3 ms·
The default voice sounds robotic for several reasons. It has a low sample rate to conserve space. It is built using a mix of techniques that make it difficult t
by rhdunn 2y ago
The default voice sounds robotic for several reasons. It has a low sample rate to conserve space. It is built using a mix of techniques that make it difficult to reconstruct the original waveform exactly. And it uses things like artificial noise for the plosives, etc.
The default voice is optimized for space and speed instead of quality of the generated audio.
- vlovich123 2y agoI’ll suggest that’s the wrong optimization to make for an accessibility tool. Modern CPUs are more than capable of handling its speed requirements by several orders of magnitude (they can decode h265 in real time for gods sake without HW acceleration). And same goes for size. It’s simply the wrong tuning tradeoff.
- codedokode 2y agoBut today disk space is not an issue.