[Unofficial] Feature Request / Suggestion List

Shun

状元
Hello Mike,

can I ask, do Apple let you access the OS's Chinese Siri voices, as well? Those would be quite excellent, very intelligible.

Thanks, Shun
 

Shun

状元
can I ask, do Apple let you access the OS's Chinese Siri voices, as well? Those would be quite excellent, very intelligible.

I think I answered it myself—Siri voice APIs are private:

 

hunxiao

Member
The problem with those other offerings is that they’re mostly online, and rarely come with the ability to follow along word by word as they read. Also, frankly, most users seem to be happy with the system TTS, and paid TTS is not a particularly big / profitable add-on for us.

So our current plan is mostly to improve system TTS support - let you access all of the Chinese voices on iOS and juggle multiple engines / voices simultaneously on Android - and wait for somebody to release an offline Chinese TTS with the features we need.
Hi, Mike. As I understand it, Azure neural TTS is recently available on-device/disconnected and also includes word-by-word tracking. However, if most users are satisfied with legacy technology or it's not profitable to develop further in any case, I get it.
 

mikelove

皇帝
Staff member
can I ask, do Apple let you access the OS's Chinese Siri voices, as well? Those would be quite excellent, very intelligible.
The Siri ones don't seem to be accessible, no; the APIs give us access to I think 7 in total but seemingly not those.

Azure neural TTS is recently available on-device/disconnected and also includes word-by-word tracking
It's not available on iOS yet, unfortunately:


It also seems to mostly be intended as an enterprise thing, they aren't really set up to license it for use on consumer devices; the main use case they're imagining seems to be industrial systems without reliable network access. (which probably explains the lack of iOS support)

One thing we have considered is using a cloud TTS engine to generate time-keyed audio recordings for example sentences and readers, which you would then download and play locally; you'd end up with effectively the same behavior as an offline TTS but without it having to actually run in Pleco. (main downside is that it'd take up a lot more storage space, but probably not enough to be a problem on modern devices)
 

mikelove

皇帝
Staff member
Yeah, I don't know whether they actually don't want us using them or whether they simply work in some slightly different way from the other voices that means they're not compatible with the AVSpeechSynthesizer APIs, but either way, not available to us.
 
Top