Recommendation for good TTS for Asian language

Does anyone please have an up-to-date recommendation for a TTS that is maybe optimised for Asian languages, specifically Northern Vietnamese.

I don’t mind paying for good quality (ie good audio, able to inject some intelligence/emotion into the audio, eg articulating words in a phrase as a natural group), but would prefer one-off payment rather than monthly sub.

(I’m finding that I need to remake the Vietnamese mini-story lessons for several reasons eg sentence mode is just not well enough synchronised.)

There are a few options that come to mind.

TTSMaker

https://ttsmaker.com/

Supports 5 Vietnamese Voices - Is unlimited free use on their website (with a captcha) or you can buy a subscription for easier access. With the paid version you adjust a whole range of options for emotion / custom SSML.

image
image
image
image

Microsoft Edge Voices

If you use Microsoft Edge you can activate the read-along mode which has access to 2 Vietnamese voices for free.


You can also do this (without edge browser and ability to save the audio / speed of audio) with this python module

edge-tts · PyPI

Meta MMS

Demo: - MMS - a Hugging Face Space by mms-meta
Info: facebook/mms-tts · Hugging Face

  • Support for 1100+ languages (including Vietnamese)
  • Voices List

I am still at the early-beginner level for Vietnamese. I too am interested in hearing Northern speech.

I see that @roosterburton selected the ttsmaker voice “Yen 459” when preparing his image of the ttsmaker web page.

image

That voice, “Yen 459”, sounds pretty Northern to me. I compared it to the following three other voices:
ttsmaker Lan 450
Edge HoaiMy
Edge NamMinh

So for me, “Yen 459” is the one I would pick when doing text-to-speech. I wonder if any native speakers of Vietnamese can confirm that “Yen 459” has a standard Northern accent.

Thanks @roosterburton. The problem is that if you do a search for Vietnamese TTS then you get many hits from which Narakeet often seems to come near the top. I don’t really believe that there are this many different services and probably most of them are just wrappers accessing the same underlying TTS primary service. So I was really just wondering if anyone might have actually used and maybe compared any of these services and, if so, which they felt did best. I know it’s a long shot, but if you don’t ask…

I don’t need the complication of an API, this is just asking for a service that has a textbox input and allows download of the resulting audio file.

And yes, the naming of Vietnamese voices is a definite nuisance, ie the failure to distinguish between Northern/Southern/Central dialects. It’s like a non-English speaker looking for a standard UK English voice and being presented with an unlabelled choice that includes a strong Geordie or Glaswegian accent. Of course, ultimately it’s good to be able to understand all accents, but when learning as a beginner it’s difficult enough with a language like Vietnamese to achieve some useful listening comprehension even with a single consistent accent. But listening to a Southern accent when you’re used to hearing Northern accent is just adding another layer of complication that is often puzzling and unhelpful.