googleapis / googleapis/google-cloud-cpp

Adjust `end of speech` timeout on streaming speech to text

未关闭
#15,225 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看
type: feature request
主要语言
C++
星标
659
派生
462
平均合并
1 天 2 小时
30 天内合并 PR
89

描述

**What component of `google-cloud-cpp` is this feature request for?**

google/cloud/speech/speech_client

**Is your feature request related to a problem? Please describe.**

I would like to reduce timeout it takes on Google before end of speech is detected

**Describe the solution you'd like**

I would like to have parameter added to streaming recognition config (speech::v1::StreamingRecognizeRequest) that would setup this timeout, similarly to what EndSilenceTimeout is on MSFT (https://learn.microsoft.com/en-us/windows/apps/design/input/set-speech-recognition-timeouts)

> End Silence Timeout:
> This timeout is triggered after a phrase has been successfully recognized, and the service waits for further speech input before finalizing the recognition result.
> The EndSilenceTimeout property determines how long the service waits for additional speech after a recognized phrase before concluding that the speech input has ended.
> This timeout can be adjusted to accommodate various speaking styles, allowing for short pauses within a longer phrase without prematurely ending the recognition.

**Describe alternatives you've considered**

No alternatives are known to exist, but please let me know otherwise

**Additional context**

Request originates from work on AI voice assistants where I would like to experiment with slightly reduced end of speech timeouts

贡献指南

打开贡献指南

调研方向

从 google/cloud/speech/speech_client 和 speech::v1::StreamingRecognizeRequest API 定义开始,然后检查 Google Speech-to-Text 是否提供可配置的语音结束超时。完成标准是:可以为流式识别配置受支持的超时,并通过适当的测试验证其行为。

由索引模型根据 Issue 内容生成。

评估

技术栈
cpp, gcp
领域
api
Issue 类型
功能
难度
5/5
预计耗时
一周以上
活跃度
停滞
描述清晰度
基本清楚
新手友好度
35/100

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。