DL - 4G Americas

cross-talk suppression. The downlink control signaling does not indicate the presence of co-scheduled
UEs.
Figure B.10. UE-Specific Reference Signal Structure per Resource Block (RB).
For Rel-9 dual layer BF, the UE may feedback CQI back based on transmit diversity computed from the
common reference signals (CRS) and may not feedback a rank indicator. The transmit weights, MCS and
rank are computed at the eNodeB in this transmission mode. The UE is not aware of SU or MU-MIMO
transmission during decoding of PDSCH. The transmit weights (for both PDSCH and UE-specific
demodulation RS) are determined at the eNodeB based on covariance computed from either, SRS (for
TDD) or a long-term estimate of UL covariance (FDD).
B.2.8 VOCODER RATE ADAPTATION FOR LTE
The vocoder rate adaptation mechanism allows operators to be able to control the codec rate based on
load criteria. At peak hour there could be a desire to trade off some quality for additional capacity. The
purpose of this mechanism is to provide support to enable vocoder rate change in LTE networks, in
particular to let the UE select a more appropriate and radio resource friendly AMR encoder for VoIP.
Vocoder rate adaptation has also been extended to cover HSPA in Rel-10.
The vocoder rate adaptation mechanism relies on existing end-to-end schemes for Codec Rate
Adaptation (3GPP TS 26.114) to dynamically adapt an individual real-time media component to changing
conditions in the network. Those schemes are based on measurements performed on the receiving side
(e.g., packet loss, packet delay) that are reported back to the sending side via RTCP receiver reports. In
addition, the receiving side can use RTCP to explicitly control, for example, the codec rate, at the sending
side.
The key element of the vocoder rate adaptation mechanism is the adoption of the IP based Explicit
Congestion Notification (ECN) specified in [RFC3168]. The eNodeB can use ECN as an “early prewarning”
mechanism, (i.e. first set the “Congestion Experienced” (ECN-CE) codepoint in IP packets at
incipient congestion) and only start the dropping of packets on a bearer when congestion persists and/or
increases. The ECN-CE codepoint in an IP packet indicates congestion in the direction in which the IP
packets are being sent. The ECN-CE signal propagates to the receiving IP end-point and is made
available to the media/application layer receiver. The receiver can then send an application layer rate
reduction message (e.g., RTCP) to request a new send rate from the corresponding sender. Thereby, the
media/application layer should typically have sufficient reaction time (i.e. trigger a rate reduction before
packets need to be dropped at the bottleneck).
The basic mechanism is shown in Figures B.11 and B.12 for downlink and uplink, respectively.
www.4gamericas.org February 2011 Page 137