iamf/cli/proto/audio_element.proto

// Copyright (c) 2023, Alliance for Open Media. All rights reserved
//
// This source code is subject to the terms of the BSD 3-Clause Clear License
// and the Alliance for Open Media Patent License 1.0. If the BSD 3-Clause Clear
// License was not distributed with this source code in the LICENSE file, you
// can obtain it at www.aomedia.org/license/software-license/bsd-3-c-c. If the
// Alliance for Open Media Patent License 1.0 was not distributed with this
// source code in the PATENTS file, you can obtain it at
// www.aomedia.org/license/patent.

syntax = "proto2";

package iamf_tools_cli_proto;

import "iamf/cli/proto/obu_header.proto";
import "iamf/cli/proto/param_definitions.proto";

enum AudioElementType {
  AUDIO_ELEMENT_INVALID = 0;
  AUDIO_ELEMENT_CHANNEL_BASED = 1;
  AUDIO_ELEMENT_SCENE_BASED = 2;
}

message ParamDefinitionExtension {
  optional uint32 param_definition_size = 1;
  optional bytes param_definition_bytes = 2;
}

message AudioElementParam {
  optional uint32 deprecated_param_definition_type = 1 [deprecated = true];
  optional ParamDefinitionType param_definition_type = 5;
  oneof param_definition {
    DemixingParamDefinition demixing_param = 2;
    ReconGainParamDefinition recon_gain_param = 3;
    ParamDefinitionExtension param_definition_extension = 4;
  }

  // Next ID: 6
}

// Valid proto enums start at index 1, which are different from the
// corresponding enums in C++, e.g. kLayoutMono = 0.
enum LoudspeakerLayout {
  LOUDSPEAKER_LAYOUT_INVALID = 0;
  LOUDSPEAKER_LAYOUT_MONO = 1;
  LOUDSPEAKER_LAYOUT_STEREO = 2;
  LOUDSPEAKER_LAYOUT_5_1_CH = 3;
  LOUDSPEAKER_LAYOUT_5_1_2_CH = 4;
  LOUDSPEAKER_LAYOUT_5_1_4_CH = 5;
  LOUDSPEAKER_LAYOUT_7_1_CH = 6;
  LOUDSPEAKER_LAYOUT_7_1_2_CH = 7;
  LOUDSPEAKER_LAYOUT_7_1_4_CH = 8;
  LOUDSPEAKER_LAYOUT_3_1_2_CH = 9;
  LOUDSPEAKER_LAYOUT_BINAURAL = 10;
  LOUDSPEAKER_LAYOUT_RESERVED_10 = 11;
  LOUDSPEAKER_LAYOUT_RESERVED_14 = 15;
  LOUDSPEAKER_LAYOUT_EXPANDED = 16;
}

enum ExpandedLoudspeakerLayout {
  EXPANDED_LOUDSPEAKER_LAYOUT_INVALID = 0;
  EXPANDED_LOUDSPEAKER_LAYOUT_LFE = 1;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_S = 2;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_SS = 3;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_RS = 4;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_TF = 5;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_TB = 6;
  EXPANDED_LOUDSPEAKER_LAYOUT_TOP_4_CH = 7;
  EXPANDED_LOUDSPEAKER_LAYOUT_3_0_CH = 8;
  EXPANDED_LOUDSPEAKER_LAYOUT_9_1_6_CH = 9;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_F = 10;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_SI = 11;
  EXPANDED_LOUDSPEAKER_LAYOUT_STEREO_TP_SI = 12;
  EXPANDED_LOUDSPEAKER_LAYOUT_TOP_6_CH = 13;
  EXPANDED_LOUDSPEAKER_LAYOUT_RESERVED_13 = 14;
  EXPANDED_LOUDSPEAKER_LAYOUT_RESERVED_255 = 256;
}

message ChannelAudioLayerConfig {
  optional uint32 deprecated_loudspeaker_layout = 1 [deprecated = true];
  optional LoudspeakerLayout loudspeaker_layout = 10;
  optional uint32 output_gain_is_present_flag = 2;
  optional uint32 recon_gain_is_present_flag = 3;
  optional uint32 reserved_a = 4;
  optional uint32 substream_count = 5;
  optional uint32 coupled_substream_count = 6;
  optional uint32 output_gain_flag = 7;
  optional uint32 reserved_b = 8;
  optional int32 output_gain = 9;
  optional ExpandedLoudspeakerLayout expanded_loudspeaker_layout = 11;
}

message ScalableChannelLayoutConfig {
  optional uint32 num_layers = 1;
  optional uint32 reserved = 2;
  repeated ChannelAudioLayerConfig channel_audio_layer_configs = 3;
}

enum AmbisonicsMode {
  AMBISONICS_MODE_INVALID = 0;
  AMBISONICS_MODE_MONO = 1;
  AMBISONICS_MODE_PROJECTION = 2;
}

message AmbisonicsMonoConfig {
  optional uint32 output_channel_count = 1;
  optional uint32 substream_count = 2;
  repeated uint32 channel_mapping = 3 [packed = true];
}

message AmbisonicsProjectionConfig {
  optional uint32 output_channel_count = 1;
  optional uint32 substream_count = 2;
  optional uint32 coupled_substream_count = 3;
  repeated int32 demixing_matrix = 4 [packed = true];
}

message AmbisonicsConfig {
  optional AmbisonicsMode ambisonics_mode = 1;
  oneof ambisonics_config {
    AmbisonicsMonoConfig ambisonics_mono_config = 2;
    AmbisonicsProjectionConfig ambisonics_projection_config = 3;
  }
}

message AudioElementConfigExtension {
  optional uint32 audio_element_config_size = 1;
  optional bytes audio_element_config_bytes = 2;
}

message AudioElementObuMetadata {
  optional uint32 audio_element_id = 1;
  optional AudioElementType audio_element_type = 2;
  optional uint32 reserved = 3;
  optional uint32 codec_config_id = 4;
  optional uint32 num_substreams = 5;
  repeated uint32 audio_substream_ids = 6 [packed = true];
  optional uint32 num_parameters = 7;
  repeated AudioElementParam audio_element_params = 8;
  oneof config {
    ScalableChannelLayoutConfig scalable_channel_layout_config = 9;
    AmbisonicsConfig ambisonics_config = 10;
    AudioElementConfigExtension audio_element_config_extension = 11;
  }
  optional ObuHeaderMetadata obu_header = 12;
}