Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Data privacy...? #138

Open
hamuchiwa opened this issue Feb 20, 2019 · 9 comments
Open

Data privacy...? #138

hamuchiwa opened this issue Feb 20, 2019 · 9 comments

Comments

@hamuchiwa
Copy link

Hello, when sending requests to Google Web Speech API, does Google collect the data? If so, are we able to opt-out of Google data collection?

@BingLingGroup
Copy link

BingLingGroup commented Feb 21, 2019

@hamuchiwa
About speech-to-text api used in autosub (not the same one in cloud speech api) #1
And to the translation api, you can find some kind of tos on the offcial website I guess. (google translation api website)

@BingLingGroup
Copy link

BingLingGroup commented Feb 21, 2019

@hamuchiwa 感觉你没看到这层回复,圈你一下。
才发现好像是国人啊,我就直说了,autosub里面用的语音转文字api是从chromium源码里逆向出来的(出处见楼上1号问题),可以在源码的constant.py(手机回复,我就不加链接了)里看到请求url,api key是以明文的形式写在url里的,大概是谷歌官方的某个公共api(譬如你搜索框旁边的小话筒一类),所以谈什么隐私好像意义不大……除此之外,如果想使用最新版的cloud speech api(还有最新版的翻译api,以及什么机器学习模式的翻译等等),可能需要自己改代码了,autosub没有加这方面的功能(感觉是作者太懒)。

@hamuchiwa
Copy link
Author

@BingLingGroup 谢谢回复,这么说数据保护什么的想也不要想了。。。还有一事不要明白,谷歌的这个语音转文本不是收费的嘛,作者找的这个api key还能无限量的用?

@BingLingGroup
Copy link

BingLingGroup commented Feb 22, 2019

@hamuchiwa 我不是说了这个api的来源了吗……一号issue里面有讲啊,chromium逆向出来的,看楼上。
这个api不是cloud speech-to-text那个api,不是用python cloud speech-to-text client进行请求的,而是用url请求的,估计就是网页服务一类的。
这个google web speech-to-text api用法参考 #50#135

@hamuchiwa
Copy link
Author

@BingLingGroup 请问有没有其他的语音转中文字幕的工具推荐,我用这个autosub转了一个视频,有些字幕不是很准确。

@BingLingGroup
Copy link

BingLingGroup commented Feb 23, 2019

@hamuchiwa 这个应该是免费里面最准确的,而且我觉得可能也是付费里面最准确的……你指的不准是说普通话识别成粤语吗?还是时间轴的不准?还是单纯只是文字内容不准?
autosub在speech-to-text上面有两点需要改进,一个是时间轴切割,目前时间轴切割只是用音频功率判断的,没有更精确的方法,因为api只返回识别文本,不返回时间轴。一个是需要加入对付费的google cloud speech api的支持,目前的api是逆向出来的免费版,虽然我个人认为不会有显著差异。
文字内容不准的问题一般是背景音比较嘈杂或者说话人声音不够清晰准确导致的。
如果是普通话识别成粤语的问题,可以参考 #136 的修改版代码。
时间轴如果不准的话,可以考虑曲线救国,把视频传到youtube上让youtube识别(如果可以用youtube的话)

@hamuchiwa
Copy link
Author

@BingLingGroup 文字内容大体正确,个别词语识别有误。我大概也上网搜了搜, 免费的除了这个也没找着别的,其他的都是收费的。有一个叫“人人译世界”的应用,可以免费试用生成一分钟的字幕,用的同样一个中文视频,100%识别正确,就是速度很慢。Youtube好像还不能识别中文吧。
我觉得准确率很重要,数据的保护也很重要啊,要是有能线下转字幕的就好了,曲线救国,多绕几个圈我也愿意~~

@BingLingGroup
Copy link

BingLingGroup commented Feb 24, 2019

@hamuchiwa youtube可以识别中文……youtube能识别的语言和cloud speech-to-text api基本一致,而且还能检测时间轴,只是需要你上传视频时设置好视频的语言
我个人认为能有一定准确率就可以了,至少能达到我再校对字幕的时候没有从零开始做那么累就行了,难道你还不会做字幕?Aegisub了解一下。
离线转换是不存在的……这玩意要是做出离线软件那这些公司还怎么卖服务啊。除非你去买实体产品什么的,譬如翻译机(

@roipoussiere
Copy link

Related to #83: using Mozilla's DeepSpeech project and it's commonVoice dataset, users could install the speech recognition software locally and avoid sending private data to big techs.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants