Play.ht for me has been amazing, it has emotion control and the quality output it's amazingly good, also the interface is very optimised to produce a whole script in one go in separate dialogs, I grew to 2 million followers using their service and now I'm very sad that they announce they'll shutting down and while looking for an alternative I came with this post and THANK YOU, MiniMax has a very close quality output to what I was looking for. I don't mind to pay, just that the AI voice of my brand is a remix of a popular elevenlabs voice and eventually they asked me to verify it and of course I couldn't. I'm so mad and frustrated at play.ht
Yeah it's funny, because PlayHT actually did quite well in my original comparison, but somehow was quite subpar in this specific test. Then again, I can't rule out that the paid version does a much better job than my free test would indicate. Happy you find this useful and hope MiniMax serves as a passable replacement for you!
For me PlayHT was the best one, they had emotional control even before elevenlabs just that they didn't marketed like that as you couldn't control it and also, sometimes, voice generations came with a little laugh (depending on the context) that seem so organic and authentic, I haven't seem anything like it. For months I felt that PlayHT was an undervalued gem and I just find out that they shut down because Meta bought them. Anyway thank you so much again for your contribution, you just got a new follower 🙏
Yeah that really sucks to hear! But thankfully there are many other options out there nowadays and the TTS models are only getting better. I appreciate the follow, welcome aboard!
Looks like I’m not the only one. PlayHT was by far the best voice cloning tool, so I’m surprised that it only gets 6/10. Then again, this doesn’t seem to be a full testing of that tool. Anyway, that doesn’t matter anymore since they’re shutting down. Tried MiniMax and it’s just far inferior compared to PlayHT. Still looking for an alternative.
Yeah, I had better prior experiences with PlayHT before as well, so perhaps this one-off test didn't quite do it justice. But like you said, it's being phased out, so that's a moot point.
Someone in the comments mentioned Naturalreader, and in my brief test, the voice clone sounded quite a bit like me, although the audio quality was a bit underwhelming.
So if you haven't already, give it a spin and see if it does the trick!
Btw, I appreciate that you included actual samples from the tools you tested. I don’t recall seeing that in other articles. As constructive feedback, though, it would’ve been even better if the samples were the best outputs after multiple regenerations, so we could compare the tools at their best.
Oh, I’d also like to add that Cartesia, in terms of voice cloning, is a solid alternative. It’s better than most I’ve tried, though still inferior to Play.HT.
Haven't heard of it before, but after just looking it up, it seems to be an open-source model by Resemble AI, which was on my "rejected" list because of its paywall and because the 5-second preview sample didn't sound like me. Maybe the standalone open-source model is better. Have you been using it? What have your impressions been?
Hi Daniel! I've not tried it but heard it's a good alternative to ElevenLabs. No free plan although a pay as you go that seems cheap and the open source option that can be a hassle for many...
Ah, got it! I'm sure the more tech-savvy people actually prefer the open-source option to incorporate into their own projects, but like you said, it's probably overkill for most average users like me who just want something that works off-the-shelf.
Just briefly ran it through the same test, and it suffers from a bit of that same "muffled telephone line" syndrome as some entries on this list. But it sounded quite close to me, so it'd probably end up at around 7/10 in my above grading. (EDIT: Apparently, it also doesn't let you download the resulting MP3 Audio on a free account, so that's another notch against it for our purposes.)
Thanks as always for an excellent review. My own personal response to all these reviews is that unless I have a compelling reason to use any of these services now, my choice is just to patiently wait while they continue to improve.
Evan Ratliff of Shell Game used his AI voice extensively in season 1 and it was really good. He would call his friends (and wife) to talk and they would be fooled for a bit. But, he had hours of podcasts to train it on and I gotta think that helps. I think he used ElevenLabs.
He did, and I believe he mainly used Vapi to power the voice agent. I listened to the entire season #1 after your recommendation, and it was a fun ride!
Yeah I recall a call center-ish app at the middle of it all orchestrating the calls and recording and such. Maybe even an interview with Vapi CEO iirc. Are they still around?
Relatively. I actually am a paying customer to eleven labs now in part due to what you’ve written about them. I’ve been using them to “spruce up” some podcast stuff I’m doing. You should get a commission! How is you?
Play.ht for me has been amazing, it has emotion control and the quality output it's amazingly good, also the interface is very optimised to produce a whole script in one go in separate dialogs, I grew to 2 million followers using their service and now I'm very sad that they announce they'll shutting down and while looking for an alternative I came with this post and THANK YOU, MiniMax has a very close quality output to what I was looking for. I don't mind to pay, just that the AI voice of my brand is a remix of a popular elevenlabs voice and eventually they asked me to verify it and of course I couldn't. I'm so mad and frustrated at play.ht
Yeah it's funny, because PlayHT actually did quite well in my original comparison, but somehow was quite subpar in this specific test. Then again, I can't rule out that the paid version does a much better job than my free test would indicate. Happy you find this useful and hope MiniMax serves as a passable replacement for you!
For me PlayHT was the best one, they had emotional control even before elevenlabs just that they didn't marketed like that as you couldn't control it and also, sometimes, voice generations came with a little laugh (depending on the context) that seem so organic and authentic, I haven't seem anything like it. For months I felt that PlayHT was an undervalued gem and I just find out that they shut down because Meta bought them. Anyway thank you so much again for your contribution, you just got a new follower 🙏
Yeah that really sucks to hear! But thankfully there are many other options out there nowadays and the TTS models are only getting better. I appreciate the follow, welcome aboard!
Looks like I’m not the only one. PlayHT was by far the best voice cloning tool, so I’m surprised that it only gets 6/10. Then again, this doesn’t seem to be a full testing of that tool. Anyway, that doesn’t matter anymore since they’re shutting down. Tried MiniMax and it’s just far inferior compared to PlayHT. Still looking for an alternative.
Yeah, I had better prior experiences with PlayHT before as well, so perhaps this one-off test didn't quite do it justice. But like you said, it's being phased out, so that's a moot point.
Someone in the comments mentioned Naturalreader, and in my brief test, the voice clone sounded quite a bit like me, although the audio quality was a bit underwhelming.
So if you haven't already, give it a spin and see if it does the trick!
I’ll give it a try. Thanks!
Btw, I appreciate that you included actual samples from the tools you tested. I don’t recall seeing that in other articles. As constructive feedback, though, it would’ve been even better if the samples were the best outputs after multiple regenerations, so we could compare the tools at their best.
Oh, I’d also like to add that Cartesia, in terms of voice cloning, is a solid alternative. It’s better than most I’ve tried, though still inferior to Play.HT.
Have you tried chatterbox?
Haven't heard of it before, but after just looking it up, it seems to be an open-source model by Resemble AI, which was on my "rejected" list because of its paywall and because the 5-second preview sample didn't sound like me. Maybe the standalone open-source model is better. Have you been using it? What have your impressions been?
Hi Daniel! I've not tried it but heard it's a good alternative to ElevenLabs. No free plan although a pay as you go that seems cheap and the open source option that can be a hassle for many...
Ah, got it! I'm sure the more tech-savvy people actually prefer the open-source option to incorporate into their own projects, but like you said, it's probably overkill for most average users like me who just want something that works off-the-shelf.
Awesome review! I've played around with Naturalreader and enjoyed it but I will definitely check a few of these out.
Nice, thanks for the tip!
Just briefly ran it through the same test, and it suffers from a bit of that same "muffled telephone line" syndrome as some entries on this list. But it sounded quite close to me, so it'd probably end up at around 7/10 in my above grading. (EDIT: Apparently, it also doesn't let you download the resulting MP3 Audio on a free account, so that's another notch against it for our purposes.)
Thanks as always for an excellent review. My own personal response to all these reviews is that unless I have a compelling reason to use any of these services now, my choice is just to patiently wait while they continue to improve.
"While the voice is vaguely Daniel-ish, the narrator sounds like he was told to finish his lines before a bomb goes off."
hahahahahaha...I really often look at your posts and laugh loudly!
Happy to entertain, I do try to keep things light!
Evan Ratliff of Shell Game used his AI voice extensively in season 1 and it was really good. He would call his friends (and wife) to talk and they would be fooled for a bit. But, he had hours of podcasts to train it on and I gotta think that helps. I think he used ElevenLabs.
He did, and I believe he mainly used Vapi to power the voice agent. I listened to the entire season #1 after your recommendation, and it was a fun ride!
Yeah I recall a call center-ish app at the middle of it all orchestrating the calls and recording and such. Maybe even an interview with Vapi CEO iirc. Are they still around?
The site is very much around but I’ve never used AI voice agents, so I’m not sure whether Vapi is the gold standard or not.
This post hit me at exactly the right time. Thank you!
Awesome to hear that! And great to hear from you, it's been ages - hope you're doing well?
Relatively. I actually am a paying customer to eleven labs now in part due to what you’ve written about them. I’ve been using them to “spruce up” some podcast stuff I’m doing. You should get a commission! How is you?
I'll make sure to demand my ElevenLabs commission ASAP! I'll drop you a private message shortly.
This is how I’m using AI
https://open.substack.com/pub/hamtechautomation/p/a-battle-tested-sredevops-engineers?r=64j4y5&utm_medium=ios