Upload dataset.md
Browse files- dataset.md +89 -0
dataset.md
ADDED
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# DreamVoiceDB: Voice Timbre Dataset
|
2 |
+
|
3 |
+
The DreamVoiceDB dataset is an integral part of our research on voice timbre and its diverse characteristics. This page provides a detailed overview of the dataset creation process, the criteria for speaker selection, the methodology for keyword annotation, and our approach to enhancing the dataset's richness and authenticity.
|
4 |
+
|
5 |
+
## Overview
|
6 |
+
|
7 |
+
Voice timbre is shaped by a myriad of factors, including age, gender, physical properties of the vocal tract and vocal cords, and perceptual characteristics. To effectively encapsulate the multifaceted nature of voice timbre, we employed a comprehensive three-stage process.
|
8 |
+
|
9 |
+
|
10 |
+
#### Figure: Survey Method
|
11 |
+
<img src="./data/survey_method.png" alt="Schematic diagram of DreamVoiceDB survey method" width="600"/>
|
12 |
+
|
13 |
+
*Figure 1: Schematic diagram of DreamVoiceDB survey method.*
|
14 |
+
|
15 |
+
|
16 |
+
## Details
|
17 |
+
|
18 |
+
- **Datasets Used**: LibriTTS-R and VCTK
|
19 |
+
- **Number of Speakers**: 900 [100 from VCTK and 800 from LibriTTS-R, sampled randomly]
|
20 |
+
- **Keyword Selection Process**:
|
21 |
+
- Guided by an expert voice actor
|
22 |
+
- Total of 10 keywords identified
|
23 |
+
- Keywords categorized based on relative objectivity and subjectivity
|
24 |
+
- **Annotator Profile**:
|
25 |
+
- Total Annotators: 8 (4 females and 4 males)
|
26 |
+
- Expertise: Speech pathology, accent coaching, singing coaching, transcription, and translation
|
27 |
+
|
28 |
+
You can find our sample survey page used for data collection here - [DreamVoiceDB Survey Link](https://sm7orzxyhu.cognition.run/)
|
29 |
+
|
30 |
+
## Keywords and Descriptors
|
31 |
+
|
32 |
+
The DreamVoiceDB dataset utilizes a variety of keywords and descriptors to annotate voice timbre. These are grouped into relatively objective and subjective categories, along with additional aspects related to suitability for various voice-related professions.
|
33 |
+
|
34 |
+
### Relatively Objective Keywords [Multtple Options Single Choice]
|
35 |
+
- **Age**: 'Teenager or Young', 'Adult or Neutral', and 'Senior or Old'.
|
36 |
+
- **Gender**: 'Definitely Female', 'Somewhat Ambiguous / Neutral'.
|
37 |
+
- **Brightness**: Ranging from 'Very Dark (Rich)' to 'Very Bright (Thin)'.
|
38 |
+
- **Roughness**: From 'Rough (Raspy)' to 'Smooth (Mellow)'.
|
39 |
+
|
40 |
+
Reference Audio can be found in the following powerpoint presentation: [DreamVoiceDB Reference Audio](./data/reference_audio.pptx)
|
41 |
+
|
42 |
+
### Relatively Subjective Keywords [Multiple Options Multiple Choice]
|
43 |
+
- **Nasal**: A noticeable resonance or vibration in the nasal passages.
|
44 |
+
- **Breathy**: You can hear the flow of air as the person speaks.
|
45 |
+
- **Weak**: Lack of energy and confidence.
|
46 |
+
- **Strong**: Full, resonant, and confident.
|
47 |
+
- **Cute**: Perceived as young and endearing.
|
48 |
+
- **Attractive**: A voice that draws pleasure in listeners.
|
49 |
+
- **Cool**: Laid-back, confident, and collected.
|
50 |
+
- **Warm**: A voice that conveys kindness, sincerity, and comfort.
|
51 |
+
- **Authoritative**: A voice that carries weight and command.
|
52 |
+
- **Others**: Other characteristics but not related to the speaker's emotion.
|
53 |
+
|
54 |
+
### Additional Aspects
|
55 |
+
This includes evaluating the voice's fit for professions such as singing, acting, public speaking, and other vocations where vocal qualities are paramount. Specific examples include:
|
56 |
+
- **Public Presentation**: Public Speaker, Motivational Speaker, Teacher, Tour Guide
|
57 |
+
- **Storytelling**: Podcaster, Audiobook Narrator
|
58 |
+
- **Client and Public Interaction**: Customer Service, Healthcare Communicator, Retail Associate, Receptionist
|
59 |
+
- **Diplomacy and Judiciary**: Diplomat, Judge
|
60 |
+
- None of the above
|
61 |
+
|
62 |
+
Each keyword and descriptor plays a crucial role in accurately capturing and conveying the unique characteristics of each voice sample in the dataset.
|
63 |
+
|
64 |
+
### Analysis and Enhancement
|
65 |
+
|
66 |
+
- Agreement scores used to prioritize keywords
|
67 |
+
- Rigorous reassessment for moderately agreed keywords
|
68 |
+
- For each speaker we atleast have age and gender keywords
|
69 |
+
- Other keywords are selected based on the Agreement scores
|
70 |
+
- We do not use any fine grained annotations for darkness and roughness in the model training
|
71 |
+
- Use of OpenAI's GPT4 for natural language descriptors
|
72 |
+
- Generation of descriptors based on all keyword combinations
|
73 |
+
|
74 |
+
|
75 |
+
Generated Descriptors Examples:
|
76 |
+
|
77 |
+
| Speaker ID | WAV File | Keywords | Prompt |
|
78 |
+
|------------|-----------------|-----------------------|-----------------------------------------------------|
|
79 |
+
| 2562 | [File 2562](#) | Male, Adult, Dark | An adult male voice, with a dark timbre. |
|
80 |
+
| 30 | [File 30](#) | Female, Teenager, Bright | A teenage girl's voice, radiating brightness and energy. |
|
81 |
+
| 2156 | [File 2156](#) | Male, Senior, Strong | A senior male voice, characterized by strength and power. |
|
82 |
+
|
83 |
+
|
84 |
+
## Dataset Statistics
|
85 |
+
|
86 |
+

|
87 |
+

|
88 |
+

|
89 |
+
|