Emotional TTS
Research

Results & Audio Samples

Audio samples generated by the emotional TTS system under different datasets, emotion control methods, and emotion settings. Compare reference audio with model-generated speech.

3
Datasets
2
Methods
6
Emotions
24
Total Samples

About These Results

Emotion Control Modes

  • Reference-Based: Emotional style extracted from a reference audio sample
  • Vector-Based: Emotion controlled via predefined embedding vectors

Evaluation Datasets

  • CV3_Eval & EmoVoiceDB: Zero-shot evaluation (unseen speakers)
  • ESD: Seen-speaker evaluation (speakers included in training)

Filters

(24 samples)
CV3_EvalReference-basedAngry#1

Always resorting to underhanded tactics and cheap tricks, it's just so unethical!

Reference Audio

Reference

Generated Audio

Generated

CV3_EvalReference-basedAngry#2

The professor was furious, his face turning bright red like a spark falling into a pool of gasoline.

Reference Audio

Reference

Generated Audio

Generated

CV3_EvalReference-basedAngry#3

If you'd ever treated me like a human being, would they have dared to do this?

Reference Audio

Reference

Generated Audio

Generated

CV3_EvalReference-basedHappy#1

I bought my first mountain bike with my own earnings, a Merida Warrior 500! Go me!

Reference Audio

Reference

Generated Audio

Generated

CV3_EvalReference-basedHappy#2

As the sunlight warmed my face and a gentle breeze brushed past, it felt like all my worries were blown away.

Reference Audio

Reference

Generated Audio

Generated

CV3_EvalReference-basedSad#1

The research institute won't renew my contract, so I have to go back to my country. I don't even have a job anymore.

Reference Audio

Reference

Generated Audio

Generated

CV3_EvalReference-basedSad#2

Why did this happen? I really tried my best. Why did I still lose everything?

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBReference-basedFear#1

The power might go out during the storm.

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBReference-basedFear#2

The children are playing in the yard.

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBReference-basedSad#1

If I had been more determined back then, the three of us wouldn't have ended up like this.

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBReference-basedSurprise#1

He finished reading the document yesterday.

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBVector-basedAngry#1

You never show any gratitude — always making others worry about you!

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBVector-basedAngry#2

If these examinations are held orally, they may be known colloquially as orals.

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBVector-basedHappy#1

I discovered an amazing place today! I'll take you there next time!

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBVector-basedSad#1

Stop acting like such special snowflakes.

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBVector-basedSurprise#1

He arrived earlier than everyone else, even though he usually comes late.

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBVector-basedSurprise#2

Wait, today is your birthday?!

Reference Audio

Reference

Generated Audio

Generated

EmoVoiceDBVector-basedAfraid#1

Reference Audio

Reference

Generated Audio

Generated

ESDReference-basedAngry#1

Do you have any idea how worried I was when you left without saying a word?

Reference Audio

Reference

Generated Audio

Generated

ESDReference-basedAngry#2

The party stands for economic liberalism and Federal Peronism.

Reference Audio

Reference

Generated Audio

Generated

ESDReference-basedHappy#1

I ran into a teacher I hadn't seen in years at the coffee shop today.

Reference Audio

Reference

Generated Audio

Generated

ESDReference-basedHappy#2

I saw a little dog playing with its owner on the street today.

Reference Audio

Reference

Generated Audio

Generated

ESDReference-basedSad#1

I've known him for eight years, and he still cheated on me.

Reference Audio

Reference

Generated Audio

Generated

ESDReference-basedSurprise#1

Really? He actually came first in the exam?

Reference Audio

Reference

Generated Audio

Generated