Ikibazo gihuza ibibazo · Hagati Urwego
Guhinduranya Ikibanza nigihe cyo Kurambura Ikibazo
Tandukanya ikibanza gihindagurika uhereye igihe urambuye kandi ugereranye amajwi ya majwi, PSOLA nuburyo bwimitsi hamwe nibikorwa byabo.
Ikibazo 1 Bya 8
Mugihe cyo guhinduranya ikibanza, ni uwuhe mutungo ugomba guhinduka mugihe igihe gikomeza kuba gihamye?
Reba buri kibazo8 ibibazo
Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.
- Mugihe cyo guhinduranya ikibanza, ni uwuhe mutungo ugomba guhinduka mugihe igihe gikomeza kuba gihamye?
- Change duration while preserving pitch
- Change loudness while preserving both
- Change perceived pitch while preserving duration
- Convert stereo audio to mono
- Mugihe cyo kurambura, ni uwuhe mutungo ugomba guhinduka mugihe ikibanza kigumye gihamye?
- Change pitch while preserving duration
- Change duration while preserving pitch
- Remove all transients
- Change only the sample-rate metadata
- Bigenda bite iyo igipimo cyo gukina gihinduwe neza nta ndishyi?
- Only pitch changes
- Only duration changes
- Neither changes
- Duration and pitch both change
- Ni ubuhe buryo bwo guhagararira icyiciro gisanzwe-amajwi arahindura?
- Short-time Fourier transform domain
- A text-token sequence from a speech recognizer
- A pitch-class histogram with all phase discarded
- An unframed average power spectrum with no time information
- Ni ikihe kimenyetso gishobora kwerekana gusebanya byigihe gito mumajwi ya majwi?
- A long steady sinusoid without a transient onset in the evaluated region
- A slowly varying sustained tone
- Percussive audio with sharp attacks
- A stationary segment whose amplitude and frequency remain constant
- What does PSOLA rely on for voiced speech processing?
- An arbitrary frame grid unrelated to voicing
- Pitch-synchronous marks or periods
- Only the transcript's word boundaries
- Only the average spectrum of the whole recording
- During ordinary sample-rate conversion, what happens when the output sample rate is correctly updated to represent the same signal?
- Pitch necessarily rises at every new sample rate
- Duration always halves during any rate conversion
- Pitch must change even when the output represents the same time span
- Pitch and duration are preserved rather than intentionally shifted
- How should an audio augmentation transformation be chosen?
- Keep changes plausible and preserve the task label
- Use the largest possible stretch
- Apply it to test data until results improve
- Choose a method without listening