|
|
Conversion of Whispered to Normal Voice
Gajda, Richard ; Černocký, Jan (referee) ; Brukner, Jan (advisor)
Cílem této práce je vyvinout funkční program, který konvertuje vstupní šeptanou řeč na neutrální za pomoci predikce hlasového buzení, která je získána pomocí neuronových sítí. Práce je založena na studii z Indian Institute of Science v indickém Bengalúru. Řešení je provedeno následovně: nejprve získáme trénovací dataset řečníků, poté implementujeme zpracování řeči a její parametrizaci za pomoci vokodéru WORLD, vytvoříme a natrénujeme neuronovou síť, provedeme experimenty, které vyhodnotíme, a nakonec navrhneme použití pro budoucí aplikace a vylepšení.
|
|
|
Penetration Tests of Speaker Verification System
Nguyen, QuangTrang ; Rohdin, Johan Andréas (referee) ; Plchot, Oldřich (advisor)
The aim of this bachelor thesis is to create a penetration tests of speaker verification system with the use of the speech synthesis method. This work studies methods of spoofing against automatic speaker verification system. Before designing of the test set, the system and it's components that were used in this work are described. The last chapters of this work include a description of the process of designing the test set, realization of the designed test and the last part contains evaluation of the results and answers the question if it is possible to penetrate a verification system with the use of speech synthesis.
|
|
|
Efficient neural speech synthesis
Vainer, Jan ; Dušek, Ondřej (advisor) ; Hajič, Jan (referee)
While recent neural sequence-to-sequence models have greatly improved the quality of speech synthesis, there has not been a system capable of fast training, fast inference and high-quality audio synthesis at the same time. In this the- sis, we present a neural speech synthesis system capable of high-quality faster- than-real-time spectrogram synthesis, with low requirements on computational resources and fast training time. Our system consists of a teacher and a student network. The teacher model is used to extract alignment between the text to synthesize and the corresponding spectrogram. The student uses the alignments from the teacher model to synthesize mel-scale spectrograms from a phonemic representation of the input text efficiently. Both systems utilize simple convo- lutional layers. We train both systems on the english LJSpeech dataset. The quality of samples synthesized by our model was rated significantly higher than baseline models. Our model can be efficiently trained on a single GPU and can run in real time even on a CPU. 1
|
|
|
Multilingual speech synthesis
Nekvinda, Tomáš ; Dušek, Ondřej (advisor) ; Peterek, Nino (referee)
This work explores multilingual speech synthesis. We compare three models based on Tacotron that utilize various levels of parameter sharing. Two of them follow recent multilingual text-to-speech systems. The first one makes use of a fully-shared encoder and an adversarial classifier that removes speaker-dependent information from the encoder. The other uses language-specific encoders. We introduce a new approach that combines the best of both previous methods. It enables effective parameter sharing using a meta- learning technique, preserves encoder's flexibility, and actively removes speaker-specific information in the encoder. We compare the three models on two tasks. The first one aims at joint multilingual training on ten languages and reveals their knowledge-sharing abilities. The second concerns code-switching. We show that our model effectively shares information across languages, and according to a subjective evaluation test, it produces more natural and accurate code-switching speech.
|
|
|
Voice Conversion
Lukáč, Peter ; Glembek, Ondřej (referee) ; Černocký, Jan (advisor)
Predmetom tejto práce je konverzia hlasu. Konverzia hlasu predstavuje preberanie reči jedného rečníka, ktorého nazývame zdrojový rečník a transformovanie tejto reči na reč ktorá znie ako reč druhého rečníka, ktorého nazývame cieľový rečník. Toto je dosiahnuté pomocou systému pre konverziu hlasu, ktorý je popísaný v tejto práci. Ako framework pre analýzu a syntézu reči používame STRAIGHT, ktorý bol dominantne používaný vo Voice Conversion Challenge 2016. Náš system pre konverziu hlasu je založený na konverzii spectra použitím doprednej neurónovej siete a paralelného trénovania.
|
|
|
Improving text-to-speech in spoken dialogue systems by employing user's feedback
Hudeček, Vojtěch ; Žabokrtský, Zdeněk (advisor) ; Peterek, Nino (referee)
Although spoken dialogue systems have greatly improved, they still cannot handle communications involving unknown topics. One of the problems is, that they experience difficulties when they should pronounce unknown words. We will investigate methods that can improve spoken dialogue systems by correcting the pronunciation of unknown words. This is a crucial step to provide a better user experience, since for example mispronounced proper nouns are highly undesirable. Incorrect pronunciation is caused by imperfect phonetic representation of the word. We aim to detect incorrectly pronounced words, use knowledge about the pronunciation and user's feedback and correct the transcriptions accordingly. Furthermore, the learned phonetic transcriptions can be added to the speech recognition module's vocabulary. Thus extracting correct pronunciations benefits both speech recognition and text-to-speech components of the dialogue systems.
|
|
| |
|
|
Implementing and Improving a Speech Synthesis System
Beněk, Tomáš ; Szőke, Igor (referee) ; Hannemann, Mirko (advisor)
Tato práce se zabývá syntézou řeči z textu. V práci je podán základní teoretický úvod do syntézy řeči z textu. Práce je postavena na MARY TTS systému, který umožňuje využít existujících modulů k vytvoření vlastního systému pro syntézu řeči z textu, a syntéze řeči pomocí skrytých Markovových modelů natrénovaných na vytvořené řečové databázi. Bylo vytvořeno několik jednoduchých programů ulehčujících vytvoření databáze a přidání nového jazyka a hlasu pro MARY TTS systém bylo demonstrováno. Byl vytvořen a publikován modul a hlas pro Český jazyk. Byl popsán a implementován algoritmus pro přepis grafémů na fonémy.
|
|
|
Sound interface in application for the blind
Načeradský, Hynek ; Samec, Marek (advisor) ; Vacek, Martin (referee)
This work addresses use of sound interface in application for blind users. Correct application design is achieved by analysis of similar applications and by assesment of problematic areas in creation of applications for blind users. This work suggests rules for solving these problematic areas, which are underlined in practical part. Practical part of this work is mainly focused on design and creation of application, which offers alternative approach to file management. Application is based on speech synthesis and accessibility for blind users, which is achieved by use of open source libraries. The main benefit of this work is suggesting rules, which can serve in creating accessible applications and realization of application, which is based on those rules.
|
|
|
Speech Processing
Vích, Robert
Proceedings of the Workshop on Speech Processing is a periodic publication collecting the contributions presented at the Czech-German Workshop organized every year in September in Prague This proceedings volume includes 22 papers by 41 authors. Papers are devoted to phonetics and prosody, construction of dialogs, speech analysis, synthesis, and recognition and voice conversion.
|
guest ::
