#TITLE_ALTERNATIVE#

<p align="justify">Speech's quality in conversations using VoIP technology is really dependent to the condition and the characteristics of IP network that is used. To establish the strategic implementation, we should have the knowledge concerning the IP network's characteri...

Full description

Saved in:
Bibliographic Details
Main Author: (NIM 13203003), KURNIAWAN
Format: Final Project
Language:Indonesia
Online Access:https://digilib.itb.ac.id/gdl/view/10785
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Institut Teknologi Bandung
Language: Indonesia
Description
Summary:<p align="justify">Speech's quality in conversations using VoIP technology is really dependent to the condition and the characteristics of IP network that is used. To establish the strategic implementation, we should have the knowledge concerning the IP network's characteristic, such as delay, jitter, bandwidth limitations, congestion, and packet loss, to qualify the Quality of Service (QoS) requirements. One way to determine its attribute is by means of network emulator. Nowadays, NIST Net Emulator, an open-source software to emulate the IP network and its characteristics, is already presented.<p align="justify"><p>This final assignment report discuss briefly about the study of IP network's characteristic which affects the implementation of an VoIP system. This system implemented based on three categories of protocol, i.e. SIP, H323 and IAX; using three types of speech codecs, which are G.711, GSM, and iLBC.<p align="justify"><p>Here, research is done by sending some speech signals which has been recorded while passing them through certained IP network, of which characteristic determined using the NIST Net, followed by recording the obtained speech. The result of these speech signals are then compared with the original speech signal quantitatively and objectively using Perceptual Evaluation for Speech Quality (PESQ).<p align="justify"><p>The outline of the IP network's structure implemented in this research there are four PC used here: two PC used as clients, one used as a NIST Net server and router, also the other one used as an Asterisk server which functions as an IP-PBX. The speech quality measured here is an end-to-end point, i.e. from the source of speech workstation to its destination.<p align="justify"><p>By altering some parameters affecting the characteristic of the examined IP network's structure, we are able to construct relational graph among its parameters, VoIP network protocol, and speech compression used regarding PESQ as well as its correlated coefficients.<p align="justify"><p>The research indicates that packet loss boundary resulting acceptable speech quality vary from +2% (IAX, iLBC) up to +13% (SIP, u-law); as for jitter, without using jitter buffer, vary between +9.5ms (SIP, GSM) and +15ms (IAX, u-law). Using jitter buffer, jitter which is still acceptable for speech quality can reach up to +45ms (H.323, a-law). Quantitative data resulted here can be used as references or comparators to establish one way speech’s quality with Indonesian language and a constrain of IP network's parameters recommendation where VoIP communication technologies are to be used appropriately. <br />