I recently came across some anki plugins for TTS that use google's TTS engine. I imagine Japanese lends itself fairly well to TTS, at least way better than English (vocaloids kinda proved that), but do you really want to memorize computer-generated audio? Can a computer really get the inflection and accent right? I feel like that's kind of a dangerous thing to use, especially for something like an anki vocab deck where you're being exposed to new words for the first time and you don't want to learn them wrong. I feel like it'd be better to just read for anki (or download audio of a Japanese person saying the text), and then get listening practice with anime, audiobooks, podcasts, or whatever else you use.

The one thing I feel like TTS would be really good for is reading novels out loud, since it's impossible to find actual audiobooks for most things in Japanese. That way I could read a new chapter on my own, and then have the TTS read it to me while I multi-task later as passive learning. The pronunciation wouldn't have to be perfect because I'm not memorizing new words so much as reviewing sentence patterns.

Also, I'm not super tech-savvy (anki constantly pushes my limits), and I know nothing about TTS. I was trying to google around to see if I could use google's TTS online, without having to start downloading 5 different plugins, and test it out to judge for myself. But... I couldn't find anything. Is it the same engine that will read aloud the translation on google translate? I also found some chrome plugins but I don't think they're by google, and some Japanese voices but they just read English text with an accent... I don't want to get into the technical side in this topic too much, but I guess, is there a place you can preview the engine that the anki plugins use?

Thanks for your input everyone!

I don't have much specific information, but I would suspect you would end up speaking a little strangely. Even though TTS is getting a lot better, it still sounds strange and robotic. You would probably internalize some of that strangeness if you used TTS extensively as a learning tool.

One of the major problems with TTS is that that aren't very good at replicating tonality and volume modulation which adds extra information that is not printed in the text. For instance, when we ask a question, our voice usually goes up at the end. Better TTS voices deal with this, but I know that there are other problems that haven't been addressed. I wasn't familiar with vocaloids, but a singing TTS would be able to mask some of these issues either by explicit manipulation by the musician, or software manipulation (think autotune). Additionally, it is a common understanding that foreign accents are largely masked in song.

I don't know much about the google plugin, but Misaka (the "best" japanese computer generated voice, from what I know) can be found if one searches (certain russian places), which can then be used as some sort of standalone program where you can easily copy/paste or write japanese words/sentences and generate audio with a couple of clicks.

If you're getting words/sentences from audio sources like podcasts or even films/anime, then something like "SolveigMM WMP Trimmer Plugin" for Windows Media Player will easily let you cut out audio segments with a few clicks and save as mp3 files (ranging from perhaps 20-80kb in size) which you can then insert into Anki.

somewhat off topic.. but does anyone know if the voice used for NHK easy news articles is real or tts? I've always thought it was tts. It would be great to have that for example sentences, etc. in Anki. If it's not tts, my apologies to whoever it is that does those recordings, for thinking she sounds like a robot.

jasberg Wrote:but does anyone know if the voice used for NHK easy news articles is real or tts?

She is TTS but I can’t see any reason why you would not want to listen to her. ATST, I would be cautious about not trying to imitate her and for the speaking practise would use audio done by the real people.

jessem Wrote:Also, I'm not super tech-savvy (anki constantly pushes my limits), and I know nothing about TTS. I was trying to google around to see if I could use google's TTS online, without having to start downloading 5 different plugins, and test it out to judge for myself.

I'm sure it's available on Google too, but it took me about 10 seconds to install and use the TTS plugin with Anki. It's only one plugin, you can add it in three easy steps:
http://www.streetsmartlanguagelearning.c...reign.html

yogert909 Wrote:I don't have much specific information, but I would suspect you would end up speaking a little strangely. Even though TTS is getting a lot better, it still sounds strange and robotic. You would probably internalize some of that strangeness if you used TTS extensively as a learning tool.

One of the major problems with TTS is that that aren't very good at replicating tonality and volume modulation which adds extra information that is not printed in the text. For instance, when we ask a question, our voice usually goes up at the end. Better TTS voices deal with this, but I know that there are other problems that haven't been addressed. I wasn't familiar with vocaloids, but a singing TTS would be able to mask some of these issues either by explicit manipulation by the musician, or software manipulation (think autotune). Additionally, it is a common understanding that foreign accents are largely masked in song.

All true, but if the option is between that or no audio, I always go with TTS. Hearing a word read out loud helps memorize it.