Peregrine: True Realistic Text to Speech Model by Play.ht(play.ht)
play.ht
Peregrine: True Realistic Text to Speech Model by Play.ht
https://play.ht/blog/introducing-truly-realistic-text-to-speech-with-emotion-and-laughter/
5 comments
This is amazing. Holy moly. The first sample audio clip could easily pass for a clip from an audiobook. And without knowing about this you'd find it hard to convince me that the Joe Rogan clip isn't actually Joe
Very impressive... except the voice cloning does nothing to correct the effect when a person says words they wouldn't normally say in a completely fluid way, it sounds perceptibly wrong. The Elon Musk example is pretty jarring in the second half. The JFK wasn't bad though.
Their pricing model is kind of abysmal. They use words per year. Who has a realistic understanding of how many words they might want read per year. It's also surprisingly low. 480 pages per year for $171 per year. Just a bit over. page and a third per day. Basically their entry level and medium tier are "teaser" plans, pushing to the $891 plan that has the sort of actually useful level of quota.
They likely have to charge like this due to the hardware involved in running the model, but its also really unfortunate since its a huge disincentive to try out the lower plans, I'm not about to pay to try and use a tool that even if its good enough I want to use it, will cost too much to use.
Their pricing model is kind of abysmal. They use words per year. Who has a realistic understanding of how many words they might want read per year. It's also surprisingly low. 480 pages per year for $171 per year. Just a bit over. page and a third per day. Basically their entry level and medium tier are "teaser" plans, pushing to the $891 plan that has the sort of actually useful level of quota.
They likely have to charge like this due to the hardware involved in running the model, but its also really unfortunate since its a huge disincentive to try out the lower plans, I'm not about to pay to try and use a tool that even if its good enough I want to use it, will cost too much to use.
Really very good samples
This is scary good
We believe in a future where all content creation will be generated by AI but guided by humans, and the most creative work will depend on the human ability to articulate their desired creation to the model.
This reads to me as if everyone should have the chance to be a movie producer, but people who write/ act/ paint/ play music are somewhat suspect, as if they were talent monopolists or something. I doubt this is what was intended. It's great to have it available as a tool such that imagination rather than financial resources is the primary engine of creative works, but a world in which there's no particularly incentive to get good at things because something better can always be synthesized on demand sounds like an anomie factory.