84 lines
1.9 KiB
JSON
84 lines
1.9 KiB
JSON
{
|
|
"codec_start": {
|
|
"type": "codec.start",
|
|
"version": 1,
|
|
"utterance_index": 0,
|
|
"sentence_index": 0,
|
|
"sentence_text": "I can't believe how beautiful the sunset looks from up here on the mountain.",
|
|
"model": "Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice",
|
|
"codebooks": 16,
|
|
"codebook_size": 2048,
|
|
"frame_rate_hz": 12.5,
|
|
"sample_rate": 24000,
|
|
"decoder_id": "qwen3_tts_tokenizer_v2_12hz",
|
|
"header_format": "<IHB"
|
|
},
|
|
"schedule": [
|
|
2,
|
|
4,
|
|
8,
|
|
16,
|
|
25
|
|
],
|
|
"total_frames": 67,
|
|
"ttfa_ms": 151.47794736549258,
|
|
"playback_start_ms": 151.47845214232802,
|
|
"stall_at_ttfa_ms": 0,
|
|
"audio_s": 5.36,
|
|
"total_s": 2.755943182390183,
|
|
"rtf": 0.514168504177273,
|
|
"decode_ms_total": 696.1517208255827,
|
|
"chunks": [
|
|
{
|
|
"idx": 0,
|
|
"t_ms": 151.47845214232802,
|
|
"arrival_ms": 106.2490432523191,
|
|
"frames": 2,
|
|
"samples": 3840,
|
|
"decode_ms": 45.2048322185874
|
|
},
|
|
{
|
|
"idx": 1,
|
|
"t_ms": 308.50786017253995,
|
|
"arrival_ms": 249.9086633324623,
|
|
"frames": 4,
|
|
"samples": 7680,
|
|
"decode_ms": 58.58577275648713
|
|
},
|
|
{
|
|
"idx": 2,
|
|
"t_ms": 618.9267742447555,
|
|
"arrival_ms": 536.8196251802146,
|
|
"frames": 8,
|
|
"samples": 15360,
|
|
"decode_ms": 82.09230238571763
|
|
},
|
|
{
|
|
"idx": 3,
|
|
"t_ms": 1249.835402239114,
|
|
"arrival_ms": 1119.5730213075876,
|
|
"frames": 16,
|
|
"samples": 30720,
|
|
"decode_ms": 130.24557195603848
|
|
},
|
|
{
|
|
"idx": 4,
|
|
"t_ms": 2243.4333111159503,
|
|
"arrival_ms": 2046.2959483265877,
|
|
"frames": 25,
|
|
"samples": 48000,
|
|
"decode_ms": 197.09910824894905
|
|
},
|
|
{
|
|
"idx": 5,
|
|
"t_ms": 2754.7398009337485,
|
|
"arrival_ms": 2571.790866088122,
|
|
"frames": 12,
|
|
"samples": 23040,
|
|
"decode_ms": 182.92413325980306
|
|
}
|
|
],
|
|
"device": "cpu",
|
|
"wav": "/data/zhoutaichang/embedding_infer/vllm-omni/benchmarks/tts/edge_results/codec_stream/20260909-164759/client_3.wav",
|
|
"codes": "/data/zhoutaichang/embedding_infer/vllm-omni/benchmarks/tts/edge_results/codec_stream/20260909-164759/client_3.codes.npy"
|
|
} |