What thread is Ooma referring to? - audio delay fix (?)

[Bobby B]

For those users experiencing high delay during calls, it'd help us if you could measure how much delay you've been experiencing. Measuring this delay is somewhat technical - but it'd definitely help us further troubleshoot this problem.

Here's how to measure your delay:

1. Open a sound recording program on your computer. A free one that you can download is Audacity.
2. After verifying that your attached or built-in microphone is working, press the record button.
3. Bring your phone next your microphone or laptop, and dial test number 909-390-0003 using the speakerphone on your phone.
4. Speak a short word like "one" - you'll see the microphone pick up your word and the echo response.
5. Stop the recording, and then highlight the waveforms that contain your word and the echo response. Click the magnifying glass to zoom in on the area and give you a more precise timeline measurement.
6. Afterwards, measure between the start of the your word and the echo response. To make this measurement easier, you may delete any audio that was recorded prior to your word, so that time "0.00" aligns with the start of your word.

In the below example, there was approx 500ms of delay between your spoken word and the echo response.

After you take the measurement, please post your delay measurement to this thread.

[murphy]

I made three measurements.

Using my Panasonic cordless phone:
Telo 450 ms
Hub 500 ms

Using my Telo handset
600 ms

I have minor delay problems.
It's more noticeable on the hub than it is on the Telo.

[cmeisel]

The delay (cable line, 10mb upload and 30mb down speed) the delay with the Telo (which is plugged directly into the cable model) is always around 530-580, sometimes over 600. telo and telo handset

[southsound]

Using Rev: 1.41189

Telo handset: 539 ms / 609 ms / 530 ms

Uniden handset: 560 ms / 568 ms / 560 ms

Wired handset: 528 ms / 510 ms / 490 ms

Instead of saying a word, I clicked a pencil against the phone housing - this gave a very sharp waveform that allowed me to be pretty excact. I did this several times during each 90 second call. The numbers above are from the average clicks during 3 separate calls for each device. I am a little surpprised at the difference between several calls. I was more surprised that during the same call, the difference could be as much as 15 ms in a duration of about 90 seconds of clicking. The first click took the longest to return - others were a little bit shorter. Maybe part of the echo cancellation software? Call quality is great today - but the delay numbers don't look as good as the calls sounded.

In case it matters:
VoIP test statistics
--------------------
Jitter: you --> server: 0.1 ms
Jitter: server --> you: 3.5 ms
Packet loss: you --> server: 0.0 %
Packet loss: server --> you: 0.0 %
Packet discards: 0.0 %
Packets out of order: 0.0 %
Estimated MOS score: 4.0

Speed test statistics
---------------------
Download speed: 1286144 bps
Upload speed: 768176 bps
Download quality of service: 92 %
Upload quality of service: 99 %
Download test type: socket
Upload test type: socket
Maximum TCP delay: 186 ms
Average download pause: 10 ms
Minimum round trip time to server: 89 ms
Average round trip time to server: 90 ms
Estimated download bandwidth: 1286144bps
Route concurrency: 1.0
Download TCP forced idle: 0 %
Maximum route speed: 5890784bps

Totally unrelated trivia: Delay on my iPod Touch using Vonage Mobile is over 600 ms.

[gjwhitehorn]

Using my Telo through a cheap GE cordless phone, I got:

1st call: 1 iteration, ~600 ms delay
2nd call: 1 iteration, ~700 ms delay
3rd call: 1 iteration, ~800 ms delay
4th call: 8 iterations as follows -- 500, 500, 750, 750, 430, 700, 720, 730

And from the only computer wired through the Telo:

VoIP test statistics
--------------------
Jitter: you --> server: 0.5 ms
Jitter: server --> you: 2.5 ms
Packet loss: you --> server: 0.0 %
Packet loss: server --> you: 0.0 %
Packet discards: 0.0 %
Packets out of order: 0.0 %
Estimated MOS score: 4.1

Speed test statistics
---------------------
Download speed: 2277896 bps
Upload speed: 438000 bps
Download quality of service: 50 %
Upload quality of service: 99 %
Download test type: socket
Upload test type: socket
Maximum TCP delay: 437 ms
Average download pause: 5 ms
Minimum round trip time to server: 55 ms
Average round trip time to server: 56 ms
Estimated download bandwidth: 2880000bps
Route concurrency: 1.2643247
Download TCP forced idle: 0 %
Maximum route speed: 9532360bps

Also, in my modem I've forwarded all the Ooma ports to the Telo (which really didn't seem to make much difference).

Totally unrelated trivia: Delay on my iPod Touch using Vonage Mobile is over 600 ms.

I'd be interested to see delay times from your actual Vonage box.

[southsound]

Totally unrelated trivia: Delay on my iPod Touch using Vonage Mobile is over 600 ms.

I'd be interested to see delay times from your actual Vonage box.

Actually, I don't have Vonage - just the free Vonage iTouch app. It tells me that I have a $0.00 balance but I can still make calls with it. Funny thing is the caller ID - 88351XXX1577X18 - on some phones it shows up as that number for my name as well. The X's are really digits, but since it always shows up the same, I've replaced them with X's. Another forum user alerted us to the free Vonage thing - it keeps wanting me to add money to my account but I don't and it still works. I don't use it much because I don't do a lot of calling and have a great cell plan - but it is a fun way to call those who might not accept a call from one of my ooma numbers.

[murphy]

Can someone who has a landline with free long distance capability (It is a 900 number) make the measurement?
I'm wondering what the delay is on a copper circuit.
500 ms seems high by a factor of 10.

[tfk]

Thank you for the tip on how to more precisely measure the latency. OK, I tried two calls and measured: 575 ms and then 600 ms. My Telo is connected between the cable modem and the router. Using Comcast cable broadband from Lansing, MI. I tried traceroute to ooma proxy and get about 90 ms for that. My copper line is much faster, even though I have no speaker phone hooked up today to measure it.

I am a new user, I am on the fence about keeping it because of this delay. It would be helpful if Ooma would try to define a service level for this latency (i.e. maximum audio delay) the company would at least *try* to maintain for all its users.

I wonder if users closer to the west coast will have lower latency results?

[gjwhitehorn]

Can someone who has a landline with free long distance capability (It is a 900 number) make the measurement?
I'm wondering what the delay is on a copper circuit.
500 ms seems high by a factor of 10.

I called the echo number from my landline and was amazed at how fast the echo was.

It's almost instant -- too low to measure, at least with the method mentioned above. I'd say that's because the PSTN is a circuit-switched network, not packet-switched, like the internet. With packet-switched, every packet has to get routed individually, which takes a small bit of time I'm guessing.

Also, it's not a "900" number like you're probably thinking. 1-900 numbers, the ones that charge exorbitant fees for connecting, ONLY use the area code 900 -- not 901, 902, 903, or anything else -- and those fees would apply to anyone, even people with free long distance. My area code is 903. 909 is just another area code. And I'm curious, how do you not have free long distance? I thought most companies offered that. We're on at&t and ours is unlimited; we pay $35/month for the phone plan (before taxes and fees, of course).

[Bill D]

Hat's off to Bobby B's elegant delay test. After having 3 Hubs for over a year, I never was bothered by the delay until recently. Starting about a month ago, on my Hub in California, the delay seemed excessive on many calls, but got better about a week ago.

I just measured it and it's a consistent 410 ms (with a Uniden Handset). If I'm bothered by delay again, I'll measure it right after the call. I'll be in Florida Wednesday and I'll also measure my 2 hubs there to compare. I've not been bothered by delay there.

I assume the 909-390-0003 is a land line and the measurement test represents the delay from Ooma to a land line. Does anyone know if calling another VOIP from an Ooma line gives "double delay" compared to calling a land line from an Ooma line?

Bill