VOICEVOX_CORE 関連の短いプログラム若しくはファイル等

voicevox-core-cli.c

/*
CLI より VOICEVOX CORE の C API 経由でテキスト音声合成を行い、
音声ファイルを出力する。簡単なテスト目的のプログラムである。
Copyright (C) 2026 z80oolong (zool@zool.jpn.org)

Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:

The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <stdbool.h>

#include "voicevox_core.h"

#define HOMEBREW_PREFIX                  "/home/linuxbrew/.linuxbrew"
#ifndef VOICEVOX_CORE_PREFIX
#define VOICEVOX_CORE_PREFIX             HOMEBREW_PREFIX "/opt/voicevox-core"
#endif

#ifndef DEFAULT_ONNXRUNTIME_PATH
#define DEFAULT_ONNXRUNTIME_PATH         VOICEVOX_CORE_PREFIX "/lib/voicevox/libonnxruntime.so"
#endif
#ifndef DEFAULT_OPENJTALK_DIC_DIR
#define DEFAULT_OPENJTALK_DIC_DIR        VOICEVOX_CORE_PREFIX "/share/voicevox/dic"
#endif
#ifndef DEFAULT_VVM_MODEL_PATH
#define DEFAULT_VVM_MODEL_PATH           VOICEVOX_CORE_PREFIX "/share/voicevox/models/sample.vvm"
#endif
#define DEFAULT_SPEAKER_ID               0
#define DEFAULT_OUTPUT_FILE              "./audio.wav"

static void print_usage(const char *progname) {
    fprintf(stderr, "Usage: %s [-x libonnxruntime_so] [-d open_jtalk_dic_dir] [-v vvm_model] [-i speaker_id] [-o output_wav] TEXT\n", progname);
    fprintf(stderr, "  -x : ONNX Runtime .so ファイル (default: %s)\n", DEFAULT_ONNXRUNTIME_PATH);
    fprintf(stderr, "  -d : OpenJTalk辞書ディレクトリ (default: %s)\n", DEFAULT_OPENJTALK_DIC_DIR);
    fprintf(stderr, "  -v : VVMモデルファイル (.vvm) (default: %s)\n", DEFAULT_VVM_MODEL_PATH);
    fprintf(stderr, "  -i : 話者ID (default: %d)\n", DEFAULT_SPEAKER_ID);
    fprintf(stderr, "  -o : 出力WAVファイル (default: %s)\n", DEFAULT_OUTPUT_FILE);
    fprintf(stderr, "  TEXT: 合成するテキスト\n");
}

int main(int argc, char *argv[]) {
    const char *onnx_path      = DEFAULT_ONNXRUNTIME_PATH;
    const char *dic_dir        = DEFAULT_OPENJTALK_DIC_DIR;
    const char *vvm_path       = DEFAULT_VVM_MODEL_PATH;
    VoicevoxStyleId speaker_id = DEFAULT_SPEAKER_ID;
    const char *output_file    = DEFAULT_OUTPUT_FILE;
    char *text = NULL;

    int opt;
    while ((opt = getopt(argc, argv, "x:d:v:i:o:h")) != -1) {
        switch (opt) {
            case 'x':
                onnx_path = optarg;
                break;
            case 'd':
                dic_dir = optarg;
                break;
            case 'v':
                vvm_path = optarg;
                break;
            case 'i':
                speaker_id = (VoicevoxStyleId)atoi(optarg);
                break;
            case 'o':
                output_file = optarg;
                break;
            case 'h':
                print_usage(argv[0]);
                return 0;
            default:
                print_usage(argv[0]);
                return 1;
        }
    }

    if (optind >= argc) {
        fprintf(stderr, "Error: TEXT が指定されていません。\n");
        print_usage(argv[0]);
        return 1;
    }
    text = argv[optind];

    VoicevoxResultCode result = VOICEVOX_RESULT_OK;
    const VoicevoxOnnxruntime *onnxruntime = NULL;
    OpenJtalkRc *open_jtalk = NULL;
    VoicevoxSynthesizer *synthesizer = NULL;
    VoicevoxVoiceModelFile *model = NULL;
    uint8_t *wav = NULL;
    size_t wav_size = 0;

    /* ONNX Runtime ロード */
    VoicevoxLoadOnnxruntimeOptions ort_opts = voicevox_make_default_load_onnxruntime_options();
    ort_opts.filename = onnx_path;
    result = voicevox_onnxruntime_load_once(ort_opts, &onnxruntime);
    if (result != VOICEVOX_RESULT_OK) {
        fprintf(stderr, "ONNX Runtime ロード失敗: %s\n", voicevox_error_result_to_message(result));
        goto cleanup;
    }

    /* OpenJTalk */
    result = voicevox_open_jtalk_rc_new(dic_dir, &open_jtalk);
    if (result != VOICEVOX_RESULT_OK) {
        fprintf(stderr, "OpenJTalk 初期化失敗: %s\n", voicevox_error_result_to_message(result));
        goto cleanup;
    }

    /* Synthesizer */
    VoicevoxInitializeOptions init_opts = voicevox_make_default_initialize_options();
    result = voicevox_synthesizer_new(onnxruntime, open_jtalk, init_opts, &synthesizer);
    if (result != VOICEVOX_RESULT_OK) {
        fprintf(stderr, "Synthesizer 作成失敗: %s\n", voicevox_error_result_to_message(result));
        goto cleanup;
    }

    /* VVMモデルロード */
    result = voicevox_voice_model_file_open(vvm_path, &model);
    if (result != VOICEVOX_RESULT_OK) {
        fprintf(stderr, "VVMモデルロード失敗 (%s): %s\n", vvm_path, voicevox_error_result_to_message(result));
        goto cleanup;
    }

    /* VVMモデル適用 */
    result = voicevox_synthesizer_load_voice_model(synthesizer, model);
    if (result != VOICEVOX_RESULT_OK) {
        fprintf(stderr, "モデル適用失敗: %s\n", voicevox_error_result_to_message(result));
        goto cleanup;
    }

    /* 音声合成 */
    VoicevoxTtsOptions tts_opts = voicevox_make_default_tts_options();
    result = voicevox_synthesizer_tts(synthesizer, text, speaker_id, tts_opts, &wav_size, &wav);
    if (result != VOICEVOX_RESULT_OK) {
        fprintf(stderr, "音声合成失敗: %s\n", voicevox_error_result_to_message(result));
        goto cleanup;
    }

    /* WAVファイル出力 */
    FILE *fp = fopen(output_file, "wb");
    if (fp == NULL) {
        fprintf(stderr, "出力ファイルを開けません: %s\n", output_file);
        goto cleanup;
    }

    if (fwrite(wav, 1, wav_size, fp) != wav_size) {
        fprintf(stderr, "ファイル書き込み中にエラーが発生しました: %s\n", output_file);
        fclose(fp);
        goto cleanup;
    }

    fclose(fp);
    printf("音声合成完了: %s (%zu bytes)\n", output_file, wav_size);

cleanup:
    if (wav) voicevox_wav_free(wav);
    if (model) voicevox_voice_model_file_delete(model);
    if (synthesizer) voicevox_synthesizer_delete(synthesizer);
    if (open_jtalk) voicevox_open_jtalk_rc_delete(open_jtalk);

    return (result == VOICEVOX_RESULT_OK) ? 0 : 1;
}

voicevox-core-cli.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
CLI より VOICEVOX CORE の Python API 経由でテキスト音声合成を行い、
音声ファイルを出力する。簡単なテスト目的のプログラムである。
なお、本プログラムは C API 用のプログラムの移植版である。
Copyright (C) 2026 z80oolong (zool@zool.jpn.org)

Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:

The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
"""

import argparse
import sys
from pathlib import Path

from voicevox_core.blocking import (
    Onnxruntime,
    OpenJtalk,
    Synthesizer,
    VoiceModelFile,
)


def main():
    parser = argparse.ArgumentParser(
        description="VOICEVOX CORE Python API でテキストを音声合成してWAV出力"
    )
    parser.add_argument("text", nargs="?", help="合成するテキスト")
    parser.add_argument("-x", "--onnxruntime", default=None,
                        help="ONNX Runtime .so ファイルのパス")
    parser.add_argument("-d", "--dict-dir", default=None,
                        help="OpenJTalk辞書ディレクトリ")
    parser.add_argument("-v", "--vvm", default=None,
                        help="VVMモデルファイル (.vvm)")
    parser.add_argument("-i", "--speaker-id", type=int, default=0,
                        help="話者ID (style_id)")
    parser.add_argument("-o", "--output", default="audio.wav",
                        help="出力WAVファイルパス")

    args = parser.parse_args()

    if not args.text:
        print("Error: TEXT が指定されていません。", file=sys.stderr)
        parser.print_usage()
        sys.exit(1)

    # パス設定（Homebrew prefixに合わせる）
    HOMEBREW_PREFIX = Path("/home/linuxbrew/.linuxbrew")
    VOICEVOX_CORE_PREFIX = HOMEBREW_PREFIX / "opt/voicevox-core"

    onnx_path = args.onnxruntime or str(VOICEVOX_CORE_PREFIX / "lib/voicevox/libonnxruntime.so")
    dict_dir = args.dict_dir or str(VOICEVOX_CORE_PREFIX / "share/voicevox/dic")
    vvm_path = args.vvm or str(VOICEVOX_CORE_PREFIX / "share/voicevox/models/sample.vvm")
    output_path = Path(args.output)

    print(f"ONNX Runtime: {onnx_path}")
    print(f"OpenJTalk辞書: {dict_dir}")
    print(f"VVMモデル: {vvm_path}")
    print(f"話者ID: {args.speaker_id}")
    print(f"出力ファイル: {output_path}")
    print(f"テキスト: {args.text}")

    try:
        # ONNX Runtime ロード
        ort = Onnxruntime.load_once(filename=onnx_path)

        # OpenJTalk
        open_jtalk = OpenJtalk(dict_dir)

        # Synthesizer作成
        synthesizer = Synthesizer(ort, open_jtalk)

        # VVMモデルロード & 適用
        with VoiceModelFile.open(vvm_path) as model:
            synthesizer.load_voice_model(model)

        # 音声合成 (TTS)
        wav_bytes = synthesizer.tts(
            text=args.text,
            style_id=args.speaker_id,
            # tts_options=... で各種オプションも指定可能
        )

        # WAV保存
        output_path.write_bytes(wav_bytes)
        print(f"[OK] 音声合成完了: {output_path} ({len(wav_bytes):,} bytes)")

    except Exception as e:
        print(f"[ERR] エラー: {e}", file=sys.stderr)
        sys.exit(1)


if __name__ == "__main__":
    main()

voicevox_core.pc.in

prefix=@prefix@
exec_prefix=${prefix}
libdir=${prefix}/lib
includedir=${prefix}/include

Name: voicevox_core
Description: VOICEVOX CORE Library
Version: 0.16.4
URL: https://voicevox.hiroshiba.jp/

Libs: -L${libdir} -lvoicevox_core -ldl
Cflags: -I${includedir}