|
0:00:15
|
So now let’s take a look at troubleshooting some in point issues.
|
|
0:00:21
|
So first let’s just take a look at defining some issues with
|
|
0:00:24
|
voice calls and what we want to do here is basically go
|
|
0:00:26
|
some of the definitions of some of the types of
|
|
0:00:31
|
problems that users, either out on the PSTN reporting
|
|
0:00:35
|
to the people that they're talking to internal,
|
|
0:00:38
|
to CUCM that run IP Phones are experiencing
|
|
0:00:42
|
PSTN users experiencing or else
|
|
0:00:44
|
issues that the IP phone users themselves are experiencing
|
|
0:00:47
|
and we wanted to find some of these and take a look at some of the
|
|
0:00:51
|
general problems and what these might be referred or called
|
|
0:00:55
|
and that way throughout the rest of the videos as we’re
|
|
0:00:59
|
taking a look at all the configuration and hands on,
|
|
0:01:04
|
we can begin to see how we might
|
|
0:01:07
|
troubleshoot and deal with these issues
|
|
0:01:09
|
and then of course there will be a lot more
|
|
0:01:12
|
that we’ll go over
|
|
0:01:14
|
in additional slides if you're watching this for or the CCNP Voice.
|
|
0:01:21
|
For CCNA we’ll just take a look at these
|
|
0:01:24
|
as a general overview and then
|
|
0:01:27
|
move on to al the hands on where we’ll take some more
|
|
0:01:32
|
look at these. So first of all,
|
|
0:01:35
|
we have just defined as general noise, it could be silence or clicking,
|
|
0:01:40
|
crackling, or crosstalk where
|
|
0:01:43
|
we actually hear part of someone else’s conversation,
|
|
0:01:45
|
so this is not the same as echo
|
|
0:01:47
|
where we hear our own words or the caller on the PSTN hears their words
|
|
0:01:51
|
repeated but we’re actually hearing someone else’s conversation,
|
|
0:01:56
|
could be sort of a hissing static or
|
|
0:01:59
|
a low frequency buzz or hum.
|
|
0:02:04
|
OK? So that would be defined under noise.
|
|
0:02:07
|
Under echo, obviously this would be repeated
|
|
0:02:10
|
voice, so I would hear myself or the PSTN caller would hear themselves,
|
|
0:02:15
|
or it could also possibly sound like a tunnel voice,
|
|
0:02:18
|
so kind of sound like I'm in a tunnel,
|
|
0:02:21
|
and you can’t hear what I'm saying,
|
|
0:02:22
|
but it’s really right difficult to tell what I'm saying.
|
|
0:02:26
|
OK?
|
|
0:02:27
|
Could have garbled voice,
|
|
0:02:30
|
something like that
|
|
0:02:33
|
or possibly sound like it’s underwater and
|
|
0:02:36
|
I can’t really make that sound that well, but
|
|
0:02:38
|
you know sort of an underwater sound,
|
|
0:02:40
|
almost like there's water in someone’s mouth
|
|
0:02:43
|
or they’re robotic voice
|
|
0:02:46
|
where there are seemingly buckets dropping.
|
|
0:02:49
|
OK? Something we would define those are garbled.
|
|
0:02:52
|
Also possibly volume distortion, so we might have a really
|
|
0:02:56
|
high frequency tiny sound
|
|
0:02:58
|
to someone’s voice or someone’s voice might be really soft
|
|
0:03:04
|
or it might be kind of fuzzy which is kind of hard to describe
|
|
0:03:07
|
or even a bit of a muffled voice
|
|
0:03:09
|
so you can hear what they're saying, but it sounds muscled.
|
|
0:03:12
|
OK? So those would be defined as volume distortion.
|
|
0:03:15
|
And then another issue, one way audio.
|
|
0:03:18
|
OK so
|
|
0:03:20
|
I hear you,
|
|
0:03:22
|
but you don’t hear me or vice versa,
|
|
0:03:25
|
you hear me, but I can’t hear anything you're saying.
|
|
0:03:31
|
So looking at noise and breaking it down a little bit further,
|
|
0:03:34
|
silence or hissing,
|
|
0:03:36
|
typically is caused by overaggressive VAD or
|
|
0:03:40
|
Voice Activity Detection
|
|
0:03:42
|
Clicking of sorts are typically caused by clock slips,
|
|
0:03:46
|
so a TDM clock on a digital circuit,
|
|
0:03:50
|
you know maybe we don’t have timing set up properly
|
|
0:03:53
|
or other TDM circuit airs.
|
|
0:03:55
|
Crackling, this could just be bad electrical connections, so
|
|
0:03:59
|
so maybe we have a TDM circuit, a PRI
|
|
0:04:01
|
and instead of being fully punched down
|
|
0:04:04
|
on a 66 or 110 block,
|
|
0:04:07
|
we have you know maybe a wire that was not fully punched down
|
|
0:04:12
|
and it’s kind of arching, so the signal sometimes gets through,
|
|
0:04:15
|
but it’s not quite perfect and this would actually
|
|
0:04:18
|
probably be less on a digital circuit, I probably shouldn’t have used that
|
|
0:04:21
|
as an example, but rather an analog circuit,
|
|
0:04:25
|
again punch down probably on a 66 or 110 block
|
|
0:04:29
|
or possibly at a work area outlet.
|
|
0:04:31
|
Cross talk of sorts, so this is not enough shielding,
|
|
0:04:35
|
so we have signals leaking from one circuit to another,
|
|
0:04:38
|
wires too close, but then again the wire should have
|
|
0:04:41
|
sufficient shielding, so maybe the shielding,
|
|
0:04:43
|
the case has been broke,
|
|
0:04:46
|
maybe they're even touching.
|
|
0:04:49
|
Static of some sort,
|
|
0:04:50
|
this is typically due to a codec mismatch which
|
|
0:04:53
|
can be worsened or exacerbated by VAD or Voice Activity Detection.
|
|
0:04:58
|
Remember for the most part? VAD is bad.
|
|
0:05:02
|
Echo,
|
|
0:05:03
|
so we need to know if it’s listener echo or talker echo.
|
|
0:05:07
|
Listener echo, the person listening hears the talker twice.
|
|
0:05:11
|
OK, so it could be talker echo where either the
|
|
0:05:14
|
party on the PTSN or
|
|
0:05:17
|
myself, you know the party internal to the IP Phone
|
|
0:05:19
|
hears their own voice, that’s talker echo in both cases,
|
|
0:05:23
|
it’s just a different direction of
|
|
0:05:25
|
who’s hearing the talker echo and that’s important as well,
|
|
0:05:30
|
so if you're
|
|
0:05:31
|
if you're gaining information from users, it’s important to
|
|
0:05:34
|
first of all differentiate between listener and talker
|
|
0:05:37
|
that is, am I hearing the listener twice
|
|
0:05:41
|
am I hearing the talker twice?
|
|
0:05:43
|
or are they hearing themselves twice?
|
|
0:05:46
|
If they are hearing themselves twice or I'm hearing myself twice,
|
|
0:05:50
|
that’s talker echo,
|
|
0:05:52
|
but if I'm hearing the other person talking twice,
|
|
0:05:55
|
or they're hearing me twice, that’s listener echo,
|
|
0:05:58
|
and then it’s also important to note who on the
|
|
0:06:00
|
circuit is hearing that
|
|
0:06:02
|
and this is typically caused by either
|
|
0:06:05
|
no ECAN or Echo Canceller
|
|
0:06:08
|
or potentially too short
|
|
0:06:11
|
of an Echo Canceller or too long of an echo tail, so really
|
|
0:06:15
|
these really are the same thing that I just said,
|
|
0:06:18
|
so
|
|
0:06:19
|
so the idea of echo is of course
|
|
0:06:22
|
that the transmit side
|
|
0:06:24
|
of the wire or communication is being linked over to the received
|
|
0:06:27
|
side if we’re dealing with talker echo
|
|
0:06:32
|
and
|
|
0:06:34
|
nd so what we do, is an ECAN or Echo Canceller,
|
|
0:06:38
|
basically sits there and buffers and delays
|
|
0:06:42
|
the speech just a little bit, you know maybe up to,
|
|
0:06:46
|
we can go out as long as we can find Echo Cancellers to
|
|
0:06:50
|
provision for this and sometimes you can even get
|
|
0:06:52
|
external third party ones if you have a really bad case of it,
|
|
0:06:55
|
obviously going too long gets you into a sort of a satellite
|
|
0:06:58
|
type phone communication where
|
|
0:07:00
|
you end up saying something and then you have to wait to,
|
|
0:07:03
|
for the other person to hear what you're saying
|
|
0:07:06
|
and then they talk and you end up stepping over each other words,
|
|
0:07:09
|
if the Echo Canceller is too long, but let’s say
|
|
0:07:12
|
an Echo Canceller of 32 milliseconds
|
|
0:07:14
|
or 64 milliseconds or even a 128 milliseconds,
|
|
0:07:18
|
they even have that go up to 256,
|
|
0:07:20
|
but however many milliseconds we’re buffering the voice,
|
|
0:07:23
|
and during that time, the Echo Canceller is taking a look
|
|
0:07:27
|
and it’s saying OK
|
|
0:07:30
|
I’m buffering the receive side
|
|
0:07:33
|
or I'm buffering the transmit side and I want to, or I'm sending it out,
|
|
0:07:37
|
but I'm also keeping that information in my
|
|
0:07:40
|
Echo Canceller buffer for a while to see if the signal that’s
|
|
0:07:42
|
that’s coming back on the receive side
|
|
0:07:45
|
sounds or really
|
|
0:07:48
|
from a wave form perspective,
|
|
0:07:50
|
you know sort of an siloscope waveform perspective, if that
|
|
0:07:54
|
waveform looks identical to
|
|
0:07:57
|
what was just transmitted 128 or 256 milliseconds ago or
|
|
0:08:01
|
64 milliseconds ago and if it does, we’re gonna cancel that out,
|
|
0:08:05
|
now obviously there are some things that send
|
|
0:08:08
|
and receive various similar waveforms like faxes and modems,
|
|
0:08:12
|
so Echo Cancellers are,
|
|
0:08:14
|
they will kill faxes and modems.
|
|
0:08:15
|
You should never use an Echo Canceller
|
|
0:08:18
|
or Voice Activity Detection or anything like that on a
|
|
0:08:22
|
fax or modem
|
|
0:08:24
|
connection, dial peer, or whatever,
|
|
0:08:30
|
but again that Echo Canceller, if the echo is longer,
|
|
0:08:36
|
so the actual echo tail, the actual
|
|
0:08:40
|
transmit is actually going out longer than the
|
|
0:08:43
|
Echo Canceller has buffered, so this is where I'm saying too long
|
|
0:08:46
|
of an echo tail, the actual echo,
|
|
0:08:48
|
or too short of an ECAN, the actual Echo Canceller is not
|
|
0:08:51
|
buffering long enough out,
|
|
0:08:53
|
then this is where echo will be still present
|
|
0:08:58
|
even if the Echo Canceller has been configured so
|
|
0:09:03
|
either enabling or replacing a defective echo canceller
|
|
0:09:06
|
or lengthening that Echo Canceller
|
|
0:09:10
|
ERL or Echo Return Loss.
|
|
0:09:13
|
OK so
|
|
0:09:15
|
So that can be a solution for that. Tunnel problem,
|
|
0:09:17
|
typically this is an echo problem but a very tight echo.
|
|
0:09:20
|
OK so
|
|
0:09:22
|
the reason it sounds kind of tunneled is because the echo is so
|
|
0:09:25
|
close that it’s almost imperceptible,
|
|
0:09:27
|
but it’s perceptible enough that the phenomenon,
|
|
0:09:31
|
the audible phenomenon actually
|
|
0:09:33
|
ends up sounding a bit like a tunnel.
|
|
0:09:37
|
Garbled voice, so if we have choppy voice
|
|
0:09:40
|
as one of the sub-functions or sub-definitions of garbled,
|
|
0:09:43
|
so again,
|
|
0:09:46
|
so this is typically packet loss or excessively delayed packets which
|
|
0:09:50
|
could be caused by an Echo Canceller, but also other things
|
|
0:09:54
|
such as jitter or the difference between latencies.
|
|
0:09:58
|
We could have synthetic or robotic voice,
|
|
0:10:01
|
this could just be maybe single packet loss or
|
|
0:10:04
|
every once in a while of packet loss or a packet
|
|
0:10:07
|
delay beyond the limit of the De-Jitter Buffer.
|
|
0:10:11
|
Underwater voice, which is very hard to mimic
|
|
0:10:14
|
is typically a G729 codec that’s being used
|
|
0:10:18
|
on both sides, but there's an annex codec mismatch,
|
|
0:10:21
|
so maybe G729 is used on one
|
|
0:10:24
|
and G729A is, or G729B
|
|
0:10:28
|
annex B is used on another.
|
|
0:10:32
|
Dealing with volume distortion,
|
|
0:10:34
|
let’s say we have fuzzy voice,
|
|
0:10:37
|
this could be on an analog circuit,
|
|
0:10:39
|
too much signal gain has been applied.
|
|
0:10:42
|
So we have the ability to gain
|
|
0:10:44
|
or attenuate a signal
|
|
0:10:47
|
loosely for the layman analogous to
|
|
0:10:52
|
gain, analogous to volume up,
|
|
0:10:54
|
attenuation analogous to volume down, not exactly the same.
|
|
0:10:59
|
OK.
|
|
0:11:00
|
Gain is really an amplification of the circuit or the signal in a way,
|
|
0:11:04
|
attenuation is the loss of a signal across a given medium,
|
|
0:11:09
|
but loosely saying, so we’re basically
|
|
0:11:11
|
turning the volume up essentially on a FXO port probably
|
|
0:11:15
|
too much.
|
|
0:11:16
|
Muffled voice, this can be
|
|
0:11:18
|
by an overdriven signal
|
|
0:11:20
|
possibly you have your hand over the handset or speaker phone mic
|
|
0:11:26
|
and so essentially what's happening is a reduction in the frequencies
|
|
0:11:30
|
and the key voice range
|
|
0:11:32
|
and the key voice range which is important to know is 440 hertz
|
|
0:11:36
|
up to 3500 hertz or 3.5 kilohertz, whichever way you want to say it,
|
|
0:11:41
|
but again, as I'm covering my mouth talking into the mic,
|
|
0:11:44
|
I'm still saying everything in the exact same way that I was right now,
|
|
0:11:48
|
but I have my hand off, and now I have my hand over. I'm just reducing
|
|
0:11:52
|
the frequencies that are being picked up,
|
|
0:11:54
|
whether by this recording mic or by a phone mic.
|
|
0:11:58
|
Soft voice, so this is just an attenuated signal.
|
|
0:12:02
|
Again layman’s terms, reduction in volume
|
|
0:12:05
|
probably an analog circuit, probably an FXO,
|
|
0:12:08
|
probably want to
|
|
0:12:12
|
not deal with you know or not deal, but
|
|
0:12:17
|
reverse whatever attenuation and the reason I'm looking for the right word,
|
|
0:12:23
|
is that when we’re dealing with, first of all when we’re dealing with
|
|
0:12:26
|
signals or audio signals,
|
|
0:12:29
|
if you're not already familiar, it’s important to note that
|
|
0:12:32
|
we never talk about positive numbers, well we do, but if we are,
|
|
0:12:36
|
we’re talking about over driving a signal,
|
|
0:12:38
|
so we’ll start and if you’ve looked at any
|
|
0:12:41
|
stereo amplifier receiver, then you’ve probably noticed this.
|
|
0:12:46
|
At least any
|
|
0:12:48
|
decently mid range to higher end one,
|
|
0:12:51
|
if you turn up the volume,
|
|
0:12:53
|
first of all the volume is almost always,
|
|
0:12:56
|
and if it’s not, then you probably have blown speakers,
|
|
0:12:59
|
but it’s almost always a negative number,
|
|
0:13:02
|
a negative integer. So you're seeing,
|
|
0:13:05
|
you know maybe -60, that’s a very very very low signal,
|
|
0:13:10
|
-30 is a relatively you know decent signal and these are in the unit
|
|
0:13:15
|
of measurement of decibels or dB,
|
|
0:13:20
|
and -0 or 0,
|
|
0:13:24
|
or also referred to as digital zero, also referred to as unity,
|
|
0:13:29
|
nothing to do with voicemail,
|
|
0:13:32
|
this is effectively the maximum that you can
|
|
0:13:37
|
have the signal for a given amplifier or
|
|
0:13:40
|
you know whatever without overdriving the signal.
|
|
0:13:43
|
Anytime you go past
|
|
0:13:45
|
digital zero into the positive number, you're overdriving a signal
|
|
0:13:51
|
and possibly overdriving an amplifier,
|
|
0:13:54
|
but the important thing to note here is that when we’re dealing with
|
|
0:13:56
|
attenuation
|
|
0:13:58
|
versus gain, if we are
|
|
0:14:03
|
ncreasing gain, then we want to basically go to lower numbers,
|
|
0:14:09
|
but again they are higher numbers, they are lower negative numbers.
|
|
0:14:12
|
So sort of a double negative there.
|
|
0:14:14
|
OK, attenuating the signal, we want to, I'm sorry
|
|
0:14:18
|
I said that backwards, we’re wanting to gain, we’re wanting to come to,
|
|
0:14:23
|
it is right, they are lower numbers. We’re wanting to go to like
|
|
0:14:27
|
you know -8 or -1 or -0.
|
|
0:14:33
|
I was thinking in my head of looking at it in a
|
|
0:14:38
|
vertical fashion,
|
|
0:14:39
|
but if we’re attenuating a signal, we’re trying to go -40, -60,
|
|
0:14:44
|
and really not that much
|
|
0:14:46
|
but we want to make sure that if we are attenuating a signal,
|
|
0:14:50
|
if we put let’s say an attenuation, let’s back up, if we put a gain
|
|
0:14:55
|
of -6,
|
|
0:14:58
|
that’s actually gonna be reduction in signal.
|
|
0:15:01
|
So if we put a gain of let’s say 3 then that’s,
|
|
0:15:07
|
we’re adding 3 decibels, if it was already at -20,
|
|
0:15:10
|
we’re now at -17 , because we added 3 decibels.
|
|
0:15:14
|
OK
|
|
0:15:15
|
If we are attenuating, we’re also wanting to put in a positive number
|
|
0:15:20
|
because
|
|
0:15:22
|
attenuation is the loss of a signal,
|
|
0:15:24
|
so as we’re configuring or Echo Cancellers on FXO
|
|
0:15:28
|
voice ports, we want to say attenuation, let’s say 10,
|
|
0:15:33
|
that’s gonna turn down the volume a lot, because we’re
|
|
0:15:35
|
attenuating or decreasing the decibel by 10
|
|
0:15:39
|
versus if we put attenuate
|
|
0:15:41
|
to -3, now we’re actually
|
|
0:15:44
|
doing the opposite, we’re actually increasing the volume.
|
|
0:15:47
|
OK? So be aware
|
|
0:15:49
|
of the voice terms that are used specific to
|
|
0:15:54
|
analogue voice waveforms really not to
|
|
0:15:58
|
unified communications in anyway and when you're configuring FXO ports.
|
|
0:16:04
|
OK, enough about that, so a tiny voice is typically caused
|
|
0:16:07
|
by an overdriven signal again, so again hand over the mic set
|
|
0:16:11
|
or something like that, but this is unfortunately this is
|
|
0:16:15
|
or not unfortunately, opposite to the muffled voice,
|
|
0:16:18
|
this is a reduction in the outside frequency.
|
|
0:16:20
|
So the key range staying the same,
|
|
0:16:23
|
we’re now at the tiny voice, we’re reducing the
|
|
0:16:27
|
outside frequencies,
|
|
0:16:31
|
we’re reducing the highs, the really highs, above 3500
|
|
0:16:35
|
and we’re reducing typically above,
|
|
0:16:38
|
you know 3500 and the muffled voice, we’re reducing the inside
|
|
0:16:43
|
frequency, so between those 4400 or 40 hertz and 3500 hertz.
|
|
0:16:52
|
So good to note and good to be aware of these types of
|
|
0:16:56
|
endpoint related issues and be on the lookout
|
|
0:17:00
|
and we’ll talk about how to mitigate them even more
|
|
0:17:03
|
than what we’ve talked about as time goes on.
|
