Enter your first and last name to begin. If you only use one name, fill the first box and leave the second empty.
Thank you so much for your time to help us in this project! Please read the instructions below carefully before you begin.
Assume you are explaining a video of a scientific experiment to a grad student with visual impairment who cannot see the scene, but needs to know the critical steps. You should describe everything in a way that allows the student to reconstruct the experiment.
Your explanation should be complete, structured, and sequential, as if you are narrating what is happening in real time.
Walk the student through the procedure the same way you'd walk them through it in person. For every step, try to convey:
Only describe what you can actually see in the video. Please don't fill things in from your own knowledge of how the protocol "should" go, even if you're confident. We want the annotation to reflect the video, not the textbook.
A few cases come up often:
[Not Visible Parameter] where that value would appear in the sentence.[Not visible] in the step and describe what you can see.Each video is divided into segments (from the spreadsheet next to the file when available; otherwise fixed-length chunks). For every segment you complete the four panels on the right. In the downloaded or submitted JSON, that becomes one object per segment with (among other metadata) these fields:
Optional Extra_Notes: Click the pink Extra Note button next to Replay at the exact moment you care about. A popup opens (same purple header as the Instructions window); the tool records the video time at that click and pauses playback. Type your remark and choose Save note. Entries belong to the current segment only. In JSON they appear under Extra_Notes: each key is a time string in m:ss form (for example "1:05"), and each key maps to a list of note strings. You can save multiple notes at the same timestamp or at many different times. Remove notes from the list inside the popup if needed.
Special labels (type them exactly so automated checks can find them):
[Not visible]: this field might apply but you cannot clearly see the relevant detail, or (in Steps) the action is ambiguous. Say what you can see; in a step you may write [Not visible: adds a white powder from a spatula].N/A: this field does not apply to this segment.[Not Visible Parameter]: use in Steps when a number or unit belongs in the sentence but is not visible in the video (duration, rpm, temperature, volume, and similar). Put the tag exactly where that value would appear, as in section 3.videos/<name>.xlsx or .csv) with start and end times in seconds per step; if neither file is present, the app uses fixed 30-second segments.Extra_Notes for the current segment (see Output Format).Full reference: same content as on the welcome screen. Reopen anytime from Instructions in the top bar.
Assume you are explaining a video of a scientific experiment to a grad student with visual impairment who cannot see the scene, but needs to know the critical steps. You should describe everything in a way that allows the student to reconstruct the experiment.
Your explanation should be complete, structured, and sequential, as if you are narrating what is happening in real time.
Walk the student through the procedure the same way you'd walk them through it in person. For every step, try to convey:
Only describe what you can actually see in the video. Please don't fill things in from your own knowledge of how the protocol "should" go, even if you're confident. We want the annotation to reflect the video, not the textbook.
A few cases come up often:
[Not Visible Parameter] where that value would appear in the sentence.[Not visible] in the step and describe what you can see.Each video is divided into segments (from the spreadsheet next to the file when available; otherwise fixed-length chunks). For every segment you complete the four panels on the right. In the downloaded or submitted JSON, that becomes one object per segment with (among other metadata) these fields:
Optional Extra_Notes: Click the pink Extra Note button next to Replay at the exact moment you care about. A popup opens (same purple header as the Instructions window); the tool records the video time at that click and pauses playback. Type your remark and choose Save note. Entries belong to the current segment only. In JSON they appear under Extra_Notes: each key is a time string in m:ss form (for example "1:05"), and each key maps to a list of note strings. You can save multiple notes at the same timestamp or at many different times. Remove notes from the list inside the popup if needed.
Special labels (type them exactly so automated checks can find them):
[Not visible]: this field might apply but you cannot clearly see the relevant detail, or (in Steps) the action is ambiguous. Say what you can see; in a step you may write [Not visible: adds a white powder from a spatula].N/A: this field does not apply to this segment.[Not Visible Parameter]: use in Steps when a number or unit belongs in the sentence but is not visible in the video (duration, rpm, temperature, volume, and similar). Put the tag exactly where that value would appear, as in section 3.videos/<name>.xlsx or .csv) with start and end times in seconds per step; if neither file is present, the app uses fixed 30-second segments.Extra_Notes for the current segment (see Output Format).The note is saved at the video time shown below (the time when you opened this window).
Time in video: 0:00
Here are a few instruction hints to keep in mind: