Conectronica Revista Maganize industrial de electronica

The voice over Protocol Internet (Voice to over IP, VoIP) is extending by a wide form as alternative of low cost for the users in international telephone calls of long distance. Nevertheless, for the customary users to the high quality of audio that offers the service of the Exchanged Public Wire net (PSTN, Public Switched Telephone Network), the quality of VoIP leaves generally damaged in comparison.

The quality of voice in the VoIP networks varies much as a result of several factors. This article will show how the suppliers of services can avoid degradation in a network VoIP and provide a quality of comparable voice, or even superior, at the PSTN levels. In the PSTN, the quality of voice he is intelligible, it sounds natural, it allows that the users identify the interlocutors and undergoes only small misalignments that are annoying. The voice is provided like an analogical signal, that it finalizes in the Operator of Digital Curl (Digital Loop Carrier, DLC) either Central office (Central Office, CO) the more near. The DLC analogical convert a signal samples digitalises for communication long-distance, and next in the end point another DLC with it again spills the digital samples to an analogical format in telephone interface TIP-RING. The voice of high quality is obtained because the analogical signals of voice are sampled to 8000 only 8 Hertz and they are compressed to bit by sample having used the Modulation of Code of Pulse ITU-T-G711.

Several factors that allow the PSTN to maintain the quality of the voice, among them exist:

- Samples and compression of voice: The transmission of the original DLC to the DLC of destiny is digital. The digital transmissions with G711 are controlled with layer clocks that assure the synchronism the provided sample. Rarely losses happen or errors in the compressed sample and, therefore, are not perceived by the users.

- Retardations: The retardations of end to end must challenging them them of physical transmission in the PSTN. There is memory nor process of no packages in the transmission, which does that the conversation of interactive voice is more natural and pleasant.

- Echo: Although the same analogical telephones in calls are used on PSTN and VoIP, the echo in service PSTN is not perceivable due to the minimum retardations. The balance of the losses of end to end in the DLC and the CO of the PSTN helps to reduce the disturbances by echo. For international calls and of long distance where retardations can be introduced, echo erasers on the part of the operator are gotten up long-distance.

- Conformity to TR-57: TR-57 [1] is the North American standard for the transmission directives and commutation in DLC. The systems and the analogical stages of entrance of DLC that control telephone cables TIP-RING fulfill TR-57 and the equivalent characteristics of transmission in each country, which optimizes the quality of transmission. The DLC and the CO of PSTN also assure reduced retardations, which gives like result one better perception of the beginning and the completions of the call, as well as the use of the services of call.

- Control of Equivalent Number of Call (Ring Equivalent Number, REN) and impedance matching: The DLC are designed habitually to offer support up to three to five telephones with the same physical wiring TIP-RING. This allows that the PSTN maintains the quality in spite of the simultaneous use of the telephone and to diminish the disturbance due to the impedance misalignment. The PSTN is not completely free of failures. In some international calls the quality of voice can be seen affected for different reasons. The causes could be the result of unsuitable completions, retardations of the connections by satellite and several stages of transcodificación G711, deficiency of the suitable canceladores of echo in the completion nodes, lengths local curls, use of several telephones with impedance misalignments and use of intermediate connectivity based on packages between the nodes of long distance. When it is compared with PSTN, the quality of voice based on VoIP that is perceived sees influenced by several factors.

These factors - among them the retardation, echo, compression of voice, loss of packages, G711-PLC, transcodificación, losses in the footbridge and losses of Connection of Terminal - are classified following standard TIA-TSB 116-A [to 2]. Voip1En the architecture VoIP, the Equipment in the Facilities of the User (Customer Premises Equipment, CPE) includes services of voice like the DLC of the PSTN. But between the CPE of VoIP also functions of the CO of the PSTN are included, like generation of tone of call, detection of marked digits, initiation of the call establishment and functions of support. The voice VoIP is provided like packages through a network IP. Figure 1 indicates that there are several parameters that control the quality of voice and how network IP introduces some obstacles, like the fall of packages, fluctuation (to jitter), errors of package and fragmentation of these. The VoIP service also is provided through CPE in the same way that in the traditional VoIP adapters, residential footbridges of high range and special telephones IP.

The devices depend on the interfaces of the local network (LAN), like wireless Ethernet and LAN (WLAN) or networks of ample area (WAN) through a digital local line (DSL). The bandwidth available in these interfaces of network, as well as the architecture of the device and the incorporated mechanisms of QoS, are united to help to reduce the retardations of packages between ends. Many devices CPE manage the QoS in ascending sense, whereas the QoS of slope is managed through Supplier of Services of Internet (Internet Service Provider, ISP). The CO of the route of the data, as the Multiplexers of Access of Digital Line of Abonado (DSLAM) VDSL are available already like support to QoS, which aid to eliminate the falls of VoIP packages in completions IP. The CPE of VoIP are designed considering that hardware allows to reduce the cost of the devices in order that the service is economic for the consumers. For example, the clock references cannot be equaled to the clocks of PSTN layer and some of the devices in the stages of entrance of the CPE cannot fulfill specifications TR-57 totally. Nevertheless, several manufacturers have recently presented/displayed devices of telephone interface that low cost that incorporates characteristics TR-57, which allows one better quality of voice. Still multiple impedances, tones, functions of deviation of call must be gotten up and fits thus planned of losses for each country, adapting to the local values of sonorous level to improve the quality of voice as a whole. In some countries, the services of Internet and VoIP provide different suppliers. In order to reduce the expenses, the subscribers can select the service of Internet with smaller bandwidth. In order to surpass these limitations in the bandwidth, the suppliers of VoIP use codecs like G729A instead of G711. The compression of voice G729A is eight times superior to the one of G711, which gives like result a sound of less quality. In addition, the calls of VoIP can cross several footbridges and transcodificadores, or conversions of a system from compression to another one, to arrive to the end users in others re- DES VoIP, the PSTN or the networks ina- lámbricas. Due to the network congestion, the fall of packages is inevitable. The algorithms of Concealment of Losses of Packages (Packet Loss Concealment, PLC) are gotten up to the CPE to manage the unexpected falls of packages. During the periods of silence, buffers of fluctuation (to jitter) adjusts to optimize the packages available from the network interface and to diminish the buffer retardations. This activity also improves the difficulties that present/display the resulting packages of drifts of the clock of end to end. The echo is a critical problem that influences in the quality of voice. The sonorous retardations of call of VoIP of end to end and levels also add echo. A retardation inferior helps to improve the echoes. The retardations can improve with a greater bandwidth, the mechanisms of QoS and the CPE designed for smaller retardations of process of packages. The sonorous level settles down mainly by the telephones of the end user, the impedances and losses of end to end and in the country in which it has settled. The quality of voice VoIP can be improved canceladores more incorporating of the operator in the process of the CPE. In the VoIP networks, the quality of voice must be monitored continuously. The software of monitoring of Quality of Voice (VQmon) is used in several facilities, but all the devices CPE cannot provide an improvement of the dynamic quality and cradle in the monitored parameters of. In order to reach the highest quality, devices CPE and the unfolded infrastructure of end to end must be able to offer support to a dynamic improvement of the quality based on the monitored parameters. Since it has been explained in this article has detailed, a series of improvements is being carried out to assure a quality comparable voice to the one of the PSTN from the point of view of the experience of the users when these use VoIP services. Nevertheless, it is possible that the quality of VoIP can surpass to the one of PSTN when broadband voice is used. In the traditional services of PSTN and VoIP, the telephone acoustics and the process limit the frequencies of voice of 300 3400 Hertz. In the broadband voice, the acoustic interfaces and the process support frequencies of 50 to 7000 Hertz, which allows to offer conversation with natural sound with presence sensation. The audio broadband tablet with the bandwidth with códec G722 needs the same bandwidth that G711. Others broadband codecs as 3 G729EV [] work with half of bandwidth that G711. The perception obtained through some codecs often surpasses to the quality of the PSTN of G711.

The quality level of voice `R' of the voice in broadband is around a 20% better than G711 based on updates of a model with a bandwidth of model R following the recommendations ITU-T-G107 and 4.5 G113 []. The voice in broadband also is unfolding in Europe, with the support of broadband telephones DECT (Digital Enhanced Cordless Telecommunications). These facilities are based on the hardware of the entrance stage in broadband and the acoustics. The stage of telephone entrance connected to the Circuit of Interface of the Local line and the devices of hardware sampling provide an ample bandwidth. At present several models of telephone for conferences, telephones by software in PC and telephones IP exist that offer support to voice in broadband. The acoustic interfaces on digital telephones, Bluetooth, DECT and WiFi also can use for services of voice of they go away it gives widens. A series is making sure improvements to assure that the quality of voice is comparable to the one of PSTN when VoIP services are used, and that the quality of VoIP can surpass at the PSTN levels when they are used broadband codecs. In the traditional services of PSTN and VoIP, the telephone acoustics and the processing limit the frequencies of voice of 300 3400 Hertz. In the voice for broadband, the acoustic interfaces and the process support frequencies of 50 to 7000 Hertz, that allow to offer the conversation with the natural sound with presence sensation. The voice compressed in broadband needs less bandwidth than G711. The perceptions of listening of many broadband codecs surpass to those of the quality of G711 for the PSTN. The quality level of voice `R' of the voice in broadband is approximately a 20% better than G711 based on updates of a model with a bandwidth of model R following the recommendations ITU-T-G107 and G113.