Hindi Voice Recording Contributions

Role Type: Freelance / Independent Contributor Location: Ahemdabad, India Language: Hindi (Native / Near-Native Fluency) Project Duration: 8–10 weeks (batch-wise delivery) Commitment: Minimum 30 minutes of final approved audio; up to 4–5 hours per speaker --- About the Project We are building a high-quality expressive Hindi speech dataset for text-to-speech (TTS), conversational AI, and voice model training. The project involves recording natural, emotionally expressive Hindi speech across a wide range of domains and speaking styles — from podcasts and storytelling to sports commentary and voice acting. Recordings must meet studio-quality audio standards and will undergo strict quality validation before acceptance. You may participate in monologue recordings, conversational (two-speaker) recordings, or both. --- Key Responsibilities - Record high-quality expressive Hindi speech in your assigned domains and speaking styles (read, spontaneous, narrative, conversational, persuasive, energetic). - Deliver natural emotional variation across recordings — e.g., happy, sad, angry, surprised, fearful, cheerful, empathetic, commanding, whisper, hesitation — as per project allocation. Emotional delivery must sound natural, never exaggerated or theatrical. - Produce non-verbal vocalizations where required (laughs, sighs, gasps, coughs, breathing, etc.) integrated naturally into speech. - Vary prosody as directed: fast/slow pacing, loud/soft tones, energetic and relaxed conversational delivery. - For conversational tasks: engage in natural two-speaker dialogues, maintain clear speaker turns, and avoid cross-talk/overlapping speech where specified. - Record utterances of 5–30 seconds each; avoid speech segments shorter than 5 seconds in conversations. - Maintain identical microphone settings and recording environment across all sessions. - Follow file naming conventions (e.g., `[login to view URL]`) and complete all required metadata (age group, profession, accent/region, recording device, domain, speaking style, etc.). - Re-record files rejected during quality checks (clipping, noise, long silences >3 seconds, unnatural delivery). - Sign an informed consent form covering data usage. --- ## Required Skills Language & Voice - Native or near-native Hindi fluency with clear diction and natural pronunciation. - Ability to read Hindi scripts fluently (Devanagari) for read-speech tasks. - Strong expressive range — capable of authentic emotional delivery across 20+ emotion/style categories without sounding forced. - Comfortable with spontaneous, unscripted speaking (interviews, commentary, general conversation). Performance & Versatility - Voice acting / character voice ability (preferred for select domains). - Storytelling, narration, or audiobook-style delivery skills. - Ability to modulate pace, volume, energy, and tone on direction. - Experience in any of: podcasting, radio, theatre, dubbing, voice-over, stand-up, anchoring, or content creation is a strong plus. Technical & Process Discipline - Basic audio recording know-how: setting input levels, avoiding clipping, monitoring for background noise/echo. - Ability to record and export audio in **WAV format at 48 kHz sampling rate**. - Consistency and attention to detail — same setup, same settings, every session. - Reliability in meeting batch deadlines (weekly delivery cadence). **Eligibility** - Minimum age: **18 years** (no exceptions). - All genders encouraged to apply — the project maintains a balanced male/female speaker ratio. - Speakers across age groups (18–60+), regional accents, and backgrounds are actively sought for dataset diversity. --- ## Equipment & Setup Required - Microphone: Studio-quality condenser or dynamic USB/XLR microphone (with audio interface if XLR). Built-in laptop/phone mics are **not acceptable**. - Recording environment: Quiet, acoustically treated or well-dampened room — no background noise, echo, or reverberation. Clipping-free, distortion-free capture. - Recording software: Any DAW or recorder capable of exporting **48 kHz WAV** (e.g., Audacity, Adobe Audition, Reaper). - Headphones: Closed-back headphones for monitoring playback (recommended). - Pop filter / windscreen: Recommended for clean plosive-free recordings. - Computer & internet: Stable internet connection for uploading large WAV files and working on the designated recording/annotation platform. - Consistency requirement: The same microphone, settings, and environment must be used for the entire project duration. --- ## Quality Standards - All submissions undergo 100% quality review; only validated recordings count toward paid hours. - No clipping, no unintended silences longer than 3 seconds, clear intelligible speech throughout. - Unnatural or exaggerated emotional delivery, noisy recordings, and out-of-spec audio will be rejected. - No sensitive, offensive, or personally identifiable content in recordings. --- ## Compensation - Paid per approved audio hour (rates shared during onboarding, based on task type — monologue vs. conversational). - Payment released post batch validation and client acceptance. --- ## How to Apply Submit a short voice sample (1–2 minutes) in Hindi demonstrating at least two contrasting emotions/styles (e.g., cheerful narration + neutral news-read), along with details of your recording setup (microphone model, software, room environment).

Project ID: 40566178

Location: Ahmedabad, India