Voice Rating Poll
Listen to the audio samples and rate score according to category; headphones and quiet room recommended. You might want to turn down your volume a bit.
Method 1: TensorflowTTS/Tacotron2 & MelGAN-STFT
Listen to the eight samples. How would you rate overall for:
Audio quality
*
1
2
3
4
5
6
7
8
9
10
Terrible
Excellent
1 is Terrible, 10 is Excellent
Pronunciation
*
1
2
3
4
5
6
7
8
9
10
Terrible
Excellent
1 is Terrible, 10 is Excellent
Sentence naturalness and flow
*
1
2
3
4
5
6
7
8
9
10
Terrible
Excellent
1 is Terrible, 10 is Excellent
Method 2: NVIDIA/Tacotron2 & WaveGlow
Listen to the eight samples. How would you rate them overall for:
Audio quality
*
1
2
3
4
5
6
7
8
9
10
Terrible
Excellent
1 is Terrible, 10 is Excellent
Pronunciation
*
1
2
3
4
5
6
7
8
9
10
Terrible
Excellent
1 is Terrible, 10 is Excellent
Sentence naturalness and flow
*
1
2
3
4
5
6
7
8
9
10
Terrible
Excellent
1 is Terrible, 10 is Excellent
Conclusion
After listening to the samples, which method do you think is better overall?
*
Method 1
Method 2
Both are roughly equal
Submit
Should be Empty: