Model-generated English synthetic speech database

1. Product Overview

The model-generated English synthetic speech database comprises a total of 1.51 million entries, collected from 61 speakers. The synthesis engines used are from Volcano, Microsoft, and OpenAI: Volcano (24 kHz, 16-bit); Microsoft (48 kHz, 16-bit); and OpenAI (24 kHz, 32-bit).

2. Product Directory Structure

Directory Structure

Data directory structure

└─spker-info.xlsx (voice tone information)
└─SPEECHDATA (data folder)
│├─HS (synthesis platform)
│├─common_Ava (voice tone name)
││ ││ common_Ava_01250001.wav (audio file)
││ ││ common_Ava_01250001.txt (speech text content)
│├─openai (synthesis platform)

Naming Conventions

CORPUS/USAGE/PLATFORM/SPEAKER/SPEECH_IDe.g./SPEECHDATA/HS/common_Ava/ common_Ava_01250001.wav

Directory Name · Content · Notes

CORPUS · · Speech database ID

USAGE · SPEECHDATA · Folder name

PLATFORM · HS · Synthesis interface platform

SPEAKER · common_Ava · Emotion/Voice Tone

SPEECH_ID · common_Ava_01250001.wav · WAV file

3. Speaker Information

The gender distribution of the speakers is shown in the table below. In total, there are 61 voice tones, covering various platforms as well as 19 distinct emotional styles. Each speaker contributes 10,000 recordings per emotional style, with details provided as follows:

Serial Number · Platform · Speaker · Style Type

1 · Lazy Female Voice – Ava · General ·

2 · HS · Lazy Female Voice – Ava · Sadness

3 · Emotional Female Voice – Lawrence · Narrator – Calm ·

Sample Data Download

Click to download the sample data (sample_data.zip)