引導式測驗 · 中等 等級
UnivNet 多重解析度聲碼器測驗
研究 UnivNet 如何在多個時頻解析度下使用自適應卷積和頻譜圖鑑別器進行語音合成。
問題 1 的 8
UnivNet的判別器的簽章特性是什麼?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- UnivNet的判別器的簽章特性是什麼?
- 它計算不同解析度下的多個 STFT 譜圖並對它們進行區分
- 它僅使用單一波形鑑別器
- 它使用語言模型來判斷音頻
- 它忽略了高頻
- UnivNet 特別針對早期 GAN 聲碼器中的哪些問題?
- 慢速自回歸取樣
- 高頻低沉或充滿偽影
- 無法在 GPU 上運行
- 缺乏任何歧視者
- UnivNet 中的位置變數卷積 (LVC) 是什麼?
- 僅在訊號開始時應用卷積
- 修復了所有層共享的內核
- 根據輸入特徵動態預測內核的捲積
- 框架之間的循環連接
- 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
- 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
- UnivNet 的產生器產生波形的條件是什麼?
- 原始文字
- 梅爾譜圖
- 僅雜訊頻譜圖
- 音素持續時間表
- 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
- 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