TensorFlowTTS: Real-Time State-of-the-art Speech Synthesis for Tensorflow 2 demo

This is the demonstration page of TensorFlowTTS: Real-Time State-of-the-art Speech Synthesis for Tensorflow 2 demo.

Include:

  • Tacotron-2 based on Tensorflow 2
  • FastSpeech based on Tensorflow 2
  • MelGAN based on Tensorflow 2
  • MelGAN STFT based on Tensorflow 2

Github: https://github.com/TensorSpeech/TensorFlowTTS

🎧 Model samples


Tacotron-2 samples

Peter Piper picked a peck of pickled peppers. How many pickled peppers did Peter Piper pick?

Tacotron-2 + MelGAN Tacotron-2 + MelGAN-STFT Tacotron-2 + Multi-band MelGAN

There’s a way to measure the acute emotional intelligence that has never gone out of style.

Tacotron-2 + MelGAN Tacotron-2 + MelGAN-STFT Tacotron-2 + Multi-band MelGAN

The Commission further recommends that the Secret Service coordinate its planning as closely as possible with all of the Federal agencies from which it receives information.

Tacotron-2 + MelGAN Tacotron-2 + MelGAN-STFT Tacotron-2 + Multi-band MelGAN

Unless you work on a ship, it's unlikely that you use the word boatswain in everyday conversation, so it's understandably a tricky one. The word - which refers to a petty officer in charge of hull maintenance is not pronounced boats-wain Rather, it's bo-sun to reflect the salty pronunciation of sailors, as The Free Dictionary explains.

Tacotron-2 + MelGAN Tacotron-2 + MelGAN-STFT Tacotron-2 + Multi-band MelGAN

FastSpeech samples

President Trump met with other leaders at the Group of twenty conference.

FastSpeech (V1) + MB-MelGAN FastSpeech (V2) + MB-MelGAN FastSpeech (V3) + MB-MelGAN

Scientists at the CERN laboratory, say they have discovered a new particle.

FastSpeech (V1) + MB-MelGAN FastSpeech (V2) + MB-MelGAN FastSpeech (V3) + MB-MelGAN

There’s a way to measure the acute emotional intelligence that has never gone out of style.

FastSpeech (V1) + MB-MelGAN FastSpeech (V2) + MB-MelGAN FastSpeech (V3) + MB-MelGAN

The Senate's bill to repeal and replace the Affordable Care-Act is now imperiled.

FastSpeech (V1) + MB-MelGAN FastSpeech (V2) + MB-MelGAN FastSpeech (V3) + MB-MelGAN

Speed control

For a while the preacher addresses himself to the congregation at large, who listen attentively.

FastSpeech (V3) + MB-MelGAN + 25% slower FastSpeech (V3) + MB-MelGAN + normal FastSpeech (V3) + MB-MelGAN + 25% faster

FastSpeech-2 samples (BBC news)

The Rhodes Must Fall campaigners said the announcement was hopeful, but warned they would remain cautious until the college had actually carried out the removal.

FastSpeech-1 (V3) + MB-MelGAN FastSpeech-2 (V1) + MB-MelGAN Tacotron-2 (V1) + MB-MelGAN

The nation's tourism minister has also encouraged Australian's to take their holidays within the country this year.

FastSpeech-1 (V3) + MB-MelGAN FastSpeech-2 (V1) + MB-MelGAN Tacotron-2 (V1) + MB-MelGAN

More films are scheduled to be released by streaming companies over the coming months, as the prospects of cinemas opening in India in the near future look dim.

FastSpeech-1 (V3) + MB-MelGAN FastSpeech-2 (V1) + MB-MelGAN Tacotron-2 (V1) + MB-MelGAN

As the virus has spread around the world, countries have reached different stages of their outbreaks at different times, In some places the number of excess deaths may still increase in the coming weeks and months, especially as figures are revised, while in others the number of deaths is beginning to return to normal levels.

FastSpeech-1 (V3) + MB-MelGAN FastSpeech-2 (V1) + MB-MelGAN Tacotron-2 (V1) + MB-MelGAN

This doesn't give you the same rush as watching a movie's Friday release in the cinema. But we are now reconciled to the fact that we may not catch a movie in the cinemas in the near future

FastSpeech-1 (V3) + MB-MelGAN FastSpeech-2 (V1) + MB-MelGAN Tacotron-2 (V1) + MB-MelGAN

One wonders how long this state of affairs can be allowed to persist while social media platforms continue to provide a worldwide distribution mechanism for medical misinformation.

FastSpeech-1 (V3) + MB-MelGAN FastSpeech-2 (V1) + MB-MelGAN Tacotron-2 (V1) + MB-MelGAN

MelGAN samples

GroundTruth MelGAN MelGAN-STFT Multi-band MelGAN

References

Author

Minh Nguyen Quan Anh, Eren Gölge, Kuan Chen, Takuya Ebata
e-mail: nguyenquananhminh@gmail.com, erengolge@gmail.com, azraelkuan@gmail.com, meguru.mokke@gmail.com