Please read these instructions and follow them completely in order for your submission to be successful. The settings on your phone should be set for the specifics in these instructions. Please set your camera/video settings on your phone to 1920×1080 and 30fps. Your upload speed needs to be at least 10 Mbps to upload your video within an hour or less. Any slower than that would cause some long transfer times.
Objective: This emotion video project requires every participant film a monologue video in North American English for at least 20 minutes, or slightly longer. We cannot take shorter videos. Your recording will be used to build facial/emotion recognition models for a smartphone app. This video will not be used for any other reason, nor will it be published to any website.
The goal is to collect intuitive and convincing visual data of a variety of emotions in everyday life.
Participants Needed: Each participant can only record one video. For this project we need individuals who were born in either the US or Canada whose first language was English.
Note: While you are recording your video, please do not talk about your location. We want your personal information to remain private.
- Stable Internet connection, wifi in use for mobile phones (for video upload)
- Mobile device with fully operational microphone and camera OR a camera capable of our resolution and FPS requirements. No headsets.
- Camera on your mobile device or other video camera – 1920×1080 at 30FPS Anything less is rejected immediately. Please do not submit QHD(2k) and UHD(4k) video due to extreme sizes.
- Camera must be stationary. Ideally you should use a tripod, but any item that can help hold the camera steady is fine.
- Quiet room (No sounds of A/C portable or window units, fans, other loud appliances i.e. washers/dryers, outside noises, running water. These are all examples.)
- No sunglasses, masks, or any other facial accessories that obstruct your face
- No obstacles between you and the camera
- No obvious shadows on your face
- Sufficient Lighting
- Must record the video indoors
Content: Do not use a script or read from a book. You can choose whatever topics you want to discuss or talk about. Personal relationships, politics, religion, etc., are fine as long as the appropriate emotions are displayed. Best case scenario (all following emotions displayed): angry, happy, sad, disappointed, bored, disgusted, excited, fear, surprise or neutral. Minimum requirement (4 basic emotions displayed): happy, angry, sad, neutral.The video should be 20 minute in length or slightly more. Less is considered invalid. We cannot combine small videos for you to make it add up.
Suggested Topics: Describe moments when you were happy, sad, angry, excited, embarrassed or any other situations that gave you an emotional response.
- The resolution must be set to 1920×1080 and 30fps. Anything less will not be accepted. Please do not send UHD or QHD 2k and 4k videos. Increasing your resolution will increase the file size exponentially. Go into the camera/video settings of your device to change your settings, especially if you have a newer camera. Please only upload if you are wired or on WiFi and have made sure that the file is less than 2GB, although normal HD videos with this resolution are less than 1GB.
- Your upload speed needs to be at least 10 Mbps. If you would like to see what your current download and upload speed is, you can do so at Speedtest. This is incredibly important, because you do not want to spend hours uploading a file.
- You must take your video with the phone in an upright portrait or box position. No wide-screen. On mobile, do not tilt the phone to record this video. If your video is in a tilted/landscape position, it will not be approved
- No background noises — noise must be controlled within 15dB SNR (this should be fine if you film alone in the room) Traffic/AC/fan sounds would cause your video to be invalid.
- Good lighting on face, with little to no shadows.
- You need to look into the camera when videoing. Your eyes need to be clearly visible. The phone needs to be set at eye-level for proper facial recognition. Turning your head away a lot could cause it to be rejected.
- No vibrations or movement of recording device. Make sure it is fixed in a location and remains stable throughout the filming process. Tripods are best for this, but anything that keeps your phone steady will be fine.
- Face and torso should occupy 65-80% of the screen.
- Record for at least 20 minutes. Speak normally. Natural pauses are okay. Try to keep pauses under 10 seconds and no longer than 30 seconds.
- At least 200 sentences must be spoken.
To upload your video, please complete this form!