Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo cya Vocoders Ikibazo

Ongera usuzume uburyo amajwi arangurura amajwi yubaka imiterere, kuki synthesis ya autoregressive itinda, nuburyo ivangura rirwanya ritezimbere amajwi yakozwe.

Inzira zijyanye no kuyoboraVocoders
Ikibazo 1 Bya 8

Nuwuhe murimo wibanze wijwi ryijwi?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Nuwuhe murimo wibanze wijwi ryijwi?
    • Guhindura acoustic igereranya nka mel-spectrogramu muburyo bwumvikana
    • Guhindura inyandiko mubisobanuro bitaziguye
    • Guhindura imvugo hagati yindimi
    • Gukuraho urusaku rwimbere mumajwi
  2. Niyihe moderi ya 2016 niyambere yatumye amajwi ya neural yumvikana neza?
    • Muraho-GAN
    • Umuhengeri
    • Tacotron
    • MelGAN
  3. Kuki WaveNet yumwimerere yatinze kubyara amajwi?
    • Byasabye GPU itandukanye kuri buri jambo
    • Yongeye kwitoza kuri buri cyinjijwe
    • Yahanuye amajwi icyitegererezo icyarimwe icyarimwe
    • Yatanze amashusho kuruhande rwamajwi
  4. Ni ayahe makuru mel-spectrogramu yata cyane, bigatuma akazi k'ijwi katoroshye?
    • Ijwi ryamajwi
    • Ururimi ruvugwa
    • Igipimo cy'icyitegererezo
    • Icyiciro amakuru yicyapa
  5. Nigute amajwi ashingiye kuri GAN nka HiFi-GAN atezimbere realism?
    • Ukoresheje ivangura ryinshi risuzuma ibimenyetso kumunzani zitandukanye
    • Muguhindura amajwi mwandiko mbere
    • Ugereranije ibihumbi byafashwe amajwi hamwe
    • Mugukora gusa imirongo mike
  6. Mubisanzwe bisanzwe bigezweho bya TTS, niki gitanga mel-spectrogramu amajwi akoresha?
    • Imashini itanga urusaku
    • Moderi ya acoustic nka Tacotron 2 cyangwa yihuta
    • Ijwi ubwaryo
    • Mikoro
  7. What component in HiFi-GAN combines residual blocks with different dilation patterns?
    • A phoneme tokenizer
    • A recurrent attention decoder
    • The multi-receptive-field fusion (MRF) module
    • A speaker-verification classifier
  8. What does HiFi-GAN’s multi-period discriminator help assess?
    • Whether the text transcript is grammatically correct
    • Whether the speaker’s face matches the voice
    • Whether the mel-spectrogram uses enough colors
    • Whether generated waveform segments reflect realistic periodic patterns