 Poor Voice Quality
Poor voice quality (usually heard as
stuttering or breaks in the speech) is most often caused by
insufficient network bandwidth. That is, there just isn't enough
capacity on your internet connection to handle a VoIP call. This
could just be a temporary effect (for example, someone is performing
a large download on the same internet connection) in which case the
voice quality should improve.
Using the Phone Settings form which is
available from the Options menu (an example screen shot is
shown to the right) there are a number of steps that can be
used to minimise these problems:
Silence suppression. Put simply,
this does not transmit the periods in which you are silent –
on average this reduces the transmitting bandwidth use by 50%.
Silence Suppression can be enabled by checking the option on
the Options->Phone Settings... form.
Codec selection. Articulation
supports two codecs G.711 and GSM; these are methods for encoding
the voice while it is transmitted over the internet/network. G.711
is the encoding used by the normal telephone network and GSM is most
often used on the mobile telephone network. While G.711 provides
better quality voice, it uses much more bandwidth than GSM (almost
four times more). The benefits of the reduced bandwidth usage often,
therefore, provide a better call quality with GSM. To modify the
codec selection use the Options->Phone Settings... form.
With both GSM and G711 selected Articulation and your VoIP provider
will attempt to negotiate the encoding to use.
Use the codec selection table to help you select a suitable codec for your network.
The Last Call... option from the Call menu may provide you with more information regarding the voice quality:
- Received - This shows the number of voice packets that have
been received (each packet represents 1/50th of a second of voice).
If there are a large number shown as missing, when compared with the
number received, then you may hear gaps in the speech.
There is nothing Articulation can do about this; these missing voice
packets have been lost on the network between you and the remote party.
- Max Delay - This shows the maximum delay between when a voice
packet was expected from the remote party and when it actually arrived.
If this figure is more than 200ms you may experience breaks in the
speech. If the remote party is using silence suppression this figure
may show large values.
- Sent - This shows the number of voice packets that Articulation
has sent. If there were errors sending the voice there will be a
figure following in brackets (with an error code); this indicates that
some packets could not be sent possibly due to network interference.
This will cause breaks in the voice at the remote party end.
- Overruns - These are caused when too much voice arrives at once.
This is normally caused by a 'bursty' network and is often accompanied
by a high Max Delay. If the voice packets get queued up
temporarily somewhere on the network and then all get released at the
same time this will cause overruns. This may result in a gap in the
voice followed by 'speeded up' voice.
- Underruns - This is caused when there is too big a delay between
voice packets and Articulation has run out of data to send to the
speaker. This can cause gaps in the voice. This may also be caused
by silence suppression, although this is not an error in this case.
If you have tried the steps under this section and the voice is still poor,
then your network may not be capable of providing a reliable VoIP service.
This may often be the case on cell phone wireless (GPRS/EDGE/EVDO) - these technologies have not been designed with real time communication in mind.
Note that you may still be able to use streaming applications (such as TV/Internet Radio) since these are able to use a different technique for transmission.
Streaming applications buffer an amount of the media before playing it which allows them to fix the problems described above.
If a VoIP application were to do this, the quality would be better, but there would be a couple of
seconds delay between speaking and the other end hearing what was spoken.
|
