TTS purchase

ckatt

ckatt

状元
If i want to buy the TTS at this point what is the procedure? I found the link for the hui voice in another post, bu what about Lulu? Aslo if i only wan to purchase one, which is recommended at this point? I have been flowing a few threads with people reporting different successes with the two engines. Anyone out there with an HTC device that can make a recommendation? I am mostly using an HTC One V but also have a TF101 from Asus.
Thanks
 
Sorry, here's the link for Lulu - only recently started treating them as two separate purchases. (also added this link to the announcement post)
 
Any chance you could post examples of what the voices sound like?
and is there any major functionality difference between the two of them?
 
ckatt said:
Any chance you could post examples of what the voices sound like?
and is there any major functionality difference between the two of them?
Click to expand...

There'll be samples you can preview from the 'Add-ons' screen in the finished release - haven't got the links for them yet.

Functionality-wise, Hui seems to do a slightly better job with characters-with-multiple-readings at the moment, and while we haven't yet integrated this in Pleco (having only recently even decided to use both engines - this may actually make for some fun possibilities later for TTSing dialogues and such) it also has more fine-grained support for following along with where it is in the text as it reads.

But both engines are still being actively developed and either one could add a new feature in the future that the other one doesn't have, and we never know what sorts of roadblocks we might encounter as we try to integrate a particular feature in Pleco, so we can't promise that either of them will do any specific thing in the future that it doesn't do now, and right now the only difference we're aware of is that Hui does a slightly better job with ambiguous characters.
 
so i got the Hui today and so far i like it. dosen't seem to have any of the stutter problems i have been hearing about but it dose seem to speak faster in the flashcards than in the reader. also it seems to have some trouble with the neutral tone at the end of words. for instance if there is a 了 at the end it gives to much weight to it. or if there is an 儿 it separates the er. so in 味儿 instead of weir it reads it as wei er. the old voices often do this as well but i found them to be a little more intelligible when they made the mistake. anyway i'm guessing this is an engine thing so there is nothing you can do about it. but if you get the chance to pass the feedback back along to who ever it is that made the engine, it the only thing i see as a problem so far.
PS
it would be nice to have a way to tell the tts to just keep speaking in the reader rather than having to select all of the text first.
 
ckatt said:
anyway i'm guessing this is an engine thing so there is nothing you can do about it. but if you get the chance to pass the feedback back along to who ever it is that made the engine, it the only thing i see as a problem so far.
Click to expand...

Might not be... have you observed this when playing audio in the reader too (reading entire sentences, i.e.), or only when playing dictionary headwords / flashcards?

ckatt said:
it would be nice to have a way to tell the tts to just keep speaking in the reader rather than having to select all of the text first.
Click to expand...

That one's coming, but it's trickier to program (especially with the need to have the system follow along in the document as it reads), and given all of the stability problems with TTS in earlier betas we figured we'd better nail down the basic functionality first before we get into that.
 
You must log in or register to reply here.
Back
Top