Audio Results Explained
Let me start the answer to this question with some definitions. This will help with the following text. We have 4 loudness levels (-20, -30, -40, -50 db). We also have 5 frequencies (500, 1000, 2000, 4000, 8000 Hz). We assume two ears. Let me call a 'tone' a unique combination of loudness, frequency, and ear presentation.
Each tone is presented 3 times. That makes for a total of 4 * 5 * 2 * 3 = 120 tone presentations in any game. Each game is divided into 5 rounds, thus there are 24 tones presented in a round. Tones are uniquely randomized each time the game is played, so they will always be presented in a different order.
The first computation I perform is a 2 out of 3 for each tone. A tone is considered to pass if the user touches the screen within 1.5 seconds of a tone being presented to them. The user is considered able to hear a specific tone if they can do this 2 out of the 3 times a tone is presented.
In the "Ear Summary" section of the Audio Result, you will see something like Left 13/20 and Right 14/20. This means to say that the user was able to hear 13 of the 20 unique freq/loudness combinations presented to the left ear at least 2 out of the 3 times they were presented. Similarly the user was able to hear 14 of the freq/loudness combinations presented to the right ear at least 2 out of the three times each combination was presented.
From this information I determine which is the 'stronger' ear, in this case the 'right' ear. Then in the "Ear Comparison" section of the report, I simply sum the number of passes for each frequency at all loudness levels which were presented to the user. Each frequency is presented to each ear at 4 loudness levels 3 times for a total of 12 samples. If the 'weaker' ear passes a particular freq test at least 80% as often as the 'stronger' ear, it is considered passing, otherwise it fails.
Finally, I show the most detailed results data in the 'Audio Test' section of the report. Here I show the threshold loudness to be between the lowest passing loudness (2 out of 3 passes) and the loudness just below that, which failed.
This is kind of a long winded answer, but accurate :-)