Skip to content

Add IVAS RTP implementation

Lauros Pajunenrequested to merge
1154-add-rtpdump-support-no-pi-appliance into main
  • Related issues:
  • Requested reviewers:

This is the "clean" version of the MR, see also !1809 (merged)

Usage

Encoder:

  • Use -rtpdump for the encoder to output an rtpdump stream, which is written to the output bitstream file. The rtpdump stream contains single IVAS frames packetized into IVAS RTP payload format packets. In EVS mono operating mode, the frames are packetized into IVAS payload format packets, which are interoperable with the EVS Header-Full RTP payload format.
  • -scene_orientation <file> can be used to input a scene orientation file to the encoder. The scene orientation values will be written to the PI data section of the IVAS payload. The file reading assumes one quaternion orientation value per line (same as for head orientation file reading).
  • -device_orientation <file>, same as above but for device orientations (uncompensated).

Decoder:

  • Use -VOIP_hf_only=1 to read the input rtpdump stream. This option should be able to read both IVAS payload format and EVS Header Full payload format.

Example:

IVAS_cod -rtpdump -SCENE_ORIENTATION ./scripts/testv/headrot_case00_3000_q.csv -masa 2 ./scripts/testv/stv1MASA2TC48c.met 128000 48 ./scripts/testv/stv1MASA2TC48c.wav ivas_rtpdump_test.bts
IVAS_dec -VOIP_hf_only=1 BINAURAL 48 ivas_rtpdump_test.bts ivas_rtpdump_out.wav

Related to #1154

Merge request reports