Properties and Requirements for Voip

14.3
This section illustrates the characteristics of VoIP traffic by describing the properties and functionality of the used voice codec in HSPA/LTE (Section 14.3.1) and by presenting the requirements and the used quality criteria for VoIP traffic in Sections 14.3.2 and 14.3.3, respectively.
14.3.1

Adaptive Multirate (AMR)

Adaptive Multirate (AMR) is an audio data compression scheme optimized for speech coding. AMR was adopted as the standard speech codec by 3GPP in October 1998 and is now widely used. The AMR speech codec consists of multirate speech codec, a source controlled rate scheme including a voice activity detector, a comfort noise generation system, and an error concealment mechanism to combat the effects of transmission errors and lost packets.
The multirate speech codec is a single integrated speech codec with eight source rates from 4.75 kbps to 12.2 kbps, and a low rate background noise encoding mode. For AMR the sampling rate is 8 kHz. The usage of AMR requires optimized link adaptation that selects the best codec mode to meet the local radio channel and capacity requirements. If the radio conditions are bad, source coding is reduced and channel coding is increased.
In addition to AMR audio codec, an extension of it is AMR-wideband (AMR-WB) audio codec. Sampling rate for AMR-WB is 16 kHz, which is double the sampling rate of AMR. Therefore, AMR is often abbreviated as AMR-NB (narrowband). AMR-WB is supported if 16 kbps sampling for the audio is used in the UE. Similarly, like AMR-NB, AMR-WB operates with various bit rates. For AMR-WB the bit rates range from 6.60 kbps to 23.85 kbps. It is emphasized that both AMR-NB and AMR-WB are used in HSPA and LTE radio systems. A more detailed description of AMR-NB and AMR-WB can be found in [10].
General functionality of the AMR codec is illustrated in the Figure 14.4. During voice activity periods there is one voice frame generated every 20 ms (number of bits/frame

FIGURE 14.4 Illustration of VoIP traffic.
depending on the AMR codec mode), whereas during the silent periods Silence Descriptor (SID) frames are generated once in every 160 ms.
14.3.2

Delay Requirements

Voice over Internet Protocol (VoIP) is a conversational class service and the packet delay should be strictly maintained under reasonable limits. The maximum acceptable mouth-to-ear delay for voice is in the order of 250 ms, as illustrated in the Figure 14.5. Assuming that the delay for core network is approximately 100 ms, the tolerable delay for MAC buffering/ scheduling and detection should be strictly below 150 ms. Hence, assuming that both end users are (E-)UTRAN users, the tolerable one-way delay for MAC buffering and scheduling should be under 80 ms, which is the used air interface delay for HSPA. To improve voice quality further on LTE, tolerable air interface delay was reduced to 50 ms for LTE [11]. For LTE FDD the average HARQ RTT equals 8 ms implying that at most 6 HARQ retransmissions are allowed for a VoIP packet when air-interface-delay of 50 ms is used. For HSPA the HARQ RTT would be 12 ms in downlink and 16 ms in uplink with 2 ms TTI, which means at most 6 HARQ retransmissions in downlink and 4 HARQ retransmissions in uplink with 80 ms HSPA delay budget. Similarly, at most 1 HARQ retransmission could be allowed in the uplink, corresponding to 40 ms HARQ RTT with 10 ms TTI.
14.3.3

Quality Criteria

Considering the nature of radio communication it is not practical to aim for 100% reception of all the VoIP packets in time. Instead, certain degree of missing packets can be tolerated without notably affecting the QoS perceived by the users. For VoIP traffic, the system capacity is measured as the maximum number of users who could be supported without exceeding 5% outage. During the standardization phases of HSPA and LTE in 3GPP,

FIGURE 14.5 E-Model rating as a function of mouth-to-ear delay (ms) [8].
slightly different outage criteria were used when evaluating VoIP system performance. Used outage criteria are defined according to system as follows: For HSPA, a user is in outage if more than 5% of the packets are not received within the delay budget when monitored over a 10-second window. For LTE [11], an outage is counted if more than 2% of the packets are not received within the delay budget when monitored over the whole call. It is emphasized that even though the used outage criteria differ slightly between HSPA and LTE, the performance difference originating from different outage criteria is rather small being inside the error margins of the simulations.
In UL direction there also exists an alternative method to measure VoIP performance, which is used in the context of this topic for HSPA. That is, UL VoIP capacity can be measured as the allowed number of users in a cell with an average noise rise of X dB measured at the base station. Optimal value for X may depend on the used system, and for HSPA the used value was 6 dB.