西班牙语电话谈判语音智能体实测:Pipecat、LiveKit 与 GPT-Live 对比
Latency, Speech Recognition and Turn-Taking in Voice Agents for Spanish Phone Negotiation: An Externally Measured Comparison of Pipecat, LiveKit and GPT-Live
做电话语音智能体的开发者应该看看这篇:三种主流架构实打实跑了 77 通西语谈判电话。级联方案在金额识别上错 13%,对方一句"ok"抢话就可能被当成成交,GPT-Live 两种问题都没有,但音频往返 OpenAI 多花约 1.4 秒。
一项研究对比了三种语音智能体架构在西班牙语货运费率电话谈判中的表现:Pipecat 与 LiveKit Agents 编排的级联方案(语音识别+语言模型+语音合成),以及 OpenAI 的 GPT-Live 端到端语音模型。
测试用同一说话人的 47 段录音模拟承运商,每种架构跑 77 通电话、覆盖 5 个谈判场景,在听者一侧测量延迟。内容响应中位延迟为 2,616.
0 ms(Pipecat)、3,090.0 ms(LiveKit)、2,444.
8 ms(GPT-Live),GPT-Live 相对 Pipecat 的 171.2 ms 优势置信区间包含零,统计上不显著。
机制拆解显示,GPT-Live 按自身时间戳在线路结束后 800 ms 开始应答,接近 OpenAI 公布的 0.8 s;
剩余等待中约 1,367 ms 近乎恒定,是音频在 OpenAI 服务中的往返开销。工具调用在所有架构中都增加 0.
9-1.1 s,超过架构间任何差异;
框架自报延迟比听者实际等待低 9%(Pipecat)和 22%(LiveKit)。识别词错率很低(6.
03%、4.93%、1.
47%),但级联方案在金额上的错误率达 13.01% 和 12.
61%,GPT-Live 仅 2.86%。
默认设置下 Pipecat 和 LiveKit 几乎在每次附和时都停顿并常误当作回答:对方抢话说的"ok"在 15 通 Pipecat 电话中有 3 通被登记为达成的协议,LiveKit 有 2 通,GPT-Live 为 0。
研究结论是延迟应在听者一侧测量,级联方案需要为口头金额和附和处理增加保护机制。
Latency, Speech Recognition and Turn-Taking in Voice Agents for Spanish Phone Negotiation: An Externally Measured Comparison of Pipecat, LiveKit and GPT-Live
A voice agent that negotiates by phone must answer quickly, understand spoken amounts and cope with a caller who talks over it. We compared three architectures for such an agent, negotiating freight rates in Spanish: two cascades (speech recognition, language model and speech synthesis) orchestrated by Pipecat and by LiveKit Agents, and GPT-Live, an OpenAI speech model that listens and speaks without separate stages. A program played the carrier with the same 47 recordings of one speaker for all agents, without the telephone network, in 77 calls per architecture over five negotiation scenarios, and measured latency on the listener's side with a validated instrument. Median latency to the content response was 2,616.0 ms (Pipecat), 3,090.0 ms (LiveKit) and 2,444.8 ms (GPT-Live); the confidence interval of GPT-Live's 171.2 ms advantage over Pipecat includes zero. According to its own timestamps, GPT-Live began to answer 800 ms after the end of the line, close to the 0.8 s published by OpenAI; most of the remaining wait, about 1,367 ms and nearly constant, is the audio's round trip through OpenAI's service. A tool call added 0.9-1.1 s in every architecture, more than any difference between them, and the frameworks' self-reported latency fell short of the listener's wait by 9% (Pipecat) and 22% (LiveKit). Word error rates were low (6.03, 4.93 and 1.47%), but the error on amounts was 13.01% and 12.61% in the cascades, against 2.86% in GPT-Live. With default settings, Pipecat and LiveKit stopped at almost every backchannel and often took it as the carrier's answer: an "ok" spoken over the agent ended in a registered agreement in 3 of 15 Pipecat calls, 2 of 15 LiveKit calls and no GPT-Live call (descriptive counts, not tested). Latency should therefore be measured on the listener's side, and a cascade needs safeguards for spoken amounts and backchannels.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:Pipecat、LiveKit、GPT-Live、语音智能体,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「arXiv cs.AI」(@Jovan Popoca Moran, Eduardo de Avila-Armenta, Raúl Alejandro Aquino-Ibarra, Jorge Octavio Nicolás-Díaz, Umberto Leon-Dominguez),经 AITOP 多维评分管线结构化提炼。