Matches in SemOpenAlex for { <https://semopenalex.org/work/W2973013313> ?p ?o ?g. }
- W2973013313 abstract "Unlike human speakers, typical text-to-speech (TTS) systems are unable to produce multiple distinct renditions of a given sentence. This has previously been addressed by adding explicit external control. In contrast, generative models are able to capture a distribution over multiple renditions and thus produce varied renditions using sampling. Typical neural TTS models learn the average of the data because they minimise mean squared error. In the context of prosody, taking the average produces flatter, more boring speech: an average prosody. A generative model that can synthesise multiple prosodies will, by design, not model average prosody. We use variational autoencoders (VAEs) which explicitly place the most average data close to the mean of the Gaussian prior. We propose that by moving towards the tails of the prior distribution, the model will transition towards generating more idiosyncratic, varied renditions. Focusing here on intonation, we investigate the trade-off between naturalness and intonation variation and find that typical acoustic models can either be natural, or varied, but not both. However, sampling from the tails of the VAE prior produces much more varied intonation than the traditional approaches, whilst maintaining the same level of naturalness." @default.
- W2973013313 created "2019-09-19" @default.
- W2973013313 creator A5041839891 @default.
- W2973013313 creator A5049904155 @default.
- W2973013313 creator A5062516688 @default.
- W2973013313 date "2019-09-20" @default.
- W2973013313 modified "2023-09-26" @default.
- W2973013313 title "Using generative modelling to produce varied intonation for speech synthesis" @default.
- W2973013313 cites W1522301498 @default.
- W2973013313 cites W1579853615 @default.
- W2973013313 cites W1783473872 @default.
- W2973013313 cites W1990505856 @default.
- W2973013313 cites W2043278166 @default.
- W2973013313 cites W2067752346 @default.
- W2973013313 cites W2078597717 @default.
- W2973013313 cites W2102003408 @default.
- W2973013313 cites W2112021726 @default.
- W2973013313 cites W2116046013 @default.
- W2973013313 cites W2119929864 @default.
- W2973013313 cites W2127531292 @default.
- W2973013313 cites W2129142580 @default.
- W2973013313 cites W2154920538 @default.
- W2973013313 cites W2157331557 @default.
- W2973013313 cites W2188365844 @default.
- W2973013313 cites W2188827208 @default.
- W2973013313 cites W2330979245 @default.
- W2973013313 cites W2396043161 @default.
- W2973013313 cites W2406225780 @default.
- W2973013313 cites W2471520273 @default.
- W2973013313 cites W2519430864 @default.
- W2973013313 cites W2598638573 @default.
- W2973013313 cites W2608207374 @default.
- W2973013313 cites W2608338293 @default.
- W2973013313 cites W2626778328 @default.
- W2973013313 cites W2753738274 @default.
- W2973013313 cites W2759925408 @default.
- W2973013313 cites W2794490148 @default.
- W2973013313 cites W28194048 @default.
- W2973013313 cites W2884607399 @default.
- W2973013313 cites W2888976508 @default.
- W2973013313 cites W2889129899 @default.
- W2973013313 cites W2899771611 @default.
- W2973013313 cites W2945544731 @default.
- W2973013313 cites W2951004968 @default.
- W2973013313 cites W2951051561 @default.
- W2973013313 cites W2952269766 @default.
- W2973013313 cites W2962691331 @default.
- W2973013313 cites W2962738009 @default.
- W2973013313 cites W2963568578 @default.
- W2973013313 cites W2963799213 @default.
- W2973013313 cites W2964301388 @default.
- W2973013313 doi "https://doi.org/10.21437/ssw.2019-43" @default.
- W2973013313 hasPublicationYear "2019" @default.
- W2973013313 type Work @default.
- W2973013313 sameAs 2973013313 @default.
- W2973013313 citedByCount "18" @default.
- W2973013313 countsByYear W29730133132020 @default.
- W2973013313 countsByYear W29730133132021 @default.
- W2973013313 countsByYear W29730133132022 @default.
- W2973013313 countsByYear W29730133132023 @default.
- W2973013313 crossrefType "proceedings-article" @default.
- W2973013313 hasAuthorship W2973013313A5041839891 @default.
- W2973013313 hasAuthorship W2973013313A5049904155 @default.
- W2973013313 hasAuthorship W2973013313A5062516688 @default.
- W2973013313 hasBestOaLocation W29730133132 @default.
- W2973013313 hasConcept C121332964 @default.
- W2973013313 hasConcept C127413603 @default.
- W2973013313 hasConcept C134537474 @default.
- W2973013313 hasConcept C138885662 @default.
- W2973013313 hasConcept C14999030 @default.
- W2973013313 hasConcept C151730666 @default.
- W2973013313 hasConcept C154945302 @default.
- W2973013313 hasConcept C167966045 @default.
- W2973013313 hasConcept C176217482 @default.
- W2973013313 hasConcept C204321447 @default.
- W2973013313 hasConcept C21547014 @default.
- W2973013313 hasConcept C2777530160 @default.
- W2973013313 hasConcept C2779343474 @default.
- W2973013313 hasConcept C2781045179 @default.
- W2973013313 hasConcept C28490314 @default.
- W2973013313 hasConcept C39890363 @default.
- W2973013313 hasConcept C41008148 @default.
- W2973013313 hasConcept C41895202 @default.
- W2973013313 hasConcept C542774811 @default.
- W2973013313 hasConcept C61224824 @default.
- W2973013313 hasConcept C62520636 @default.
- W2973013313 hasConcept C62897895 @default.
- W2973013313 hasConcept C86803240 @default.
- W2973013313 hasConceptScore W2973013313C121332964 @default.
- W2973013313 hasConceptScore W2973013313C127413603 @default.
- W2973013313 hasConceptScore W2973013313C134537474 @default.
- W2973013313 hasConceptScore W2973013313C138885662 @default.
- W2973013313 hasConceptScore W2973013313C14999030 @default.
- W2973013313 hasConceptScore W2973013313C151730666 @default.
- W2973013313 hasConceptScore W2973013313C154945302 @default.
- W2973013313 hasConceptScore W2973013313C167966045 @default.
- W2973013313 hasConceptScore W2973013313C176217482 @default.
- W2973013313 hasConceptScore W2973013313C204321447 @default.
- W2973013313 hasConceptScore W2973013313C21547014 @default.
- W2973013313 hasConceptScore W2973013313C2777530160 @default.