Gecko fixing missing audio

Gecko Fixing Missed Audio

The issue: Gecko’s algorithm does not detect every single word from the audio. Most commonly, the end of the recording. Example: The last clip of the audio was not detected by Gecko.

How to Fix the Missed Audio

• Highlight the missed audio

• Click the box next to Speaker on the bottom left corner in the box labelled Segment Labelling. The audio should be highlighted purple.

• Fill in the words in the box where the transcript appears

• Save this file as a JSON file and open it on your computer

• Reopen Gecko and load the new JSON file

• You should have this JSON file opened on both your computer and in Gecko

• Scroll to the missed audio section in the JSON file on your computer

• Sometimes the code will be written as:

{“text”:“_______”,“start”:_ _ _._ _,“end”:_ _ _._ _,“type”:“WORD”}, format

but it should be as:

{"start":_ _ _._ _,"end":_ _ _._ _,"text":"_______","type":"WORD"}, format

• If it is written in the wrong format, you must change it into the correct format Wrong {“text”:“_______”,“start”:_ _ _._ _,“end”:_ _ _._ _,“type”:“WORD”},

Correct {"start":_ _ _._ _,"end":_ _ _._ _,"text":"_______","type":"WORD"},

• “start”: _ _ . _ = time the audio starts
• “end”: _ _ . _ = time the audio ends
• _______ = the word that is being said
• If the time is wrong, the word will be highlighted at the wrong time when the audio plays
• Fix the audio word by word by dragging the audio back and forth while listening to when the word starts and ends.\

Example: Looking for when the word “because” starts and ends

• When you highlighted the audio, look at the time stamp in the box at the bottom left corner

Example: timestamp starts at 286.76 and ends at 287.07

• Input that time stamp into the JSON file opened on your computer

Continue these steps with all the words to correct all the time stamps Save and reupload onto Gecko to ensure all the words get highlighted at the right time