In this paper,an application independent model isproposed for the capacity evaluation of enterprise voiceapplication *** on this model,a distributedarchitecture tbr thc speech sub-system of an enterprisevoice applicat...
详细信息
In this paper,an application independent model isproposed for the capacity evaluation of enterprise voiceapplication *** on this model,a distributedarchitecture tbr thc speech sub-system of an enterprisevoice application is proposed and analyzed with queuingtheory.
CHinese spontaneous dialogue and conversation corpus(CADCC)has been collccted for the phonetic rescarch andspeech recognition and speech *** generalinformation and the orthographic and prosodic and segmentalannotation...
详细信息
CHinese spontaneous dialogue and conversation corpus(CADCC)has been collccted for the phonetic rescarch andspeech recognition and speech *** generalinformation and the orthographic and prosodic and segmentalannotation on this corpus are described in this paper.
Environmental robustness and speaker independence arefocuses of current speech recognition researh. Channeland speaker adaptation methods do the best job when theadaptation is done towards a normalized acoustic *** im...
详细信息
Environmental robustness and speaker independence arefocuses of current speech recognition researh. Channeland speaker adaptation methods do the best job when theadaptation is done towards a normalized acoustic *** improved compensation method based on SpectrumSubtraction(SS) is proposed to achieve a betterperformance in estimating the source speech signal fromthe acquired speech signal in the noisy environment.
A voice conversion algorithm based on acousticfeature transformation is *** synthsis is widely used *** such asystem,voice conversion is much more challengingthan a paramter bases synthesizer,such as *** main idea is ...
详细信息
A voice conversion algorithm based on acousticfeature transformation is *** synthsis is widely used *** such asystem,voice conversion is much more challengingthan a paramter bases synthesizer,such as *** main idea is to constructtransformations of the acoustic features between twospeakers with Maximum Likelihood Linear Regression(MLLR)based on the acoustic feature domainconcatenative speech synthesis,Some results of aseries of experiments which are based on IBMtrainablc specch synthesis system,IBM cepstralreconstruction and LSP reconstruction algorithms arepresented.
Eyes-free and hands-free operations are very desirable in mobile ***/output can perfectly meet the hands/eyes-free ***,it is a greatchallenge to apply speech technologies to the handheld devices operating in various *...
详细信息
Eyes-free and hands-free operations are very desirable in mobile ***/output can perfectly meet the hands/eyes-free ***,it is a greatchallenge to apply speech technologies to the handheld devices operating in various *** first challenge comes from the limited system resources available to speechapplications in handheld *** is well known that ASR and TTS are either computationintensive or memory hungry *** challenges come from the tough performancerequirements in mobile conditions,such as noise robustness and high accuracy of ASRapplied in the noisy mobile environments,and the naturalness of TTS in handheld *** limited system resources prohibit using complex but effective algorithms to deal with theproblems.
In the history of information revolution the user interface has been one of the key drivingforces leading to a paradigm *** greatest opportunity for the next paradigm shift is toempower people to access web Informatio...
详细信息
In the history of information revolution the user interface has been one of the key drivingforces leading to a paradigm *** greatest opportunity for the next paradigm shift is toempower people to access web Information and services anywhere,any time,and from *** technologies will provide a vital role In enabling this since speech is not onlythe most natural way for people to interact with machines but also the only consistent inputmodality that can support multiple devices such as cell phones,PDAs,AntoPCs,and PCs.
IntroductionThis paper addresses the challenges of designing user interfaces and building applications that work acrossthese multiplicities of information appliances. Amongst the key issues addressed are the user’s a...
详细信息
IntroductionThis paper addresses the challenges of designing user interfaces and building applications that work acrossthese multiplicities of information appliances. Amongst the key issues addressed are the user’s ability tointeract in parallel with the same information via a multiplicity of channels and user interfaces,and theneed to present a unified,synchronized view of information across the various channels that the userdeploys to interact with *** achieve such synchronized interactions by adopting the well-known Model,View,Controller(MVC) design paradigm and adapting it to multi-modal interactions.
Class Language Model(LM)has been proved to be usefulin the case of sparse training *** traditionally,it hasto be combined with word LM to get attractiveperformance,because the existing classification criteriaonly invo...
详细信息
Class Language Model(LM)has been proved to be usefulin the case of sparse training *** traditionally,it hasto be combined with word LM to get attractiveperformance,because the existing classification criteriaonly involve language context information or *** ovcrcome this,our paper presents a newmethod to cluster the words that have different soundingand similar language context into one *** that the class LM based on the proposed methodachieves good performance without combining with a wordLM.
暂无评论