use mobile device to record free speech in Thai

via Freelancer ·

Budget / SalaryHourly project
TypeFreelance project
LocationRemote
Posted1 hour ago
# Thai Free-Speech Recording Project

We are inviting **native Thai speakers** to participate in a solo, free-speech recording project. Speak naturally about topics of your choice, such as work, travel, food, technology, personal experiences, or everyday life, subject to the content restrictions below.

This is a **long-term project with multiple batches**. The requirements below apply to the current batch.

## 1. Recording Quantity and Deadline

- **Target per participant:** 30 hours.
- **Minimum per participant:** 10 hours.
- **Maximum per participant:** 30 hours.
- **Length of each recording:** 3–30 minutes.
- **Submission schedule:** Recordings must be submitted daily.
- **Current batch deadline:** September 22, 2026. Recording will continue until this date.

## 2. Recording Requirements

- Record in a quiet environment using a mobile phone or computer microphone.
- Speak naturally and clearly in Portuguese.
- Each recording must contain only one speaker.
- Speak spontaneously. Reading from scripts, books, articles, or any other written material is not allowed.
- Include natural expressions, interjections, emotions, and variations in your speaking style.
- Submit files in **WAV or MP3 format**.
- All recordings must be newly created by you specifically for this project.
- Do not submit copied, downloaded, AI-generated, previously published, or previously submitted audio, including audio shared on YouTube, Spotify, Apple Music, or other platforms.
- Follow all quality and content requirements provided by the project team.

## 3. Prohibited Content — Strictly Enforced

Recordings must not include:

- Threats of physical harm, descriptions of assault, or content promoting fighting or killing.
- Content encouraging self-harm, eating disorders, or injury to others.
- Slurs or content encouraging discrimination based on race, religion, ethnicity, disability, or sexual orientation.
- Pornography, sexually explicit descriptions, or references to sexual violence.
- Sensitive topics, including religion and politics.
- Any references to the harm or exploitation of minors.
- Instructions for criminal activities, drug manufacturing, hacking, or terrorism.
- Personal information about other people.
- Other parties’ trade secrets or unauthorized use of their protected intellectual property, including copyrighted, patented, or trademarked material.

## 4. Payment

Payment is calculated based on the duration of **valid and accepted audio**, rather than the total duration submitted.

## Valid Audio Duration Calculation

Payment is based on the duration of valid audio remaining after the following processing and quality checks:

1. **Pause removal and trimming:** Any pause longer than 2.5 seconds is removed in full, and the audio is trimmed.
2. **Segmentation:** The remaining audio is divided into one-minute segments.
3. **Quality checks:** Any segment containing the following is excluded:
- Excessive background noise.
- Unclear or unintelligible speech.
- More than seven filler words or sounds within one minute, such as “um,” “uh,” or “ahem.”

### Calculation Example

You submit recordings totaling **one hour**, containing:
- **10 pauses**, each lasting five seconds: **50 seconds removed**.
- **Two separate one-minute segments** with excessive background noise: **two minutes excluded**.
- **One additional one-minute segment** containing more than seven filler words or sounds: **one minute excluded**.

**Valid audio duration = 60 minutes − 50 seconds − 2 minutes − 1 minute = 56 minutes and 10 seconds.**

Each excluded segment is counted only once, even if it contains multiple quality issues.

**Recordings containing sensitive topics or other prohibited content will not be eligible for payment.** All recordings must meet the project’s quality and content requirements to qualify.
data processing training transcription
Apply on Freelancer →

Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.