I'd like to hear what people think of this. Please suggest
improvements.
----------------------------------------------------------------------------------
How to measure one-way audio latency between two phones
----------------------------------------------------------------------------------
This method scales up an interval that’s too small for a person
without fancy equipment to measure accurately (such as 353 ms)
to an interval that’s much easier to measure accurately (30 sec)
and then divide to obtain a fairly accurate latency result. It's
intended for any path: VOIP to POTS, cell to POTS, VOIP to VOIP,
and so on.
You’ll need:
- two phones
- a pen
- a watch or clock with a seconds readout
Both phones must be with you. Establish a call between the two
phones. Use the pen to tap the microphone of phone A, and make
sure that the transmitted tap is easily heard on the earpiece of
phone B. Don’t use the speakerphone on phone B because the tap
noise from there might feed back into the nearby microphone of
phone A, gumming up the test.
Here’s the test. Tap with the pen repeatedly and adjust the tapping
rate so that you’re tapping the microphone just as the sound of your
immediately previous tap is arriving at the earpiece. (That might
seem tricky, but the rhythm establishes itself easily.) Once you have
the hang of it, get a rhythm going and count the taps for 20 or 30
seconds, then stop.
Divide the test’s duration by the number of taps. For example, for a
30-second test with 85 taps:
30 seconds divided by 85 taps = 353 milliseconds per tap
So the one-way audio latency between the two phones is about 353
milliseconds.
↧