Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Amazon Polly – Lifelike Text-To-Speech

aws.amazon.com

Post reply on HN
Print view View original

51–60 of 106 posts

Re: Amazon Polly – Lifelike Text-To-Speech

#51

Post by derefr »

angstrom wrote: post #42
hacker_9 wrote: post #25

Earlier quoted context omitted.

Additionally when we speak we don't just 'replay' words; we add emotions and so on. Reminds me of this video [1] analysing how the actor Anthony Hopkins converts his lines to speech for his scenes. [1] https://www.youtube.com/watch?v=4kSGkGKwp9U

The spoken word has so many nuances. Cadence, inflexion, tone, and then the facial performances are never captured by pure text to speech. You'd almost have to be able to set boundaries for seriousness, playfulness, innuendo, indifference, sarcasm etc because none of that is conveyed in the literal text without additional descriptions or contextual analysis of the conversation.

It'd be interesting to take the Joint Many-Tasks approach here, training some of an RNN's layers on text-to-speech, and then another set of layers on sentiment analysis of speech-audio, where the error of the sentiment layer can backpropagate into the TTS layer.

Re: Amazon Polly – Lifelike Text-To-Speech

#52

Post by TazeTSchnitzel »

rckclmbr wrote: post #44
derefr wrote: post #13

One thing I always still notice about these "lifelike" speech models is that they still have random pitch variation that wouldn't be present from a real speaker—sort of a "warbling", similar to audio heavily compressed by some cellular realtime audio codecs. I assume this is due to random variation in the pitches the speakers in the training data used to say given phonemes. Would this, then, imply that speech models…

Southerns have an accent. New Jerseyans have an accent. British have an accent. Can't computers have an accent, too?

Wouldn't we describe what TTS systems do as more of a speech impediment? Sure they use a particular accent, usually that of a standard prestige dialect. But it's not the accent that's weird.

Re: Amazon Polly – Lifelike Text-To-Speech

#53

Post by martythemaniak »

zackbloom wrote: post #47

Here's a sound clip of one of the voices I pulled, if you'd like to hear it: https://soundcloud.com/zack-bloom/amazons-new-text-to-speech...

The demo files in the article sound pretty decent, but this sample is pretty bad. It sounds pretty robotic by itself (lots of artifacts, odd pitches, no pauses), but saying "during the one thousand nine hundred and seventies" is the real kicker.

Re: Amazon Polly – Lifelike Text-To-Speech

#55

Post by revelation »

splike wrote: post #41
mwcampbell wrote: post #40

Earlier quoted context omitted.

Text-to-speech systems have been expanding abbreviations for decades, and it always backfires in some cases. For example, DECtalk would expand "Sun" to "Sunday" regardless of context, with the humorous result that blind people reading tech news in the 90s would often hear about Sunday Microsystems. Edit: Yes, that example was unfair, because it's from the 90s (and actually, DECtalk was largely unchanged since the 80s…

Hence the need for intelligence

Which it fails miserably at. They probably picked the one example that worked.

Someone linked another example here:

https://soundcloud.com/zack-bloom/amazons-new-text-to-speech...

And it pronounces 1970s (the decade) as one-thousand-nine-hundred-seven-ths.

Re: Amazon Polly – Lifelike Text-To-Speech

#57

Post by toephu2 »

derefr wrote: post #13

One thing I always still notice about these "lifelike" speech models is that they still have random pitch variation that wouldn't be present from a real speaker—sort of a "warbling", similar to audio heavily compressed by some cellular realtime audio codecs. I assume this is due to random variation in the pitches the speakers in the training data used to say given phonemes. Would this, then, imply that speech models…

Chinese (Mandarin) text to speech still shares the problem of not sounding natural (sounds robotic), but it is a little better than English.

Example (click on the blue speaker icon): http://dict.baidu.com/s?wd=%E7%88%B1%E5%B1%8B%E5%8F%8A%E4%B9...

Re: Amazon Polly – Lifelike Text-To-Speech

#58

Post by PaulRobinson »

revelation wrote: post #55
splike wrote: post #41

Earlier quoted context omitted.

Hence the need for intelligence

Which it fails miserably at. They probably picked the one example that worked. Someone linked another example here: https://soundcloud.com/zack-bloom/amazons-new-text-to-speech... And it pronounces 1970s (the decade) as one-thousand-nine-hundred-seven-ths.

You're not wrong, but that's still the best real-time TTS sample I've ever heard.

Re: Amazon Polly – Lifelike Text-To-Speech

#59

Post by jldugger »

derefr wrote: post #13

One thing I always still notice about these "lifelike" speech models is that they still have random pitch variation that wouldn't be present from a real speaker—sort of a "warbling", similar to audio heavily compressed by some cellular realtime audio codecs. I assume this is due to random variation in the pitches the speakers in the training data used to say given phonemes. Would this, then, imply that speech models…

If you grab the training set for CMU-Sphinx, you'll see that the voice actors providing base input are trained to speak that way.

At a secondary level, I imagine there's a sort of pitch trajectory for a sentence, and if a word or syllable ends too far off that trajectory, the return to the pitch trajectory is fast enough to be easily detectable.

Re: Amazon Polly – Lifelike Text-To-Speech

#60

Post by mgamache »

There's a lot of negative reaction to Polly. Besides WaveNet (which is not available as an API-- and really slow), what are better alternatives? Or is the reaction just to Amazon's marketing speak using "lifelike" and "deep learning"? I find most of the voices to be okay, but Joanna is pretty good and better then anything I've heard besides WaveNet.

Also, with Amazon's FPGA investment how long before it implements WaveNet using FPGA's to reduce the speech generation time to something that is realtime?

Post reply on HN