u. s.</span> .period<o:p></o:p></p> </td> <td style="text-align: center;"><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-2-MIMO.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP2: </strong>those identified as beneficial owners hold at least ten percent of <span style="color: red;">the</span> company's equity securities</p> </td> </tr> <tr style="border-top:1px solid black"> <td rowspan="2" colspan="1"> <b>D-ASR</b> </td> <td> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-1-DASR.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> realized capital gains increased forty two percent to nine hundred nine million dollars from six hundred forty point nine million dollars</p> <p class="MsoNormal" style="text-align: justify;"><b>Predicted Location 1:</b> 307 degrees</p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-2-DASR.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><b>HYP1:</b> he said such products would be marketed by other companies with experience in that business</p> <p class="MsoNormal" style="text-align: justify;"> <b>Predicted Location 1:</b> 123 degrees</p> </td> <td style="text-align: center;"> <audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-2-DASR.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP1: </strong>coniston partners of new york said it has a six point eight percent stake in gillette and may seek to acquire the company or <span style="color: red;">gained</span> seats on its board .period</p> <p class="MsoNormal" style="text-align: justify;"><b>Predicted Location 1:</b> 325 degrees</p> </td> </tr> <tr> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/444_445_40_RIR-40_spk1-307_spk2-245_444c0212_445o030t-2-DASR.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"><b>HYP2:</b> but even before the stock market's crash ,comma <span style="color: red;">said</span> chairman greenspan <span style="color: red;">slowed</span> no signs of high inflation .period</p> <p class="MsoNormal" style="text-align: justify;"><b>Predicted Location 2:</b> 247 degrees</p> </td> <td><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_445_595_RIR-595_spk1-124_spk2-151_440c0213_445o030k-1-DASR.wav" type="audio/wav"> </audio><br> <p class="MsoNormal" style="text-align: justify;"> <b>HYP2:</b> the <span style="color: red;">democrat survived</span> utility provides gas and electric services <span style="color: red;">and</span> pennsylvania .period </p> <p class="MsoNormal" style="text-align: justify;"> <b>Predicted Location 2:</b> 149 degrees</p> </td> <td style="text-align: center;"><audio controls="controls" class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/d-asr_samples/440_441_520_RIR-520_spk1-324_spk2-77_440o030k_441c020l-1-DASR.wav" type="audio/wav"> </audio> <p class="MsoNormal" style="text-align: justify;"><strong>HYP2: </strong>those identified as beneficial owners hold at least ten percent of a company's equity securities</p> <p class="MsoNormal" style="text-align: justify;"><b>Predicted Location 2:</b> 80 degrees</p> </td> </tr> <!-- <tr style="border-top:1px solid black"> <td> <b>E2E MVDR-2 + PF + FT</b> </td> <td> <audio controls class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/HN/E2E-MVDR2+post+FT/440_446_440c0207_446c0210_air-3_ain-4.wav" type="audio/wav"> </audio> </td> <td> <audio controls class="audio-player" preload="metadata" style="width: 240px;"> <source src="audio/HN/E2E-MVDR2+post+FT/444_441_444c020h_441c0215_air-3_ain-4.wav" type="audio/wav"> </audio> </td> </tr> --> </tbody> </table> </div> <div id="footer"> <div> <i>© October 2020</i><br> <img src="img/JHU_logo.jpg" alt="" height="50px"><img src="img/tencent_icon.png" alt="" height="50px"></div> </div> </div> </body></html>