Video Recording Task for Speech Research (English only)

We are looking for participants to help us improve speech recognition technology by providing short, natural video samples. Confidentiality & NDA Requirement: Please note that this project involves proprietary research. To participate, you will be required to review and electronically sign a Non-Disclosure Agreement (NDA) before accessing the task instructions. This ensures the confidentiality of the project and the research data. Task Description: You are asked to record a total of 24 video clips (45-60 seconds each) across 4 rounds, of yourself and a partner speaking naturally about everyday topics. Each round introduces a different scenario variation (distance, head movement, movement in/out of frame, vocal effort). Detailed instructions will be provided after a short interview. Language: English only Basic instructions: Start: Begin your recording with a few seconds of silence. Solo Speech: Speak spontaneously for about 30 seconds on any topic (e.g., hobbies, weekend plans, or your favorite films). Please do not read from a script. Extra Speech: For the remainder of the recording (15–30 seconds), have your partner join you. But during entire video both participants are visible in the frame. You may speak simultaneously or take turns, depending on the specific scenario. Both speakers must be visible in the frame. You and your partner must speak about completely unrelated topics — this is not a conversation. Submissions where the two speakers interact or discuss the same topic will be rejected. End: Conclude with a few seconds of silence before stopping the recording. Scenarios – specific descriptions will be provided per round: 4 scenario combinations across 4 rounds (6, 4, 8, and 6 videos per round). 2 speech modes (overlapping and alternating). 24 videos in total. Requirements: Video Format: 1080p resolution, (16-) 30 (-60) FPS, horizontal orientation (landscape), saved as .mp4 or .mov. Vertical (portrait) videos will be rejected. Audio Format: Set your device's audio quality to the highest setting supported. Language: English, any accent. Age: Adults only. Participation of children is strictly prohibited. Supported devices: Desktops, Laptops, Tablets, Cell Phones. Camera must remain stationary during recording. Environment: Record in a natural setting (e.g., home office desk, living room table, dining room table). Coworking spaces are also suitable if you have permission to record audio/video there. In general — places where you would typically have a video call. Content: Please ensure the topics discussed are casual, spontaneous, and unrelated between you and your partner. Ideas will be provided if needed. Privacy: Do not mention any personally identifiable information (names, phone numbers, addresses, etc.). Guidelines: Avoid sensitive or controversial topics (e.g., politics, religion). Do not comment on the recording conditions, the hardware setup, the room you are in, or the sound of your voice. Note: This task requires a partner to act as a background speaker. Please ensure you have someone available to assist you before accepting this task. All submissions will be reviewed for quality and adherence to these instructions. This task is part of a 4-round series, requiring 24 videos in total. New scenarios will be proposed after succeeding in basic ones. Payment: We propose $5 per 2 videos. We will cover the fees