🎁 新人 免费注册,送 10 次调用,最高 $1,免绑卡。

Google TTS Chirp 3 HD vs Google TTS Neural2

vs

何时使用哪一个 — 经过整理的结论,而非基准测试表格

google-tts-neural2 便宜 1.9 倍(每百万字符 $16 对 $30)并且直接支持 SSML,而 google-tts-chirp3-hd 把 SSML 列为预览——但 neural2 不支持流式,公布的语言区列表也短得多,只有 17 个带命名语音 id 的语言区。两者每次请求都接受 5000 字节,也都提供数值型语音参数。在已覆盖的语言区里做 SSML 驱动的批量合成选 google-tts-neural2,要流式和语言区广度选 google-tts-chirp3-hd。

定价

Google TTS Chirp 3 HD Google TTS Neural2 Δ
每 1M 字符 $30 $16 1.9×

费率取自构建时的实时目录;每个模型页面均附有当前的费率卡。

它们的位置 — 以该计费单位计费的所有 7 个 文本转语音 模型的 每 1M 字符的价格(对数刻度)

能力

Google TTS Chirp 3 HD Google TTS Neural2
流式传输
SSML preview supported
计费单位 character character

规格

Google TTS Chirp 3 HD Google TTS Neural2
输入模态 文本 文本
输出模态 音频 音频
请求限制 5000 bytes 5000 bytes
声音 Voices are named after stars and published with a gender: Achernar, Achird, Algenib, Algieba, Alnilam, Aoede, Autonoe, Callirrhoe, Charon, Despina, Enceladus, Erinome, Fenrir, Gacrux, Iapetus, Kore, Laomedeia, Leda, Orus, Pulcherrima, Puck, Rasalgethi, Sadachbia, Sadaltager, Schedar, Sulafat, Umbriel, Vindemiatrix, Zephyr and Zubenelgenubi. A locale prefix forms the id, e.g. en-US-Chirp3-HD-Charon. Google describes the set as 30 distinct styles across many languages. Voice ids follow the locale-plus-letter pattern (en-US-Neural2-F, ja-JP-Neural2-B). Google's comparison table lists Neural2 as general purpose, generally available, controllable via SSML and not streaming-capable, and the docs state the voices are based on the same technology used to create a Custom Voice, letting anyone use Custom Voice technology without training their own.
语言 ar-XA, bn-IN, bg-BG, yue-HK, hr-HR, cs-CZ, da-DK, nl-BE, nl-NL, en-AU, en-IN, en-GB, en-US, et-EE, fi-FI, fr-CA, fr-FR, de-DE, el-GR, gu-IN, he-IL, hi-IN, hu-HU, id-ID, it-IT, ja-JP, kn-IN, ko-KR, lv-LV, lt-LT, ml-IN, cmn-CN, mr-IN, nb-NO, pl-PL, pt-BR, pa-IN, ro-RO, ru-RU, sr-RS, sk-SK, sl-SI, es-ES, es-US, sw-KE, sv-SE, ta-IN, te-IN, th-TH, tr-TR, uk-UA, ur-IN and vi-VN. A much shorter locale list than Standard. Neural2 voice ids are published for da-DK, de-DE, en-AU, en-GB, en-IN, en-US, es-ES, es-US, fr-CA, fr-FR, hi-IN, it-IT, ja-JP, ko-KR, pt-BR, th-TH and vi-VN. Google notes Neural2 voices are available on global and single-region endpoints.
语音控制
  • Voice controls are Preview: pace via speaking_rate 0.25 to 2.0
  • pause tags [pause short], [pause long] and [pause] accepted only in the markup input field, never in text, and the model may disregard tags placed unnaturally
  • custom pronunciations in IPA or X-SAMPA
  • SSML support is Preview and synchronous-only, unsupported for streaming requests, with unlisted tags ignored and say-as interpret-as=expletive or bleep not supported
  • AudioConfig controls: speakingRate 0.25 to 2.0 (1.0 native)
  • pitch -20.0 to 20.0 semitones
  • volumeGainDb -96.0 to 16.0
  • effectsProfileId device profiles for wearable, handset, headphone, small and medium Bluetooth speaker, large home entertainment, large automotive and telephony playback
  • SSML tags include speak, break, say-as, sub, mark, prosody, emphasis, phoneme, voice and lang
限制

Content limit of 5,000 total bytes per synthesize request. Default response format LINEAR16

streaming supports ALAW, MULAW, OGG_OPUS and PCM, batch adds MP3, so MP3 is not available on the streaming path. The only voice type Google's comparison table marks as streaming-capable. Generally available in the global, us and eu endpoints plus asia-southeast1, europe-west2 and asia-northeast1, but out of scope for regionalization and data residency. Billed per character including spaces and newlines, with all SSML tags except <mark> counted.

Content limit of 5,000 total bytes per synthesize request. Output LINEAR16 (with a WAV header), MP3 at 32 kbps, OGG_OPUS, or G.711 MULAW and ALAW

optional sampleRateHertz resamples and fails the request if the rate is unsupported for the encoding. Not offered over streaming synthesis. Billed per character including spaces and newlines, with all SSML tags except <mark> counted

the multi-byte-counts-once note applies to Standard and WaveNet only.

规格转录自各供应商的文档;若供应商未发布某项数据,则直接省略该行,而非进行推断。 完整来源: Google TTS Chirp 3 HD · Google TTS Neural2

只需一行代码即可在它们之间切换

两个 ID 都包含在下方的每个选项卡中 — 高亮显示的两行是唯一的修改。相同的端点,相同的密钥,相同的请求结构。

from openai import OpenAI

client = OpenAI(
    base_url="https://synthorai.io/v1",
    api_key="sk-syn-...",
)

resp = client.audio.transcriptions.create(
    model="google-tts-chirp3-hd",
    # model="google-tts-neural2",  # 取消注释此行,注释上一行
    file=open("meeting.mp3", "rb"),
    language="en",
)
print(resp.text)

获取 API 密钥 →

常见问题

Google TTS Chirp 3 HD 和 Google TTS Neural2 哪个更便宜?

在 每 1m 字符 方面,Google TTS Neural2 更便宜($16 对比 $30,相差 1.9×)。其他行可能得出相反的结论——上方表格提供了完整信息,实际成本取决于你的组合使用情况。

我可以在不进行两次集成的情况下,对 Google TTS Chirp 3 HD 和 Google TTS Neural2 进行 A/B 测试吗?

可以。两者均通过同一个兼容 OpenAI 的端点提供服务,并使用同一把 API 密钥——切换只需更改一行模型字符串,因此你可以将一部分流量路由到各个模型并直接比较账单。

文本转语音如何计费?

按输入文本的字符数计费,每次请求的字符上限如规格表所示。无论使用哪种模型,长文本都必须在多个请求中分块。

相关对比