The Blizzard Challenge 2011

By Simon King, Vasilis Karaiskos. 2011

107 citations recorded by the source corpus.

The Blizzard Challenge 2011 was the seventh annual Blizzard Challenge which was again organised by the University of Ed- inburgh with assistance from the other members of the Blizzard Challenge committee - Prof. Keiichi Tokuda and Prof. Alan Black. One English corpus was used: the 'Nancy' corpus provided by Lessac Technologies. In commone with previous challenges, participants had the option of using labels that were provided for the corpus and for the test sentences. Index Terms: Blizzard Challenge, speech synthesis, evaluation, listening test

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Xu Tan
2. Junichi Yamagishi
3. Yu Zhang
4. Tao Qin
5. Ye Jia
6. Heiga Zen
7. Wei-Ning Hsu
8. Xin Eric Wang

Citations in the corpus

• LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speech (2019, background): “The Blizzard Challenge 2011 (BC2011) [22] corpora provides a relatively large amount of reading speech from a single professional speaker.”
• Zero-Shot Multi-Speaker Text-To-Speech with State-Of-The-Art Neural Speaker Embeddings (2019, methodology): “We initialized our multi-speaker models with parameters from a well-trained speaker-dependent model trained on the “Nancy” data from Blizzard 2011 [39] for about 105k steps.”
• MelNet: A Generative Model for Audio in the Frequency Domain (2019, methodology): “To test the model’s ability to model a single speaker in a controlled environment, we utilize the Blizzard 2013 dataset [28], which consists of audiobook narration performed in a highly animated manner by a professional speaker.”
• Deep Learning Based Assessment of Synthetic Speech Naturalness (2020, methodology): “We used all available results of the Blizzard Challenge3 [22, 23, 24, 25, 26, 27, 28, 29, 30, 31], which are currently all years except for 2017 and 2018 (see Table 2).”
• Code-Switched Speech Synthesis Using Bilingual Phonetic Posteriorgram with Only Monolingual Corpora (2020, methodology): “We use an American English speech corpus [31] and a Mandarin speech corpus [32] uttered by two female speakers to build our systems.”
• The Blizzard Challenge 2013 - Indian Language Task (2013, unknown): “[1–9] are summary papers which provide information about the previous challenges.”
• GELP: GAN-Excited Liner Prediction for Speech Synthesis from Mel-spectrogram (2019, methodology): “The experiments were conducted on the “Nancy” dataset (from Blizzard Challenge 2011 [29]), spoken by a professional female US English voice talent.”
• End-to-end Code-switched TTS with Mix of Monolingual Recordings (2019, methodology): “In our experiment, one American English speech corpus [20] and one Mandarin speech corpus [21] uttered by two different female speakers are used.”
• VoiceLoop: Voice Fitting and Synthesis via a Phonological Loop (2017, methodology): “The single speaker experiments took place on the LJ (Ito, 2017a), the Nancy corpus from the 2011 Blizzard Challenge (King & Karaiskos, 2011), and the English audiobook data for the 2013 Blizzard Challenge (King & Karaiskos, 2013).”
• A deep auto-encoder based low-dimensional feature extraction from FFT spectral envelopes for statistical parametric speech synthesis (2016, methodology): “The database provided for the Blizzard Challenge 2011 [26], which contains approximately 17 hours of speech data, comprising 12K utterances, was used for the experiment.”
• The Blizzard Challenge 2013 (2013, methodology): “For background information, please refer to the previous summary papers for 2005 [1, 2], 2006 [3], 2007 [4], 2008 [5], 2009 [6], 2010 [7], 2011 [8], 2012 [8], and the Indian language tasks of 2013 [9].”
• The Blizzard Challenge 2012 (2012, methodology): “For background information, please refer to the previous summary papers for 2005 [1, 2], 2006 [3], 2007 [4], 2008 [5], 2009 [6], 2010 [7] and 2011 [8].”
• Pretraining Strategies, Waveform Model Choice, and Acoustic Configurations for Multi-Speaker End-to-End Speech Synthesis (2020, methodology): “We try the strategy of training from scratch with VCTK only, and also, following [7], warm-starting from the studio-quality, single-speaker Nancy corpus (Blizzard 2011 [22]).”
• GELP: GAN-Excited Linear Prediction for Speech Synthesis from Mel-spectrogram (2019, methodology): “The experiments were conducted on the “Nancy” dataset (from Blizzard Challenge 2011 [29]), spoken by a professional female US English voice talent.”
• Postfilters to Modify the Modulation Spectrum for Statistical Parametric Speech Synthesis (2016, background): “There are three main reasons causing the quality degradation [11]: parameterization errors in the speech analysis/synthesis stage [27], [28], [29], inaccurate modeling in the training stage [18], [30], and over-smoothness of the generated speech parameters in the synthesis stage [25], [26].”