index.html 77 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235123612371238123912401241124212431244124512461247124812491250125112521253125412551256125712581259126012611262126312641265126612671268126912701271127212731274127512761277127812791280128112821283128412851286128712881289129012911292129312941295129612971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355135613571358135913601361
  1. <!DOCTYPE html>
  2. <html>
  3. <head>
  4. <meta charset="UTF-8">
  5. <title>Enhanced Direct Speech-to-Speech Translation Using Self-supervised Pre-training and Data Augmentation</title>
  6. <link rel="stylesheet" type="text/css" href="styles.css">
  7. <script src="jquery-3.5.js"></script>
  8. </head>
  9. <body>
  10. <div class="container">
  11. <div id="text1">Enhanced Direct Speech-to-Speech Translation Using Self-supervised Pre-training and Data
  12. Augmentation</div>
  13. <div id="intro">
  14. <br>
  15. <p>
  16. Sravya Popuri<sup>&#9734;</sup>, Peng-Jen Chen<sup>&#9734;</sup>, Changhan
  17. Wang, Juan Pino, Yossi Adi,
  18. Jiatao Gu, Wei-Ning Hsu<sup>&dagger;</sup>, Ann Lee<sup>&dagger;</sup> <br>
  19. <font size="-1">(&#9734; = Equal contribution and &dagger; = Equal supervision)</font>
  20. </p>
  21. </p>
  22. <p>
  23. [<a href="https://arxiv.org/abs/2204.02967">paper</a>]
  24. </p>
  25. </div>
  26. </div>
  27. <div class="content-container">
  28. <p>
  29. We explore self-supervised pre-training with unlabeled speech data and data augmentation to improve direct
  30. speech-to-speech model training. We take advantage of a recently proposed speech-to-unit translation (S2UT)
  31. framework that encodes
  32. target
  33. speech into discrete representations, and study both speech encoder and discrete unit decoder pre-training
  34. as well as
  35. efficient partial finetuning methods. We conduct experiments under various data setups and show that
  36. self-supervised
  37. pre-training consistently improves model performance compared with multitask learning and is complementary
  38. to data
  39. augmentation techniques that apply ASR and MT models to create weakly supervised training data.
  40. </p>
  41. <ul>
  42. <li><a style="color:rgb(90, 4, 83)" href="#ES-EN Comparison with Baselines">Spanish To English</a></li>
  43. <ul>
  44. <li><a style="color:rgb(90, 4, 83)" href="#ES-EN Comparison with Baselines">Comparison with
  45. Baselines</a></li>
  46. <li><a style="color:rgb(90, 4, 83)" href="#ES-EN Different Data Setups">Different Data Setups</a></li>
  47. </ul>
  48. <li><a style="color:rgb(90, 4, 83)" href="#EN-ES Comparison with Baselines">English To Spanish</a></li>
  49. <ul>
  50. <li><a style="color:rgb(90, 4, 83)" href="#EN-ES Comparison with Baselines">Comparison with
  51. Baselines</a></li>
  52. <li><a style="color:rgb(90, 4, 83)" href="#EN-ES Different Data Setups">Different Data Setups</a></li>
  53. </ul>
  54. </ul>
  55. </div>
  56. <link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/font-awesome/4.3.0/css/font-awesome.min.css">
  57. <div id="ES-EN Comparison with Baselines" class="content-container">
  58. <script src="wavesurfer.js"></script>
  59. <div class="content-title">
  60. <font size="+5">Spanish To English</font>
  61. </div>
  62. <div class="content-subtitle">Comparison with Baselines
  63. </div>
  64. <p> We provide ground truth source and target audios with the corresponding reference text,
  65. as well as audio samples from three systems: <br>
  66. (1) <strong>S2UT+LNA-D</strong>: the proposed direct speeech-to-unit translation
  67. system initialized with wav2vec 2.0 encoder, unit mBART decoder and finetuned using LNA-D strategy<br>
  68. (2) <strong>Supervised S2UT</strong>: a baseline direct speech-to-unit translation system trained with
  69. source and target text as auxiliary task targets.
  70. <br>
  71. (3) <strong>S2T+TTS:</strong> a baseline cascaded system with a speech-to-text translation model initialized
  72. with wav2vec 2.0 encoder and a randomly initialized decoder, followed by a text-to-speech synthesis model. <br>
  73. Both (1) and (2) use an open sourced HiFi-GAN vocoder to convert units to waveforms.
  74. </p>
  75. <table border="0" class="inlineTable">
  76. <tr>
  77. <th></th>
  78. <th colspan="2">Ground truth</th>
  79. <th colspan="3">Predictions</th>
  80. </tr>
  81. <tr>
  82. <th></th>
  83. <th>Source (Spanish)</th>
  84. <th>Target (English)</th>
  85. <th>S2UT+LNA-D</th>
  86. <th>Supervised S2UT</th>
  87. <th>S2T+TTS</th>
  88. </tr>
  89. <tr>
  90. <th colspan="6" style="text-align:left">Sample 1: S2UT+LNAD performs best</th>
  91. </tr>
  92. <tr>
  93. <th></th>
  94. <th>
  95. <div id="src_waveform_1"></div>
  96. <button id="written_source__header" class="play-button-demo btn btn-primary"
  97. onclick="src_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  98. </button>
  99. <script> var src_1 = WaveSurfer.create({ container: '#src_waveform_1', waveColor: 'violet', progressColor: 'purple' });
  100. src_1.load('./audios/es-en/set1/source/11375_cv.wav'); </script>
  101. </th>
  102. <th>
  103. <div id="target_waveform_1"></div>
  104. <button id="written_target__header" class="play-button-demo btn btn-primary"
  105. onclick="tgt_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  106. </button>
  107. <script> var tgt_1 = WaveSurfer.create({ container: '#target_waveform_1', waveColor: 'violet', progressColor: 'purple' });
  108. tgt_1.load('./audios/es-en/set1/target/11375_cv.wav'); </script>
  109. </th>
  110. <th>
  111. <div id="s2ut_waveform_1"></div>
  112. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  113. onclick="s2ut_lnad_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  114. Pause </button>
  115. <script> var s2ut_lnad_1 = WaveSurfer.create({ container: '#s2ut_waveform_1', waveColor: 'violet', progressColor: 'purple' });
  116. s2ut_lnad_1.load('./audios/es-en/set1/s2ut_lnd/11375_cv.wav'); </script>
  117. </th>
  118. <th>
  119. <div id="translatotron_waveform_1"></div>
  120. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  121. onclick="s2ut_mt_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  122. Pause
  123. </button>
  124. <script> var s2ut_mt_1 = WaveSurfer.create({ container: '#translatotron_waveform_1', waveColor: 'violet', progressColor: 'purple' });
  125. s2ut_mt_1.load('./audios/es-en/set1/s2ut_mt/11375_cv.wav'); </script>
  126. </th>
  127. <th>
  128. <div id="s2ttts_waveform_1"></div>
  129. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  130. onclick="cas_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  131. </button>
  132. <script> var cas_1 = WaveSurfer.create({ container: '#s2ttts_waveform_1', waveColor: 'violet', progressColor: 'purple' });
  133. cas_1.load('./audios/es-en/set1/s2t_tts/11375_cv.wav'); </script>
  134. </th>
  135. </tr>
  136. <tr>
  137. <th> Reference: </th>
  138. <td>autobuses adicionales normalmente proporcionados por go south coast
  139. van desde bristol al festival
  140. </td>
  141. <td STYLE="text-transform:lowercase"> ADDITIONAL BUSES USUALLY PROVIDED BY GO SOUTH COAST GO FROM
  142. BRISTOL TO THE FESTIVAL</td>
  143. </tr>
  144. <tr>
  145. <th> ASR: </th>
  146. <td> </td>
  147. <td> </td>
  148. <td STYLE="text-transform:lowercase">ADDITIONAL BUSES USUALLY PROVIDED BY GO SOUTH COAST GO FROM BRISTOL
  149. TO THE FESTIVAL </td>
  150. <td STYLE="text-transform:lowercase">ADDITIONAL UP TO BORSES NORMALLY PROVIDED BY COAST SO CAST BANDS OF
  151. BRISTOL ALL FESTIVAL</td>
  152. <td STYLE="text-transform:lowercase">ADDITIONAL BUSES USUALLY PROVIDED BY GO SOUTH COAST GO FROM BRUCE
  153. TO THE FESTIVAL</td>
  154. </tr>
  155. <tr>
  156. <th colspan="6" style="text-align:left">Sample 2: S2UT+LNAD performs best</th>
  157. </tr>
  158. <tr>
  159. <th></th>
  160. <th>
  161. <div id="src_waveform_2"></div>
  162. <button id="written_source__header" class="play-button-demo btn btn-primary"
  163. onclick="src_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  164. </button>
  165. <script> var src_2 = WaveSurfer.create({ container: '#src_waveform_2', waveColor: 'violet', progressColor: 'purple' });
  166. src_2.load('./audios/es-en/set1/source/2692_cv.wav'); </script>
  167. </th>
  168. <th>
  169. <div id="target_waveform_2"></div>
  170. <button id="written_target__header" class="play-button-demo btn btn-primary"
  171. onclick="tgt_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  172. </button>
  173. <script> var tgt_2 = WaveSurfer.create({ container: '#target_waveform_2', waveColor: 'violet', progressColor: 'purple' });
  174. tgt_2.load('./audios/es-en/set1/target/2692_cv.wav'); </script>
  175. </th>
  176. <th>
  177. <div id="s2ut_waveform_2"></div>
  178. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  179. onclick="s2ut_lnad_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  180. Pause </button>
  181. <script> var s2ut_lnad_2 = WaveSurfer.create({ container: '#s2ut_waveform_2', waveColor: 'violet', progressColor: 'purple' });
  182. s2ut_lnad_2.load('./audios/es-en/set1/s2ut_lnd/2692_cv.wav'); </script>
  183. </th>
  184. <th>
  185. <div id="translatotron_waveform_2"></div>
  186. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  187. onclick="s2ut_mt_2.playPause()">
  188. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  189. </button>
  190. <script> var s2ut_mt_2 = WaveSurfer.create({ container: '#translatotron_waveform_2', waveColor: 'violet', progressColor: 'purple' });
  191. s2ut_mt_2.load('./audios/es-en/set1/s2ut_mt/2692_cv.wav'); </script>
  192. </th>
  193. <th>
  194. <div id="s2ttts_waveform_2"></div>
  195. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  196. onclick="cas_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  197. </button>
  198. <script> var cas_2 = WaveSurfer.create({ container: '#s2ttts_waveform_2', waveColor: 'violet', progressColor: 'purple' });
  199. cas_2.load('./audios/es-en/set1/s2t_tts/2692_cv.wav'); </script>
  200. </th>
  201. </tr>
  202. <tr>
  203. <th> Reference: </th>
  204. <td>encontró un país con dos gobiernos en la capital maximiliano era el
  205. emperador </td>
  206. <td STYLE="text-transform:lowercase">HE FOUND A COUNTRY WITH TWO GOVERNMENTS IN THE CAPITAL MAXIMILIAN
  207. WAS THE EMPEROR </td>
  208. </tr>
  209. <tr>
  210. <th> ASR: </th>
  211. <td> </td>
  212. <td> </td>
  213. <td STYLE="text-transform:lowercase"> HE FOUND A COUNTRY WITH TWO GOVERNMENTS IN THE CAPITAL MAXIMILIAN
  214. WAS THE EMPEROR</td>
  215. <td STYLE="text-transform:lowercase"> HE FOUND A COUNTRY WITH TWO GOVERNMENTS IN THE CAPITAL THE MOST
  216. SIMILIAN CAPITAL WAS THE EMPEROR
  217. </td>
  218. <td STYLE="text-transform:lowercase">HE FOUND A COUNTRY WITH TWO GOVERNMENTS AND THE CAPITAL MAXIMILIAN
  219. WAS AN EMPEROR</td>
  220. </tr>
  221. <tr>
  222. <th colspan="6" style="text-align:left">Sample 3: S2T+TTS performs best
  223. </th>
  224. </tr>
  225. <tr>
  226. <th></th>
  227. <th>
  228. <div id="src_waveform_3"></div>
  229. <button id="written_source__header" class="play-button-demo btn btn-primary"
  230. onclick="src_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  231. </button>
  232. <script> var src_3 = WaveSurfer.create({ container: '#src_waveform_3', waveColor: 'violet', progressColor: 'purple' });
  233. src_3.load('./audios/es-en/set1/source/1507_epst.wav'); </script>
  234. </th>
  235. <th>
  236. <div id="target_waveform_3"></div>
  237. <button id="written_target__header" class="play-button-demo btn btn-primary"
  238. onclick="tgt_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  239. </button>
  240. <script> var tgt_3 = WaveSurfer.create({ container: '#target_waveform_3', waveColor: 'violet', progressColor: 'purple' });
  241. tgt_3.load('./audios/es-en/set1/target/1507_epst.wav'); </script>
  242. </th>
  243. <th>
  244. <div id="s2ut_waveform_3"></div>
  245. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  246. onclick="s2ut_lnad_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  247. Pause </button>
  248. <script> var s2ut_lnad_3 = WaveSurfer.create({ container: '#s2ut_waveform_3', waveColor: 'violet', progressColor: 'purple' });
  249. s2ut_lnad_3.load('./audios/es-en/set1/s2ut_lnd/1507_epst.wav'); </script>
  250. </th>
  251. <th>
  252. <div id="translatotron_waveform_3"></div>
  253. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  254. onclick="s2ut_mt_3.playPause()">
  255. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  256. </button>
  257. <script> var s2ut_mt_3 = WaveSurfer.create({ container: '#translatotron_waveform_3', waveColor: 'violet', progressColor: 'purple' });
  258. s2ut_mt_3.load('./audios/es-en/set1/s2ut_mt/1507_epst.wav'); </script>
  259. </th>
  260. <th>
  261. <div id="s2ttts_waveform_3"></div>
  262. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  263. onclick="cas_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  264. </button>
  265. <script> var cas_3 = WaveSurfer.create({ container: '#s2ttts_waveform_3', waveColor: 'violet', progressColor: 'purple' });
  266. cas_3.load('./audios/es-en/set1/s2t_tts/1507_epst.wav'); </script>
  267. </th>
  268. </tr>
  269. <tr>
  270. <th> Reference: </th>
  271. <td> otro aspecto más institucional es el equilibrio de fuerzas entre
  272. el parlamento y el consejo</td>
  273. <td STYLE="text-transform:lowercase"> ANOTHER MORE INSTITUTIONAL ASPECT IS THE BALANCE OF POWER BETWEEN
  274. PARLIAMENT AND THE COUNCIL</td>
  275. </tr>
  276. <tr>
  277. <th> ASR: </th>
  278. <td> </td>
  279. <td> </td>
  280. <td STYLE="text-transform:lowercase">ANOTHER MORE INSTITUTIONAL ASPECT IS THE BALANCE OF FORCES BETWEEN
  281. PARLIAMENT AND THE COUNCIL</td>
  282. <td STYLE="text-transform:lowercase"> ANOTHER MORE INSTITUTIONAL ASPECT IS THE BALANCE OF FORCES BETWEEN
  283. PARLIAMENT AND THE COUNCIL</td>
  284. <td STYLE="text-transform:lowercase"> ANOTHER MORE INSTITUTIONAL ASPECT IS THE BALANCE OF POWER BETWEEN
  285. PARLIAMENT AND THE COUNCIL</td>
  286. </tr>
  287. <tr>
  288. <th colspan="6" style="text-align:left">Sample 4: All systems make errors</th>
  289. </tr>
  290. <tr>
  291. <th></th>
  292. <th>
  293. <div id="src_waveform_4"></div>
  294. <button id="written_source__header" class="play-button-demo btn btn-primary"
  295. onclick="src_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  296. </button>
  297. <script> var src_4 = WaveSurfer.create({ container: '#src_waveform_4', waveColor: 'violet', progressColor: 'purple' });
  298. src_4.load('./audios/es-en/set1/source/1700_epst.wav'); </script>
  299. </th>
  300. <th>
  301. <div id="target_waveform_4"></div>
  302. <button id="written_target__header" class="play-button-demo btn btn-primary"
  303. onclick="tgt_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  304. </button>
  305. <script> var tgt_4 = WaveSurfer.create({ container: '#target_waveform_4', waveColor: 'violet', progressColor: 'purple' });
  306. tgt_4.load('./audios/es-en/set1/target/1700_epst.wav'); </script>
  307. </th>
  308. <th>
  309. <div id="s2ut_waveform_4"></div>
  310. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  311. onclick="s2ut_lnad_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  312. Pause </button>
  313. <script> var s2ut_lnad_4 = WaveSurfer.create({ container: '#s2ut_waveform_4', waveColor: 'violet', progressColor: 'purple' });
  314. s2ut_lnad_4.load('./audios/es-en/set1/s2ut_lnd/1700_epst.wav'); </script>
  315. </th>
  316. <th>
  317. <div id="translatotron_waveform_4"></div>
  318. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  319. onclick="s2ut_mt_4.playPause()">
  320. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  321. </button>
  322. <script> var s2ut_mt_4 = WaveSurfer.create({ container: '#translatotron_waveform_4', waveColor: 'violet', progressColor: 'purple' });
  323. s2ut_mt_4.load('./audios/es-en/set1/s2ut_mt/1700_epst.wav'); </script>
  324. </th>
  325. <th>
  326. <div id="s2ttts_waveform_4"></div>
  327. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  328. onclick="cas_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  329. </button>
  330. <script> var cas_4 = WaveSurfer.create({ container: '#s2ttts_waveform_4', waveColor: 'violet', progressColor: 'purple' });
  331. cas_4.load('./audios/es-en/set1/s2t_tts/1700_epst.wav'); </script>
  332. </th>
  333. </tr>
  334. <tr>
  335. <th> Reference: </th>
  336. <td>además su capacidad de regeneración es muy limitada </td>
  337. <td STYLE="text-transform:lowercase"> MOREOVER THEIR CAPACITY FOR REGENERATION IS VERY LIMITED</td>
  338. </tr>
  339. <tr>
  340. <th> ASR: </th>
  341. <td> </td>
  342. <td> </td>
  343. <td STYLE="text-transform:lowercase"> MOREOVER ITS CAPACITY FOR REGENERATION IS VERY LIMITED</td>
  344. <td STYLE="text-transform:lowercase"> IN ADDITION HIS REGENERATION CAPACITY IS VERY LIMITED</td>
  345. <td STYLE="text-transform:lowercase"> MOREOVER ITS RECOVERY IS VERY LIMITED</td>
  346. </tr>
  347. </table>
  348. </div>
  349. <div id="ES-EN Different Data Setups" class="content-container">
  350. <script src="wavesurfer.js"></script>
  351. <div class="content-title">
  352. <font size="+5">Spanish To English</font>
  353. </div>
  354. <div class="content-subtitle">Different Data Setups
  355. </div>
  356. <p> We provide ground truth source and target audios with the corresponding reference text,
  357. as well as audio samples from three systems. All the three models are initialized with wav2vec 2.0 encoder,
  358. unit
  359. mBART decoder and finetuned using LNA-D strategy but use different datasets for finetuning: <br>
  360. (1) <strong>S2UT_Base</strong>: finetuned on the combination of CoVoST2, Europarl-ST, mTEDx datasets.
  361. <br>
  362. (2) <strong>S2UT_LR</strong>: finetuned on low resource setup with 50hr of data sampled from the the
  363. combination of CoVoST2, Europarl-ST, mTEDx datasets
  364. <br>
  365. (3) <strong>S2UT_Aug:</strong> finetuned on the the combination of CoVoST2, Europarl-ST, mTEDx datasets
  366. datasets plus the ASR data. <br>
  367. All models use an open sourced HiFi-GAN vocoder to convert units to waveforms.
  368. </p>
  369. <table border="0" class="inlineTable">
  370. <tr>
  371. <th></th>
  372. <th colspan="2">Ground truth</th>
  373. <th colspan="3">Predictions</th>
  374. </tr>
  375. <tr>
  376. <th></th>
  377. <th>Source (Spanish)</th>
  378. <th>Target (English)</th>
  379. <th>S2UT_Base</th>
  380. <th>S2UT_LR</th>
  381. <th>S2UT_Aug</th>
  382. </tr>
  383. <tr>
  384. <th colspan="6" style="text-align:left">Sample 1: All systems do well</th>
  385. </tr>
  386. <tr>
  387. <th></th>
  388. <th>
  389. <div id="src1_waveform1_1"></div>
  390. <button id="written_source__header" class="play-button-demo btn btn-primary"
  391. onclick="src1_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  392. Pause
  393. </button>
  394. <script> var src1_1 = WaveSurfer.create({ container: '#src1_waveform1_1', waveColor: 'violet', progressColor: 'purple' });
  395. src1_1.load('./audios/es-en/set2/source/9756_cv.wav'); </script>
  396. </th>
  397. <th>
  398. <div id="target_waveform1_1"></div>
  399. <button id="written_target__header" class="play-button-demo btn btn-primary"
  400. onclick="tgt1_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  401. Pause
  402. </button>
  403. <script> var tgt1_1 = WaveSurfer.create({ container: '#target_waveform1_1', waveColor: 'violet', progressColor: 'purple' });
  404. tgt1_1.load('./audios/es-en/set2/target/9756_cv.wav'); </script>
  405. </th>
  406. <th>
  407. <div id="s2ut_waveform1_1"></div>
  408. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  409. onclick="s2ut_lnad1_1.playPause()"> <i class="fa fa-play"></i> Play / <i
  410. class="fa fa-pause"></i>
  411. Pause </button>
  412. <script> var s2ut_lnad1_1 = WaveSurfer.create({ container: '#s2ut_waveform1_1', waveColor: 'violet', progressColor: 'purple' });
  413. s2ut_lnad1_1.load('./audios/es-en/set2/s2ut_lnd/9756_cv.wav'); </script>
  414. </th>
  415. <th>
  416. <div id="translatotron_waveform1_1"></div>
  417. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  418. onclick="s2ut_lr50_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  419. Pause
  420. </button>
  421. <script> var s2ut_lr50_1 = WaveSurfer.create({ container: '#translatotron_waveform1_1', waveColor: 'violet', progressColor: 'purple' });
  422. s2ut_lr50_1.load('./audios/es-en/set2/s2ut_lr50/9756_cv.wav'); </script>
  423. </th>
  424. <th>
  425. <div id="s2ttts_waveform1_1"></div>
  426. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  427. onclick="s2ut_asr1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  428. Pause
  429. </button>
  430. <script> var s2ut_asr1 = WaveSurfer.create({ container: '#s2ttts_waveform1_1', waveColor: 'violet', progressColor: 'purple' });
  431. s2ut_asr1.load('./audios/es-en/set2/s2ut_lnd_w_asr/9756_cv.wav'); </script>
  432. </th>
  433. </tr>
  434. <tr>
  435. <th> Reference: </th>
  436. <td>cada uno de ellos es un derecho exclusivo sujeto a ciertas
  437. limitaciones y excepciones
  438. </td>
  439. <td STYLE="text-transform:lowercase"> EACH ONE OF THEM IS AN EXCLUSIVE RIGHT SUBJECT TO CERTAIN
  440. LIMITATIONS AND EXCEPTIONS</td>
  441. </tr>
  442. <tr>
  443. <th> ASR: </th>
  444. <td> </td>
  445. <td> </td>
  446. <td STYLE="text-transform:lowercase">EACH OF THEM IS AN EXCLUSIVE RIGHT SUBJECT TO CERTAIN LIMITATIONS
  447. AND EXCEPTIONS</td>
  448. <td STYLE="text-transform:lowercase">EACH ONE OF THEM IS AN EXCLUSIVE RIGHT SUBJECT TO CERTAIN
  449. LIMITATIONS AND EXCEPTIONS</td>
  450. <td STYLE="text-transform:lowercase">EACH OF THEM IS AN EXCLUSIVE RIGHT SUBJECT TO CERTAIN LIMITATIONS
  451. AND EXCEPTIONS</td>
  452. </tr>
  453. <tr>
  454. <th colspan="6" style="text-align:left">Sample 2: S2UT_LR performs best</th>
  455. </tr>
  456. <tr>
  457. <th></th>
  458. <th>
  459. <div id="src1_waveform1_2"></div>
  460. <button id="written_source__header" class="play-button-demo btn btn-primary"
  461. onclick="src1_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  462. Pause
  463. </button>
  464. <script> var src1_2 = WaveSurfer.create({ container: '#src1_waveform1_2', waveColor: 'violet', progressColor: 'purple' });
  465. src1_2.load('./audios/es-en/set2/source/12478_cv.flac'); </script>
  466. </th>
  467. <th>
  468. <div id="target_waveform1_2"></div>
  469. <button id="written_target__header" class="play-button-demo btn btn-primary"
  470. onclick="tgt1_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  471. Pause
  472. </button>
  473. <script> var tgt1_2 = WaveSurfer.create({ container: '#target_waveform1_2', waveColor: 'violet', progressColor: 'purple' });
  474. tgt1_2.load('./audios/es-en/set2/target/12478_cv.wav'); </script>
  475. </th>
  476. <th>
  477. <div id="s2ut_waveform1_2"></div>
  478. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  479. onclick="s2ut_lnad1_2.playPause()"> <i class="fa fa-play"></i> Play / <i
  480. class="fa fa-pause"></i>
  481. Pause </button>
  482. <script> var s2ut_lnad1_2 = WaveSurfer.create({ container: '#s2ut_waveform1_2', waveColor: 'violet', progressColor: 'purple' });
  483. s2ut_lnad1_2.load('./audios/es-en/set2/s2ut_lnd/12478_cv.wav'); </script>
  484. </th>
  485. <th>
  486. <div id="translatotron_waveform1_2"></div>
  487. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  488. onclick="s2ut_lr50_2.playPause()">
  489. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  490. </button>
  491. <script> var s2ut_lr50_2 = WaveSurfer.create({ container: '#translatotron_waveform1_2', waveColor: 'violet', progressColor: 'purple' });
  492. s2ut_lr50_2.load('./audios/es-en/set2/s2ut_lr50/12478_cv.wav'); </script>
  493. </th>
  494. <th>
  495. <div id="s2ttts_waveform1_2"></div>
  496. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  497. onclick="s2ut_asr2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  498. Pause
  499. </button>
  500. <script> var s2ut_asr2 = WaveSurfer.create({ container: '#s2ttts_waveform1_2', waveColor: 'violet', progressColor: 'purple' });
  501. s2ut_asr2.load('./audios/es-en/set2/s2ut_lnd_w_asr/12478_cv.wav'); </script>
  502. </th>
  503. </tr>
  504. <tr>
  505. <th> Reference: </th>
  506. <td>esta experiencia representa un paso trascendental en la historia
  507. espacial del país </td>
  508. <td STYLE="text-transform:lowercase">THIS EXPERIENCE REPRESENTS A TRANSCENDENTAL STEP IN THE SPATIAL
  509. HISTORY OF THE COUNTRY</td>
  510. </tr>
  511. <tr>
  512. <th> ASR: </th>
  513. <td> </td>
  514. <td> </td>
  515. <td STYLE="text-transform:lowercase">THIS EXPERIENCE REPRESENTS A TRANSCENDENT STEP IN THE SPACE HISTORY
  516. OF THE COUNTRY</td>
  517. <td STYLE="text-transform:lowercase">THIS EXPERIENCE REPRESENTS A TRANSCENDENTAL STEP IN THE SPATIAL
  518. HISTORY OF THE COUNTRY</td>
  519. <td STYLE="text-transform:lowercase">THIS EXPERIENCE REPRESENTS A MOVEMENT STEP IN THE SPACE HISTORY OF
  520. THE COUNTRY</td>
  521. </tr>
  522. <tr>
  523. <th colspan="6" style="text-align:left">Sample 3: S2UT_Aug performs best</th>
  524. </tr>
  525. <tr>
  526. <th></th>
  527. <th>
  528. <div id="src1_waveform1_3"></div>
  529. <button id="written_source__header" class="play-button-demo btn btn-primary"
  530. onclick="src1_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  531. Pause
  532. </button>
  533. <script> var src1_3 = WaveSurfer.create({ container: '#src1_waveform1_3', waveColor: 'violet', progressColor: 'purple' });
  534. src1_3.load('./audios/es-en/set2/source/4109_cv.flac'); </script>
  535. </th>
  536. <th>
  537. <div id="target_waveform1_3"></div>
  538. <button id="written_target__header" class="play-button-demo btn btn-primary"
  539. onclick="tgt1_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  540. Pause
  541. </button>
  542. <script> var tgt1_3 = WaveSurfer.create({ container: '#target_waveform1_3', waveColor: 'violet', progressColor: 'purple' });
  543. tgt1_3.load('./audios/es-en/set2/target/4109_cv.wav'); </script>
  544. </th>
  545. <th>
  546. <div id="s2ut_waveform1_3"></div>
  547. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  548. onclick="s2ut_lnad1_3.playPause()"> <i class="fa fa-play"></i> Play / <i
  549. class="fa fa-pause"></i>
  550. Pause </button>
  551. <script> var s2ut_lnad1_3 = WaveSurfer.create({ container: '#s2ut_waveform1_3', waveColor: 'violet', progressColor: 'purple' });
  552. s2ut_lnad1_3.load('./audios/es-en/set2/s2ut_lnd/4109_cv.wav'); </script>
  553. </th>
  554. <th>
  555. <div id="translatotron_waveform1_3"></div>
  556. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  557. onclick="s2ut_lr50_3.playPause()">
  558. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  559. </button>
  560. <script> var s2ut_lr50_3 = WaveSurfer.create({ container: '#translatotron_waveform1_3', waveColor: 'violet', progressColor: 'purple' });
  561. s2ut_lr50_3.load('./audios/es-en/set2/s2ut_lr50/4109_cv.wav'); </script>
  562. </th>
  563. <th>
  564. <div id="s2ttts_waveform1_3"></div>
  565. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  566. onclick="s2ut_asr3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  567. Pause
  568. </button>
  569. <script> var s2ut_asr3 = WaveSurfer.create({ container: '#s2ttts_waveform1_3', waveColor: 'violet', progressColor: 'purple' });
  570. s2ut_asr3.load('./audios/es-en/set2/s2ut_lnd_w_asr/4109_cv.wav'); </script>
  571. </th>
  572. </tr>
  573. <tr>
  574. <th> Reference: </th>
  575. <td> desde la perspectiva del balance físico químico y biológico está
  576. en una posición clave</td>
  577. <td STYLE="text-transform:lowercase"> THE PERSPECTIVE OF PHYSICAL CHEMICAL AND BIOLOGICAL BALANCE IT IS
  578. IN A KEY POSITION</td>
  579. </tr>
  580. <tr>
  581. <th> ASR: </th>
  582. <td> </td>
  583. <td> </td>
  584. <td STYLE="text-transform:lowercase">FROM A PHYSICAL CHEMICAL AND BIOLOGICAL BALANCE HE IS IN A KEY
  585. POSITION</td>
  586. <td STYLE="text-transform:lowercase">FROM A PHYSICAL PERSPECTIVE OF PHYSICAL CHEMICAL AND BIOLOGICAL
  587. POSITION</td>
  588. <td STYLE="text-transform:lowercase">FROM THE PERSPECTIVE OF PHYSICAL CHEMICAL AND BIOLOGICAL BALANCE IT
  589. IS IN A KEY POSITION</td>
  590. </tr>
  591. <tr>
  592. <th colspan="6" style="text-align:left">Sample 4: S2UT_Aug performs best</th>
  593. </tr>
  594. <tr>
  595. <th></th>
  596. <th>
  597. <div id="src1_waveform1_4"></div>
  598. <button id="written_source__header" class="play-button-demo btn btn-primary"
  599. onclick="src1_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  600. Pause
  601. </button>
  602. <script> var src1_4 = WaveSurfer.create({ container: '#src1_waveform1_4', waveColor: 'violet', progressColor: 'purple' });
  603. src1_4.load('./audios/es-en/set2/source/289_epst.flac'); </script>
  604. </th>
  605. <th>
  606. <div id="target_waveform1_4"></div>
  607. <button id="written_target__header" class="play-button-demo btn btn-primary"
  608. onclick="tgt1_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  609. Pause
  610. </button>
  611. <script> var tgt1_4 = WaveSurfer.create({ container: '#target_waveform1_4', waveColor: 'violet', progressColor: 'purple' });
  612. tgt1_4.load('./audios/es-en/set2/target/289_epst.wav'); </script>
  613. </th>
  614. <th>
  615. <div id="s2ut_waveform1_4"></div>
  616. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  617. onclick="s2ut_lnad1_4.playPause()">
  618. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  619. Pause </button>
  620. <script> var s2ut_lnad1_4 = WaveSurfer.create({ container: '#s2ut_waveform1_4', waveColor: 'violet', progressColor: 'purple' });
  621. s2ut_lnad1_4.load('./audios/es-en/set2/s2ut_lnd/289_epst.wav'); </script>
  622. </th>
  623. <th>
  624. <div id="translatotron_waveform1_4"></div>
  625. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  626. onclick="s2ut_lr50_4.playPause()">
  627. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  628. </button>
  629. <script> var s2ut_lr50_4 = WaveSurfer.create({ container: '#translatotron_waveform1_4', waveColor: 'violet', progressColor: 'purple' });
  630. s2ut_lr50_4.load('./audios/es-en/set2/s2ut_lr50/289_epst.wav'); </script>
  631. </th>
  632. <th>
  633. <div id="s2ttts_waveform1_4"></div>
  634. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  635. onclick="s2ut_asr4.playPause()">
  636. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  637. Pause
  638. </button>
  639. <script> var s2ut_asr4 = WaveSurfer.create({ container: '#s2ttts_waveform1_4', waveColor: 'violet', progressColor: 'purple' });
  640. s2ut_asr4.load('./audios/es-en/set2/s2ut_lnd_w_asr/289_epst.wav'); </script>
  641. </th>
  642. </tr>
  643. <tr>
  644. <th> Reference: </th>
  645. <td>desde un punto de vista presupuestario no parece adecuada la
  646. propuesta de financiación procedente de la comisión de
  647. desarrollo ya que este dinero no existe al</td>
  648. <td STYLE="text-transform:lowercase"> IN ANY CASE GIVEN THAT THE FINANCING OF THIS NEW COOPERATION
  649. INSTRUMENT MUST BE COMPATIBLE WITH THE
  650. TWO
  651. THOUSAND SEVEN
  652. TWENTY THIRTEEN FINANCIAL FRAMEWORK IT IS WORTH</td>
  653. </tr>
  654. <tr>
  655. <th> ASR: </th>
  656. <td> </td>
  657. <td> </td>
  658. <td STYLE="text-transform:lowercase">IN ANY CASE GIVEN THAT THE FUNDING OF THIS NEW CORPORATION
  659. INSTRUMENT MUST BE COMPATIBLE WITH THE
  660. TWO THOUSAND SEVEN
  661. TWENTY THIRTEEN FINANCIAL FRAMEWORK IT IS IMPORTANT</td>
  662. <td STYLE="text-transform:lowercase">IN ANY CASE SINCE THE FINANCING OF THIS NEW INSTRUMENT OF
  663. CORPORATION MUST COMPATIBLE WITH THE
  664. FINANCIAL FRAMEWORK
  665. FOR TWENTY THIRTEEN</td>
  666. <td STYLE="text-transform:lowercase">IN ANY CASE GIVEN THAT THE FINANCING OF THIS NEW CORPORATION
  667. INSTRUMENT MUST BE COMPATIBLE WITH THE
  668. TWO THOUSAND
  669. SEVEN TWENTY THIRTEEN FINANCIAL FRAMEWORK</td>
  670. </tr>
  671. </table>
  672. </div>
  673. <div id="EN-ES Comparison with Baselines" class="content-container">
  674. <script src="wavesurfer.js"></script>
  675. <div class="content-title">
  676. <font size="+5">English to Spanish</font>
  677. </div>
  678. <div class="content-subtitle">Comparison with Baselines
  679. </div>
  680. <p> We provide ground truth source and target audios with the corresponding reference text,
  681. as well as audio samples from three systems: <br>
  682. (1) <strong>S2UT+LNA-D</strong>: the proposed direct speeech-to-unit translation
  683. system initialized with wav2vec 2.0 encoder, unit mBART decoder and finetuned using LNA-D strategy<br>
  684. (2) <strong>Supervised S2UT</strong>: a baseline direct speech-to-unit translation system trained with
  685. source and target text as auxiliary task targets.
  686. <br>
  687. (3) <strong>S2T+TTS:</strong> a baseline cascaded system with a speech-to-text translation model initialized
  688. with wav2vec 2.0 encoder and a randomly initialized decoder, followed by a text-to-speech synthesis model. <br>
  689. Both (1) and (2) use an open sourced HiFi-GAN vocoder to convert units to waveforms.
  690. </p>
  691. <table border="0" class="inlineTable">
  692. <tr>
  693. <th></th>
  694. <th colspan="2">Ground truth</th>
  695. <th colspan="3">Predictions</th>
  696. </tr>
  697. <tr>
  698. <th></th>
  699. <th>Source (English)</th>
  700. <th>Target (Spanish)</th>
  701. <th>S2UT+LNA-D</th>
  702. <th>Supervised S2UT</th>
  703. <th>S2T+TTS</th>
  704. </tr>
  705. <tr>
  706. <th colspan="6" style="text-align:left">Sample 1: S2UT+LNAD performs the best.</th>
  707. </tr>
  708. <tr>
  709. <th></th>
  710. <th>
  711. <div id="src2_waveform2_1"></div>
  712. <button id="written_source__header" class="play-button-demo btn btn-primary"
  713. onclick="src2_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  714. Pause
  715. </button>
  716. <script> var src2_1 = WaveSurfer.create({ container: '#src2_waveform2_1', waveColor: 'violet', progressColor: 'purple' });
  717. src2_1.load('./audios/en-es/set1/source/1149_epst.wav'); </script>
  718. </th>
  719. <th>
  720. <div id="target_waveform2_1"></div>
  721. <button id="written_target__header" class="play-button-demo btn btn-primary"
  722. onclick="tgt2_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  723. Pause
  724. </button>
  725. <script> var tgt2_1 = WaveSurfer.create({ container: '#target_waveform2_1', waveColor: 'violet', progressColor: 'purple' });
  726. tgt2_1.load('./audios/en-es/set1/target/1149_epst.wav'); </script>
  727. </th>
  728. <th>
  729. <div id="s2ut_waveform2_1"></div>
  730. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  731. onclick="s2ut_lnad2_1.playPause()"> <i class="fa fa-play"></i> Play / <i
  732. class="fa fa-pause"></i>
  733. Pause </button>
  734. <script> var s2ut_lnad2_1 = WaveSurfer.create({ container: '#s2ut_waveform2_1', waveColor: 'violet', progressColor: 'purple' });
  735. s2ut_lnad2_1.load('./audios/en-es/set1/s2ut_lnd/1149_epst.wav'); </script>
  736. </th>
  737. <th>
  738. <div id="translatotron_waveform2_1"></div>
  739. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  740. onclick="s2ut_mt2_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  741. Pause
  742. </button>
  743. <script> var s2ut_mt2_1 = WaveSurfer.create({ container: '#translatotron_waveform2_1', waveColor: 'violet', progressColor: 'purple' });
  744. s2ut_mt2_1.load('./audios/en-es/set1/s2ut_mt/1149_epst.wav'); </script>
  745. </th>
  746. <th>
  747. <div id="s2ttts_waveform2_1"></div>
  748. <button id="written_cas2caded_header" class="play-button-demo btn btn-primary"
  749. onclick="cas2_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  750. Pause
  751. </button>
  752. <script> var cas2_1 = WaveSurfer.create({ container: '#s2ttts_waveform2_1', waveColor: 'violet', progressColor: 'purple' });
  753. cas2_1.load('./audios/en-es/set1/s2t_tts/1149_epst.wav'); </script>
  754. </th>
  755. </tr>
  756. <tr>
  757. <th> Reference: </th>
  758. <td>this should also be an important part of our approach to the twenty
  759. twelve budget</td>
  760. <td>esto también debería ser una parte importante de nuestro enfoque
  761. del
  762. presupuesto dos mil doce
  763. </td>
  764. </tr>
  765. <tr>
  766. <th> ASR: </th>
  767. <td> </td>
  768. <td> </td>
  769. <td>esto también debería ser una parte importante de nuestro enfoque al
  770. presupuesto dos mil doce </td>
  771. <td>también debería ser una parte importante de nuestro enfoque al
  772. presupuesto dos mil doce</td>
  773. <td>esto también debería ser una parte importante de nuestro enfoque al
  774. presupuesto de dos mildos mil
  775. dos mil doce</td>
  776. </tr>
  777. <tr>
  778. <th colspan="6" style="text-align:left">Sample 2: S2UT+LNAD performs the best.</th>
  779. </tr>
  780. <tr>
  781. <th></th>
  782. <th>
  783. <div id="src2_waveform2_4"></div>
  784. <button id="written_source__header" class="play-button-demo btn btn-primary"
  785. onclick="src2_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  786. Pause
  787. </button>
  788. <script> var src2_4 = WaveSurfer.create({ container: '#src2_waveform2_4', waveColor: 'violet', progressColor: 'purple' });
  789. src2_4.load('./audios/en-es/set1/source/890_epst.wav'); </script>
  790. </th>
  791. <th>
  792. <div id="target_waveform2_4"></div>
  793. <button id="written_target__header" class="play-button-demo btn btn-primary"
  794. onclick="tgt2_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  795. Pause
  796. </button>
  797. <script> var tgt2_4 = WaveSurfer.create({ container: '#target_waveform2_4', waveColor: 'violet', progressColor: 'purple' });
  798. tgt2_4.load('./audios/en-es/set1/target/890_epst.wav'); </script>
  799. </th>
  800. <th>
  801. <div id="s2ut_waveform2_4"></div>
  802. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  803. onclick="s2ut_lnad2_4.playPause()">
  804. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  805. Pause </button>
  806. <script> var s2ut_lnad2_4 = WaveSurfer.create({ container: '#s2ut_waveform2_4', waveColor: 'violet', progressColor: 'purple' });
  807. s2ut_lnad2_4.load('./audios/en-es/set1/s2ut_lnd/890_epst.wav'); </script>
  808. </th>
  809. <th>
  810. <div id="translatotron_waveform2_4"></div>
  811. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  812. onclick="s2ut_mt2_4.playPause()">
  813. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  814. </button>
  815. <script> var s2ut_mt2_4 = WaveSurfer.create({ container: '#translatotron_waveform2_4', waveColor: 'violet', progressColor: 'purple' });
  816. s2ut_mt2_4.load('./audios/en-es/set1/s2ut_mt/890_epst.wav'); </script>
  817. </th>
  818. <th>
  819. <div id="s2ttts_waveform2_4"></div>
  820. <button id="written_cas2caded_header" class="play-button-demo btn btn-primary"
  821. onclick="cas2_4.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  822. Pause
  823. </button>
  824. <script> var cas2_4 = WaveSurfer.create({ container: '#s2ttts_waveform2_4', waveColor: 'violet', progressColor: 'purple' });
  825. cas2_4.load('./audios/en-es/set1/s2t_tts/890_epst.wav'); </script>
  826. </th>
  827. </tr>
  828. <tr>
  829. <th> Reference: </th>
  830. <td>information encourages citizens interest in public matters and
  831. their participation</td>
  832. <td>la información fomenta el interés de los ciudadanos por los
  833. asuntos
  834. públicos y su participación</td>
  835. </tr>
  836. <tr>
  837. <th> ASR: </th>
  838. <td> </td>
  839. <td> </td>
  840. <td>la información fomenta el interés de los ciudadanos en asuntos
  841. públicos y su participación</td>
  842. <td>la información y el interés de los ciudadanos alientan los
  843. intereses de las cuestiones públicas y su
  844. participación
  845. </td>
  846. <td>la información alienta el interés de los ciudadanos en asuntos
  847. públicos y en su participación</td>
  848. </tr>
  849. <tr>
  850. <th colspan="6" style="text-align:left">Sample 3: S2UT+LNAD performs the best.</th>
  851. </tr>
  852. <tr>
  853. <th></th>
  854. <th>
  855. <div id="src2_waveform2_2"></div>
  856. <button id="written_source__header" class="play-button-demo btn btn-primary"
  857. onclick="src2_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  858. Pause
  859. </button>
  860. <script> var src2_2 = WaveSurfer.create({ container: '#src2_waveform2_2', waveColor: 'violet', progressColor: 'purple' });
  861. src2_2.load('./audios/en-es/set1/source/476_epst.wav'); </script>
  862. </th>
  863. <th>
  864. <div id="target_waveform2_2"></div>
  865. <button id="written_target__header" class="play-button-demo btn btn-primary"
  866. onclick="tgt2_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  867. Pause
  868. </button>
  869. <script> var tgt2_2 = WaveSurfer.create({ container: '#target_waveform2_2', waveColor: 'violet', progressColor: 'purple' });
  870. tgt2_2.load('./audios/en-es/set1/target/476_epst.wav'); </script>
  871. </th>
  872. <th>
  873. <div id="s2ut_waveform2_2"></div>
  874. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  875. onclick="s2ut_lnad2_2.playPause()"> <i class="fa fa-play"></i> Play / <i
  876. class="fa fa-pause"></i>
  877. Pause </button>
  878. <script> var s2ut_lnad2_2 = WaveSurfer.create({ container: '#s2ut_waveform2_2', waveColor: 'violet', progressColor: 'purple' });
  879. s2ut_lnad2_2.load('./audios/en-es/set1/s2ut_lnd/476_epst.wav'); </script>
  880. </th>
  881. <th>
  882. <div id="translatotron_waveform2_2"></div>
  883. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  884. onclick="s2ut_mt2_2.playPause()">
  885. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  886. </button>
  887. <script> var s2ut_mt2_2 = WaveSurfer.create({ container: '#translatotron_waveform2_2', waveColor: 'violet', progressColor: 'purple' });
  888. s2ut_mt2_2.load('./audios/en-es/set1/s2ut_mt/476_epst.wav'); </script>
  889. </th>
  890. <th>
  891. <div id="s2ttts_waveform2_2"></div>
  892. <button id="written_cas2caded_header" class="play-button-demo btn btn-primary"
  893. onclick="cas2_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  894. Pause
  895. </button>
  896. <script> var cas2_2 = WaveSurfer.create({ container: '#s2ttts_waveform2_2', waveColor: 'violet', progressColor: 'purple' });
  897. cas2_2.load('./audios/en-es/set1/s2t_tts/476_epst.wav'); </script>
  898. </th>
  899. </tr>
  900. <tr>
  901. <th> Reference: </th>
  902. <td>his family who are my constituents are convinced of his innocence
  903. </td>
  904. <td>su familia que son mis electores está convencida de su inocencia
  905. </td>
  906. </tr>
  907. <tr>
  908. <th> ASR: </th>
  909. <td></td>
  910. <td></td>
  911. <td>su familia que son mis electores está convencida de su inocencia
  912. </td>
  913. <td>su familia que son mí circunscripciones están convencidas de estos
  914. inocentes</td>
  915. <td>su familia que son mis electores están convencidos de su inocencia
  916. </td>
  917. </tr>
  918. <tr>
  919. <th colspan="6" style="text-align:left">Sample 4: All systems make errors.</th>
  920. </tr>
  921. <tr>
  922. <th></th>
  923. <th>
  924. <div id="src2_waveform2_3"></div>
  925. <button id="written_source__header" class="play-button-demo btn btn-primary"
  926. onclick="src2_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  927. Pause
  928. </button>
  929. <script> var src2_3 = WaveSurfer.create({ container: '#src2_waveform2_3', waveColor: 'violet', progressColor: 'purple' });
  930. src2_3.load('./audios/en-es/set1/source/651_epst.wav'); </script>
  931. </th>
  932. <th>
  933. <div id="target_waveform2_3"></div>
  934. <button id="written_target__header" class="play-button-demo btn btn-primary"
  935. onclick="tgt2_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  936. Pause
  937. </button>
  938. <script> var tgt2_3 = WaveSurfer.create({ container: '#target_waveform2_3', waveColor: 'violet', progressColor: 'purple' });
  939. tgt2_3.load('./audios/en-es/set1/target/651_epst.wav'); </script>
  940. </th>
  941. <th>
  942. <div id="s2ut_waveform2_3"></div>
  943. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  944. onclick="s2ut_lnad2_3.playPause()">
  945. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  946. Pause </button>
  947. <script> var s2ut_lnad2_3 = WaveSurfer.create({ container: '#s2ut_waveform2_3', waveColor: 'violet', progressColor: 'purple' });
  948. s2ut_lnad2_3.load('./audios/en-es/set1/s2ut_lnd/651_epst.wav'); </script>
  949. </th>
  950. <th>
  951. <div id="translatotron_waveform2_3"></div>
  952. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  953. onclick="s2ut_mt2_3.playPause()">
  954. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  955. </button>
  956. <script> var s2ut_mt2_3 = WaveSurfer.create({ container: '#translatotron_waveform2_3', waveColor: 'violet', progressColor: 'purple' });
  957. s2ut_mt2_3.load('./audios/en-es/set1/s2ut_mt/651_epst.wav'); </script>
  958. </th>
  959. <th>
  960. <div id="s2ttts_waveform2_3"></div>
  961. <button id="written_cas2caded_header" class="play-button-demo btn btn-primary"
  962. onclick="cas2_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  963. Pause
  964. </button>
  965. <script> var cas2_3 = WaveSurfer.create({ container: '#s2ttts_waveform2_3', waveColor: 'violet', progressColor: 'purple' });
  966. cas2_3.load('./audios/en-es/set1/s2t_tts/651_epst.wav'); </script>
  967. </th>
  968. </tr>
  969. <tr>
  970. <th> Reference: </th>
  971. <td>of the directive on all taxes including social security
  972. contributions the automatic exchange of information and improved
  973. cooperation between the member states in matters of taxation</td>
  974. <td> de la directiva a todos los impuestos incluidas las
  975. contribuciones a
  976. la seguridad social el intercambio automático de
  977. información y la mejora de la cooperación fiscal entre los estados miembros</td>
  978. </tr>
  979. <tr>
  980. <th> ASR: </th>
  981. <td> </td>
  982. <td> </td>
  983. <td>de la directiva a todos los impuestos incluidas las contribuciones
  984. a la seguridad social el
  985. intercambio automático
  986. de información y la mejor cooperación entre los estados miembros en las cuestiones de impuestos</td>
  987. <td>de la directiva a todos los impuestos impluyendo las contribuciones
  988. de seguridad social el
  989. intercambio automático de
  990. la información y mejorar la cooperación entre los estados miembros y las cuestiones de impuestos
  991. </td>
  992. <td>de la directiva para todos los impuestos incluidos las
  993. contribuciones de seguridad social el
  994. intercambio automático
  995. de información y la mejor cooperación entre los estados miembros en la cuestión de la fiscalidad
  996. </td>
  997. </tr>
  998. </table>
  999. </div>
  1000. <div id="EN-ES Different Data Setups" class="content-container">
  1001. <script src="wavesurfer.js"></script>
  1002. <div class="content-title">
  1003. <font size="+5">English To Spanish</font>
  1004. </div>
  1005. <div class="content-subtitle">Different Data Setups
  1006. </div>
  1007. <p> We provide ground truth source and target audios with the corresponding reference text,
  1008. as well as audio samples from three systems. All the three models are initialized with wav2vec 2.0 encoder,
  1009. unit
  1010. mBART decoder and finetuned using LNA-D strategy but use different datasets for finetuning: <br>
  1011. (1) <strong>S2UT_Base</strong>: finetuned on the combination of Europarl-ST, MuST-C datasets.
  1012. <br>
  1013. (2) <strong>S2UT_LR</strong>: finetuned on low resource setup with 50hr of data sampled from the combination
  1014. of Europarl-ST, MuST-C datasets
  1015. <br>
  1016. (3) <strong>S2UT_Aug:</strong> finetuned on the combination of Europarl-ST, MuST-C datasets plus the ASR
  1017. data. <br>
  1018. All models use an open sourced HiFi-GAN vocoder to convert units to waveforms.
  1019. </p>
  1020. <table border="0" class="inlineTable">
  1021. <tr>
  1022. <th></th>
  1023. <th colspan="2">Ground truth</th>
  1024. <th colspan="3">Predictions</th>
  1025. </tr>
  1026. <tr>
  1027. <th></th>
  1028. <th>Source (English)</th>
  1029. <th>Target (Spanish)</th>
  1030. <th>S2UT_Base</th>
  1031. <th>S2UT_LR</th>
  1032. <th>S2UT_Aug</th>
  1033. </tr>
  1034. <tr>
  1035. <th colspan="6" style="text-align:left">Sample 1: All systems do well.</th>
  1036. </tr>
  1037. <tr>
  1038. <th></th>
  1039. <th>
  1040. <div id="src3_waveform3_1"></div>
  1041. <button id="written_source__header" class="play-button-demo btn btn-primary"
  1042. onclick="src3_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1043. Pause
  1044. </button>
  1045. <script> var src3_1 = WaveSurfer.create({ container: '#src3_waveform3_1', waveColor: 'violet', progressColor: 'purple' });
  1046. src3_1.load('./audios/en-es/set2/source/37_epst.wav'); </script>
  1047. </th>
  1048. <th>
  1049. <div id="target_waveform3_1"></div>
  1050. <button id="written_target__header" class="play-button-demo btn btn-primary"
  1051. onclick="tgt3_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1052. Pause
  1053. </button>
  1054. <script> var tgt3_1 = WaveSurfer.create({ container: '#target_waveform3_1', waveColor: 'violet', progressColor: 'purple' });
  1055. tgt3_1.load('./audios/en-es/set2/target/37_epst.wav'); </script>
  1056. </th>
  1057. <th>
  1058. <div id="s2ut_waveform3_1"></div>
  1059. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  1060. onclick="s2ut_lnad3_1.playPause()"> <i class="fa fa-play"></i> Play / <i
  1061. class="fa fa-pause"></i>
  1062. Pause </button>
  1063. <script> var s2ut_lnad3_1 = WaveSurfer.create({ container: '#s2ut_waveform3_1', waveColor: 'violet', progressColor: 'purple' });
  1064. s2ut_lnad3_1.load('./audios/en-es/set2/s2ut_lnd/37_epst.wav'); </script>
  1065. </th>
  1066. <th>
  1067. <div id="translatotron_waveform3_1"></div>
  1068. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  1069. onclick="s2ut_lr50_2_1.playPause()"> <i class="fa fa-play"></i> Play / <i
  1070. class="fa fa-pause"></i>
  1071. Pause
  1072. </button>
  1073. <script> var s2ut_lr50_2_1 = WaveSurfer.create({ container: '#translatotron_waveform3_1', waveColor: 'violet', progressColor: 'purple' });
  1074. s2ut_lr50_2_1.load('./audios/en-es/set2/s2ut_lr50/37_epst.wav'); </script>
  1075. </th>
  1076. <th>
  1077. <div id="s2ttts_waveform3_1"></div>
  1078. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  1079. onclick="s2ut_asr3_1.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1080. Pause
  1081. </button>
  1082. <script> var s2ut_asr3_1 = WaveSurfer.create({ container: '#s2ttts_waveform3_1', waveColor: 'violet', progressColor: 'purple' });
  1083. s2ut_asr3_1.load('./audios/en-es/set2/s2ut_lnd_w_asr/37_epst.wav'); </script>
  1084. </th>
  1085. </tr>
  1086. <tr>
  1087. <th> Reference: </th>
  1088. <td>we want to see energy poverty as a part of this debate</td>
  1089. <td>queremos ver la pobreza energética como parte de este debate
  1090. </td>
  1091. </tr>
  1092. <tr>
  1093. <th> ASR: </th>
  1094. <td> </td>
  1095. <td> </td>
  1096. <td>queremos ver la pobreza energética como parte de este deate</td>
  1097. <td>queremos ver la pobreza energética como parte de este date</td>
  1098. <td>queremos ver la pobreza energética como parte de este deate</td>
  1099. </tr>
  1100. <tr>
  1101. <th colspan="6" style="text-align:left">Sample 2: S2UT_LR has errors but S2UT_Base and S2UT_Aug got
  1102. it
  1103. right.</th>
  1104. </tr>
  1105. <tr>
  1106. <th></th>
  1107. <th>
  1108. <div id="src3_waveform3_2"></div>
  1109. <button id="written_source__header" class="play-button-demo btn btn-primary"
  1110. onclick="src3_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1111. Pause
  1112. </button>
  1113. <script> var src3_2 = WaveSurfer.create({ container: '#src3_waveform3_2', waveColor: 'violet', progressColor: 'purple' });
  1114. src3_2.load('./audios/en-es/set2/source/923_epst.wav'); </script>
  1115. </th>
  1116. <th>
  1117. <div id="target_waveform3_2"></div>
  1118. <button id="written_target__header" class="play-button-demo btn btn-primary"
  1119. onclick="tgt3_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1120. Pause
  1121. </button>
  1122. <script> var tgt3_2 = WaveSurfer.create({ container: '#target_waveform3_2', waveColor: 'violet', progressColor: 'purple' });
  1123. tgt3_2.load('./audios/en-es/set2/target/923_epst.wav'); </script>
  1124. </th>
  1125. <th>
  1126. <div id="s2ut_waveform3_2"></div>
  1127. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  1128. onclick="s2ut_lnad3_2.playPause()"> <i class="fa fa-play"></i> Play / <i
  1129. class="fa fa-pause"></i>
  1130. Pause </button>
  1131. <script> var s2ut_lnad3_2 = WaveSurfer.create({ container: '#s2ut_waveform3_2', waveColor: 'violet', progressColor: 'purple' });
  1132. s2ut_lnad3_2.load('./audios/en-es/set2/s2ut_lnd/923_epst.wav'); </script>
  1133. </th>
  1134. <th>
  1135. <div id="translatotron_waveform3_2"></div>
  1136. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  1137. onclick="s2ut_lr50_2_2.playPause()">
  1138. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  1139. </button>
  1140. <script> var s2ut_lr50_2_2 = WaveSurfer.create({ container: '#translatotron_waveform3_2', waveColor: 'violet', progressColor: 'purple' });
  1141. s2ut_lr50_2_2.load('./audios/en-es/set2/s2ut_lr50/923_epst.wav'); </script>
  1142. </th>
  1143. <th>
  1144. <div id="s2ttts_waveform3_2"></div>
  1145. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  1146. onclick="s2ut_asr3_2.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1147. Pause
  1148. </button>
  1149. <script> var s2ut_asr3_2 = WaveSurfer.create({ container: '#s2ttts_waveform3_2', waveColor: 'violet', progressColor: 'purple' });
  1150. s2ut_asr3_2.load('./audios/en-es/set2/s2ut_lnd_w_asr/923_epst.wav'); </script>
  1151. </th>
  1152. </tr>
  1153. <tr>
  1154. <th> Reference: </th>
  1155. <td>in my view one of the most important elements is the follow up of
  1156. legislative initiative requests from parliament</td>
  1157. <td>en mi opinión uno de los elementos más importantes es el
  1158. seguimiento de
  1159. las solicitudes de iniciativa legislativa del
  1160. parlamento</td>
  1161. </tr>
  1162. <tr>
  1163. <th> ASR: </th>
  1164. <td> </td>
  1165. <td> </td>
  1166. <td>n mi opinión uno de los elementos más importantes es el
  1167. seguimiento de las peticiones de la
  1168. iniciativa legislativa
  1169. por parte del pagamento</td>
  1170. <td>en mi opinión uno de los elementos más importantes es el
  1171. seguimiento de las emiendas de iniciativas
  1172. legislativas de
  1173. ley</td>
  1174. <td>en mi opinión uno de los elementos más importantes es el
  1175. seguimiento de las solicitudes de
  1176. iniciativa legislativa
  1177. del pagamento</td>
  1178. </tr>
  1179. <tr>
  1180. <th colspan="6" style="text-align:left">Sample 3: S2UT_Aug performs the best</th>
  1181. </tr>
  1182. <tr>
  1183. <th></th>
  1184. <th>
  1185. <div id="src3_waveform3_3"></div>
  1186. <button id="written_source__header" class="play-button-demo btn btn-primary"
  1187. onclick="src3_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1188. Pause
  1189. </button>
  1190. <script> var src3_3 = WaveSurfer.create({ container: '#src3_waveform3_3', waveColor: 'violet', progressColor: 'purple' });
  1191. src3_3.load('./audios/en-es/set2/source/970_epst.wav'); </script>
  1192. </th>
  1193. <th>
  1194. <div id="target_waveform3_3"></div>
  1195. <button id="written_target__header" class="play-button-demo btn btn-primary"
  1196. onclick="tgt3_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1197. Pause
  1198. </button>
  1199. <script> var tgt3_3 = WaveSurfer.create({ container: '#target_waveform3_3', waveColor: 'violet', progressColor: 'purple' });
  1200. tgt3_3.load('./audios/en-es/set2/target/970_epst.wav'); </script>
  1201. </th>
  1202. <th>
  1203. <div id="s2ut_waveform3_3"></div>
  1204. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  1205. onclick="s2ut_lnad3_3.playPause()"> <i class="fa fa-play"></i> Play / <i
  1206. class="fa fa-pause"></i>
  1207. Pause </button>
  1208. <script> var s2ut_lnad3_3 = WaveSurfer.create({ container: '#s2ut_waveform3_3', waveColor: 'violet', progressColor: 'purple' });
  1209. s2ut_lnad3_3.load('./audios/en-es/set2/s2ut_lnd/970_epst.wav'); </script>
  1210. </th>
  1211. <th>
  1212. <div id="translatotron_waveform3_3"></div>
  1213. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  1214. onclick="s2ut_lr50_2_3.playPause()">
  1215. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  1216. </button>
  1217. <script> var s2ut_lr50_2_3 = WaveSurfer.create({ container: '#translatotron_waveform3_3', waveColor: 'violet', progressColor: 'purple' });
  1218. s2ut_lr50_2_3.load('./audios/en-es/set2/s2ut_lr50/970_epst.wav'); </script>
  1219. </th>
  1220. <th>
  1221. <div id="s2ttts_waveform3_3"></div>
  1222. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  1223. onclick="s2ut_asr3_3.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1224. Pause
  1225. </button>
  1226. <script> var s2ut_asr3_3 = WaveSurfer.create({ container: '#s2ttts_waveform3_3', waveColor: 'violet', progressColor: 'purple' });
  1227. s2ut_asr3_3.load('./audios/en-es/set2/s2ut_lnd_w_asr/970_epst.wav'); </script>
  1228. </th>
  1229. </tr>
  1230. <tr>
  1231. <th> Reference: </th>
  1232. <td>we must find an open and constructive procedure on the next
  1233. financial framework</td>
  1234. <td> debemos encontrar un procedimiento abierto y constructivo en el
  1235. próximo marco financiero</td>
  1236. </tr>
  1237. <tr>
  1238. <th> ASR: </th>
  1239. <td> </td>
  1240. <td> </td>
  1241. <td>debemos encontrar un procedimiento abierto y constructivo sobre el
  1242. próximo marco financiero</td>
  1243. <td>debemos encontrar un procedimiento abierto y constructivo en el
  1244. sistema financiero financiero
  1245. financiero financiero
  1246. </td>
  1247. <td>debemos encontrar un procedimiento abierto y constructivo en el
  1248. próximo marco financiero</td>
  1249. </tr>
  1250. <tr>
  1251. <th colspan="6" style="text-align:left">Sample 4: All systems make errors</th>
  1252. </tr>
  1253. <tr>
  1254. <th></th>
  1255. <th>
  1256. <div id="src3_waveform3_5"></div>
  1257. <button id="written_source__header" class="play-button-demo btn btn-primary"
  1258. onclick="src3_5.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1259. Pause
  1260. </button>
  1261. <script> var src3_5 = WaveSurfer.create({ container: '#src3_waveform3_5', waveColor: 'violet', progressColor: 'purple' });
  1262. src3_5.load('./audios/en-es/set2/source/651_epst.wav'); </script>
  1263. </th>
  1264. <th>
  1265. <div id="target_waveform3_5"></div>
  1266. <button id="written_target__header" class="play-button-demo btn btn-primary"
  1267. onclick="tgt3_5.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1268. Pause
  1269. </button>
  1270. <script> var tgt3_5 = WaveSurfer.create({ container: '#target_waveform3_5', waveColor: 'violet', progressColor: 'purple' });
  1271. tgt3_5.load('./audios/en-es/set2/target/651_epst.wav'); </script>
  1272. </th>
  1273. <th>
  1274. <div id="s2ut_waveform3_5"></div>
  1275. <button id="written_s2ut__header" class="play-button-demo btn btn-primary"
  1276. onclick="s2ut_lnad3_5.playPause()"> <i class="fa fa-play"></i> Play / <i
  1277. class="fa fa-pause"></i>
  1278. Pause </button>
  1279. <script> var s2ut_lnad3_5 = WaveSurfer.create({ container: '#s2ut_waveform3_5', waveColor: 'violet', progressColor: 'purple' });
  1280. s2ut_lnad3_5.load('./audios/en-es/set2/s2ut_lnd/651_epst.wav'); </script>
  1281. </th>
  1282. <th>
  1283. <div id="translatotron_waveform3_5"></div>
  1284. <button id="written_translatotron_header" class="play-button-demo btn btn-primary"
  1285. onclick="s2ut_lr50_2_5.playPause()">
  1286. <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i> Pause
  1287. </button>
  1288. <script> var s2ut_lr50_2_5 = WaveSurfer.create({ container: '#translatotron_waveform3_5', waveColor: 'violet', progressColor: 'purple' });
  1289. s2ut_lr50_2_5.load('./audios/en-es/set2/s2ut_lr50/651_epst.wav'); </script>
  1290. </th>
  1291. <th>
  1292. <div id="s2ttts_waveform3_5"></div>
  1293. <button id="written_cascaded_header" class="play-button-demo btn btn-primary"
  1294. onclick="s2ut_asr3_5.playPause()"> <i class="fa fa-play"></i> Play / <i class="fa fa-pause"></i>
  1295. Pause
  1296. </button>
  1297. <script> var s2ut_asr3_5 = WaveSurfer.create({ container: '#s2ttts_waveform3_5', waveColor: 'violet', progressColor: 'purple' });
  1298. s2ut_asr3_5.load('./audios/en-es/set2/s2ut_lnd_w_asr/651_epst.wav'); </script>
  1299. </th>
  1300. </tr>
  1301. <tr>
  1302. <th> Reference: </th>
  1303. <td>of the directive on all taxes including social security
  1304. contributions the automatic exchange of information and improved
  1305. cooperation between the member states in matters of taxation</td>
  1306. <td>de la directiva a todos los impuestos incluidas las contribuciones
  1307. a la
  1308. seguridad social el intercambio automático de
  1309. información y la mejora de la cooperación fiscal entre los estados miembros</td>
  1310. </tr>
  1311. <tr>
  1312. <th> ASR: </th>
  1313. <td> </td>
  1314. <td> </td>
  1315. <td>de la directiva a todos los impuestos incluidas las contribuciones
  1316. a la seguridad social el
  1317. intercambio automático
  1318. de información y la mejor cooperación entre los estados miembros en las cuestiones de impuestos</td>
  1319. <td>la directiva sobre el impuesto de todos los contribuyentes
  1320. inpluyendo las contribuciones sociales la
  1321. introducción
  1322. automática y mejorada de los estados miembros y mejorar la cooperación entre los estados miembros
  1323. </td>
  1324. <td>de la directiva a todos los impuestos incluidas las contribuciones
  1325. a la seguridad social el
  1326. intercambio automático
  1327. de información y la mejor cooperación entre los estados miembros en materia de impuestos</td>
  1328. </tr>
  1329. </table>
  1330. </div>
  1331. <div class="content-container">
  1332. Template based on <a style="color:rgb(22, 38, 67)" href="https://speechbot.github.io/"> Textless NLP</a> and <a
  1333. style="color:rgb(22, 38, 67)" href="https://daps.cs.princeton.edu/projects/HiFi-GAN/index.php"> HiFi-GAN</a>
  1334. pages.
  1335. </div>
  1336. </body>
  1337. </html>