Single Radio - Voice Call Continuity (SRVCC) is an existing standard ([3], [4]), specifying the handover of a Voice or Video call from LTE access to CS radio access, either to GERAN (2G) or to UTRAN (3G) or other CS networks. The present document considers only eSRVCC for voice calls between 3GPP accesses. In the IMS Core Network, the voice call is typically anchored in the ATCF/ATGW (Access Transfer Control Function / Access Transfer Gate Way). The eSRVCC procedure, as specified, may cause additional transcoding between the target radio leg and the ATGW, even though in theory it would be possible to avoid it. As a result, the eSRVCC procedures may add one or more unnecessary transcoding point(s) for the call and thereby degrade the quality of the ongoing call unnecessarily. Transcoding-Less Codec Interworking (TLCI) is always desirable to achieve good voice quality. Furthermore TLCI preserves network resources, i.e. by avoiding transcoding. TLCI is especially important for HD Voice. The Mobility Management Entity (MME) of the LTE-RAN, which sends the PS-to-CS Handover Request to the Target Network, does not know the IMS Selected Codec, which is in use before the eSRVCC in the ongoing call towards the remote end. Thus the MME cannot support the Target Network for selecting the optimal Target RAN Codec. The Target Network thus selects this Target RAN Codec on own criteria; often the Target RAN Codec is then not compatible to the IMS Selected Codec. Transcoding is then the immediate reaction. While it is possible for the ATCF, based on the current procedure, to renegotiate the IMS Selected Codec with the remote end to fit any selected Target RAN Codec at call transfer, this may extend the perceived time it will take to conclude the call transfer and this might extend the speech interruption time that might result due to the time the additional negotiation with the remote end will take. The ATCF was introduced for exactly that reason: avoid renegotiation with the remote end - accelerate eSRVCC. But even worse: in a substantial number of call scenarios the remote end may not be able to support the arbitrarily chosen Target RAN Codec and the transcoding cannot even be avoided by that renegotiation. The first attempt will optimize the Target RAN Codec to fit the IMS Selected Codec. If that is impossible or not optimal, then the renegotiation with the remote end might be attempted. The last resort has to be transcoding; sometimes it is unavoidable.

Enhanced Single Radio - Voice Call Continuity (eSRVCC) is an existing standard ([3], [4]) specifying the handover of a Voice or Video call from LTE access to CS-radio access, either to GERAN (2G) or to UTRAN (3G) or other CS networks. The present document considers only enhanced SRVCC for voice calls between 3GPP accesses. This study assumes that the Codecs defined in TS 26.114 are used on the LTE access and the Codecs defined in TS 26.103 on the CS accesses. Since Rel-13, the specifications for CS networks include the Codec Type UMTS_EVS with several Configurations, called UMTS_EVS (Set 0) to UMTS_EVS (Set 3). In the IMS Core Network, the voice call is typically anchored in the ATCF/ATGW (Access Transfer Control Function/ Access Transfer Gate Way). The eSRVCC procedure, as specified, may cause additional transcoding between the target radio leg and the ATGW, even though in theory it would be possible to avoid it. As a result, the eSRVCC procedures may add one or more unnecessary transcoding point(s) for the call and thereby degrade the quality of the ongoing call unnecessarily. The main objectives of this study are to analyse example call scenarios and find potential solutions to minimize the number of transcoding cases. Another objective is to optimize the interworking and the transition between EVS and AMR-WB during eSRVCC. The study should also show the reasons and potential solutions for too long speech path interruptions during eSRVCC. The present Technical Report has the following detailed objectives:

• Identify relevant eSRVCC scenarios, especially with Codec Mode Control;
  from AMR-WB and/or EVS in VoLTE to AMR-WB and/or EVS in CS;
  but include also other important Codecs, such as AMR and G.722.
• Analyse Speech Quality Aspects and Media Handling Aspects, based on these scenarios.
• Analyse Codec Mode Control before, during and after eSRVCC;
  recently SA4 has clarified some essential details on Rate Control for AMR and AMR-WB;
  Rate Control and Audio Bandwidth Control for EVS are still under discussion to some extent.
• Analyse the existing SDP Offer - Answer protocol between Target MSC and Anchor-ATCF during eSRVCC,
  as specified in TS 23.216, Stage 2;
  This analysis will include the whole eSRVCC procedure for at least one essential scenario
  (e.g. eSRVCC to GERAN) and will identify the potential reasons for transcoding and too long speech path interruptions.
• Clarify the existing Codec Compatibility aspects for eSRVCC;
  especially the interworking between CS and IMS for AMR, AMR-WB and EVS needs to be documented.
• Propose enhancements for media and quality aspects of eSRVCC with the aims:
  1. to avoid transcoding cases as much as possible;
  2. to minimize the speech path interruption time during eSRVCC;
• Support the SA2 SETA work by SA4 expertise in speech quality and media handling.