7 Uu Interface User Plane (UE-RAN)
26.4543GPPCodec for Enhanced Voice Services (EVS)Interface to Iu, Uu, Nb and MbRelease 17TS
7.1 General
The interface between the EVS Encoder and Decoder within the UE and the Radio Access Network within the UE is an UE-internal interface and is not detailed in the present document. The mapping of Speech, SID and EVS-CMR bits corresponds to the mapping described in clause 6 for the Iu interface.
The UE shall decode every Speech or SID frame received in downlink direction, regardless of the EVS mode of operation in uplink.
The UE shall always apply in uplink the "Source Controlled Operation" (SCR), also called "Discontinuous Transmission" (DTX).
If the UE uses the EVS AMR-WB IO mode of operation to encode speech for the uplink direction, then it shall restrict changes of the Codec Mode to every other frame (40 ms grid) and to neighbouring modes.
NOTE: In case of modification of the maximum rate (e.g. by the RNC), it takes a short time period until the speech frames sent by the UE to the Radio Access Network comply with the modified maximum rate. To facilitate optimization of performance (for instance avoiding problems like audio gaps due to discarded packets) in sending direction from the UE to the network when changing mode as requested, it is a good implementation practice to allow adaptation of the UE Encoder to take effect, and to maintain currently used rate for the next few frames after the information about the rate adaptation has been sent to the UE Encoder.
7.2 Determination of the local EVS encoding mode
Table 7.2-1: Local EVS encoding mode and Interworking
|
# |
Item |
Interworking |
Motivation/comment/discussion |
|---|---|---|---|
|
1 |
Set of available rates |
The UE shall confine the EVS encoder operation to modes, whose rates are within the Set of available rates. |
The UE may otherwise not be able to send. |
|
2 |
DL EVS-CMR |
The EVS-UE shall take into account a DL EVS-CMR as soon as possible, at maximum within 60 ms after reception on the radio interface. The UE shall follow the received DL EVS CMR, unless it is in conflict with the Set of available rates, in which case the CMR conflict resolution applies. |
Shortest possible reaction time is required for optimal adaptation speed. Typically, the UE uses the highest rate and bandwidth of the DL EVS-CMR for encoding in uplink. |
|
3 |
CMR conflict resolution |
If the received DL EVS-CMR is outside the Set of available rates, then the rate requested by DL EVS-CMR shall be changed to the closest rate supported by the Set of available rates that complies with the Major operation mode of the DL EVS-CMR. That is: if DL EVS-CMR corresponds to an EVS AMR-WB IO rate, then the requested rate is changed to an allowed EVS AMR-WB IO rate; if DL EVS-CMR corresponds to an EVS Primary mode, then the requested rate is changed to an allowed EVS Primary rate and bandwidth. |
A change of the Major operation mode (EVS primary <=> EVS AMR-WB IO) can only be initiated through DL EVS-CMR via the CN, but not through another change of the Set of available rates. Otherwise, such a change of the major operation mode from EVS AMR-WB IO to EVS primary would result in an interworking failure, if the remote UE is an AMR-WB-only phone. A change from EVS primary to EVS AMR-WB IO would result in non-optimal audio quality, if the remote UE is an EVS phone. The MSC guarantees through GBR that at least the lowest AMR-WB IO mode and at least the lowest EVS Primary mode of the EVS Configuration for UE remain selectable. The RNC behaviour and GBR setting is subject to operator policy. The UE reduces/increases the requested rate to the closest rate of the requested Major operation Mode, supported by the Set of available rates. If the requested Major operation mode equals the EVS primary mode, then the rate is modified within the set of rates of 24.4, 16.4, 13.2 (non-CA), 9.6, 8, 7.2, as far as these are within the EVS Configuration for UE. If requested Major operation mode equals the EVS AMR-WB IO mode, the rate is modified within the set of rates of 12.65, 8.85 and 6.60, as far as these are within the EVS Configuration for UE. |
|
4 |
UE autonomous rate |
The control logic for the UE autonomous rate shall guarantee that at least the lowest AMR-WB IO mode and at least the lowest EVS Primary mode of the EVS Configuration for UE are included in the Set of available rates. In addition, this control logic shall guarantee that the Set of available Rates at least includes the lowest EVS Primary rate supporting the audio bandwidth requested by the DL EVS-CMR (after mapping the DL EVS-CMR into the EVS Configuration for UE). |
Otherwise, the result might be interworking failure, if the remote UE is an AMR-WB only phone, or the result may be non-optimal audio quality. |
|
5 |
EVS Channel Aware mode (EVS-CA) |
The EVS-CA mode shall solely be selected through DL EVS-CMR. The UE shall not enter or leave the EVS-CA based on RNC command or its autonomous mode. |
The remote UE may request and the CN may select or deselect the EVS CA mode, e.g. based on the perceived transmission channel condition, see also clause 11. |
|
6 |
EVS VBR mode (EVS-VBR) |
The EVS-VBR mode shall solely be selected through DL EVS-CMR. The UE shall not enter or leave the EVS VBR based on RNC command or its autonomous mode. |
The use of the EVS-VBR mode requires that the upper limit of the Set of available rates is greater or equal to 8 kbps. A prerequisite for ensuring the EVS VBR mode is a GBR of at least 8 kbps. |
7.3 Generation of UL EVS-CMR for the remote media-sender
Table 7.3-1: UL EVS-CMR
|
# |
Topic |
Interworking |
Motivation/comment |
|
1 |
UL EVS-CMR |
The UE shall always send an UL EVS-CMR with the highest bit rate and audio bandwidth it wants to receive; it shall never send a NO_REQ. The UE shall send UL EVS-CMR for modes, which are inside the EVS configuration for UE. |
The present encoding mode has no influence on the UL EVS-CMR to be sent. The UE may generate UL EVS-CMR for a different bit rate, bandwidth or for requesting the switching between non-CA and EVS-CA mode. An UL EVS-CMR for a modified audio bandwidth may be generated, if the local audio playback capability changes. The EVS-CA mode may be optimized by sending UL EVS-CMR adjusting the redundancy level or the redundancy offset. |
7.4 Interworking in case of the EVS Channel Aware Mode
In CS networks, 3G-UEs do typically not deploy de-jitter buffers for the received speech frames, since the frames arrive essentially without any jitter every 20 ms (except in speech pauses). A MGW in the speech path in PS to CS calls shall remove any PS uplink jitter, see also clause 11, before the EVS frames are sent over the Iu interface.
There are nevertheless cases, where the use of the EVS CA mode may be advantageous within the 3G-UE. One such case is that the EVA CA mode is used for improving the robustness against PS-uplink and CS-downlink frame losses in PS to CS interworking. In that case, a part of the EVS CA mode processing is done in conjunction with Jitter Buffer Management (JBM) operation in a MGW between the PS-uplink and CS-downlink. Clause 11 describes the details.
The 3G-UE shall check every received EVS CA frame, if the use_partial_copy signalling bit (see clause 11) is set to 1 or 0. If the signalling bit is set to 0, then the EVS CA primary frame shall be decoded and the redundant part is ignored. If the signalling bit is set to 1, then the redundant part of this EVS CA frame shall be decoded, as if received in (the future) frame (N+K). The difference to a PS-UE is that the 3G-UE does not have to wait for frame (N+K), but can decode immediately without additional delay. The frame received with signalling bit 1 is saved in a buffer to improve the performance of CS-downlink loss, if frame (N+K) is lost on the CS downlink.