Kuiz Vokoder Berbilang Resolusi UnivNet

Jawapan dan penjelasan muncul semasa anda mengambil kuiz.

1. Apakah ciri tandatangan pendiskriminasi UnivNet?
   • Ia mengira berbilang spektrogram STFT pada resolusi berbeza dan mendiskriminasikannya
   • Ia menggunakan diskriminator bentuk gelombang tunggal sahaja
   • Ia menilai audio menggunakan model bahasa
   • Ia mengabaikan frekuensi tinggi
2. Apakah masalah dalam vocoder GAN terdahulu yang disasarkan khusus oleh UnivNet?
   • Persampelan autoregresif perlahan
   • Frekuensi tinggi yang tersekat atau sarat artifak
   • Ketidakupayaan untuk berjalan pada GPU
   • Kekurangan sebarang diskriminasi
3. Apakah konvolusi pembolehubah lokasi (LVC) dalam UnivNet?
   • Konvolusi hanya digunakan pada permulaan isyarat
   • Inti tetap dikongsi merentas semua lapisan
   • Konvolusi yang intinya diramalkan secara dinamik daripada ciri input
   • Sambungan berulang antara bingkai
4. What adaptation potential did the UnivNet paper report for new speakers?
   • Guaranteed identical quality across every recording condition
   • No need for conditioning features or evaluation
   • Instant adaptation to every language and microphone
   • Fast adaptation without training a new vocoder from scratch
5. What synthesis capability did the UnivNet paper report?
   • High-fidelity waveform synthesis in real time
   • Only offline synthesis after a day-long wait
   • One waveform sample per hour
   • A text transcript without audio output
6. Apakah keadaan penjana UnivNet untuk menghasilkan bentuk gelombang?
   • Teks mentah
   • Satu mel-spektrogram
   • Spektrogram bunyi bising sahaja
   • Jadual tempoh fonem
7. Which adversarial loss did UnivNet use to train its generator and discriminators?
   • A word-level cross-entropy loss
   • A policy-gradient objective
   • A least-squares GAN objective
   • A contrastive image-text loss
8. What evaluation result did the UnivNet paper report across seen and unseen speakers?
   • It outperformed every vocoder on every language and dataset
   • It proved no speaker adaptation is ever needed
   • It achieved the same result regardless of input features
   • It achieved the best objective and subjective results among the compared models in its evaluation