Speech to Singing Samples
Speech : Source Speech
Singing : Template Singing
Zero-effort Baseline : Template-based Speech-to-Singing Conversion
SD-MCC Baseline : Speaker-dependent Spectral Mapping Model using Speech MCCs
SI : Speaker-independent Spectral Mapping Model
SC : Proposed Speaker-conditioned Spectral Mapping with i-vectors
SC v.s. Zero-effort Baseline
|
Speech |
Singing |
Zero-effort Baseline |
SC |
| Sample 1 |
|
|
|
|
| Sample 2 |
|
|
|
|
SC v.s. SD-MCC Baseline
|
Speech |
Singing |
SD-MCC Baseline |
SC |
| Sample 1 |
|
|
|
|
| Sample 2 |
|
|
|
|
SD-MCC Baseline v.s SD
|
Speech |
Singing |
SD-MCC Baseleine |
SD |
| Sample 1 |
|
|
|
|
| Sample 2 |
|
|
|
|
SC v.s. SD
|
Speech |
Singing |
SD |
SC |
| Sample 1 |
|
|
|
|
| Sample 2 |
|
|
|
|
SC v.s. SI
|
Speech |
Singing |
SI |
SC |
| Sample 1 |
|
|
|
|
| Sample 2 |
|
|
|
|