This is SCARY but Amazing! Highly Emotional AI Voice Cloning!

[@MattVidPro]

Use code MATTVIDPROAI at the link below to get an exclusive 60% off an annual Incogni plan: incogni.com/mattvidproai Thank you Incogni for sponsoring this video.

[@jboss1073]

It's not the fault of the thing, you wrote AI incorrectly. Don't put a space between them like A. I. That's not how you write A.I.

[@jboss1073]

@@sophia23346 Buy some punctuation.

[@jboss1073]

@@sophia23346 Why don't you use punctuation? It is not English until you abide by all the rules. You can't just write a bunch of words adjacent to each other and think that is English.

[@itzdcx]

ObamNah💀💀💀

[@mopolo]

Once they release an API it's game over for ElevenLabs

[@World-Bass_Mapper]

AI voice technology is getting so good that every time you start a video I wonder if you're trolling us with an AI voice.

[@jooptablet1727]

Haha this is definitely going to be a thing

[@Neatrior]

Noo, please. Don’t be Kwebbelkop lol 😅

[@afti03]

me too

[@reycaribe]

Riiiight. He's still got a sick voice he's just using playHT now😅. This is no coincidence.

[@ikillwithyourtruthholdagai2000]

Ur mentally ill, seek help then

[@spacekitt.n]

i have been waiting for a tool that maps on to another persons voice with all of the inflection and pauses and emphasis and all that. i knew it had to be possible.

[@kaio0777]

yup, i agree waiting for this since 2017 its scary and good honestly.

[@spacekitt.n]

@@kaio0777 it could be used for evil but also what if I want to have all my audiobooks narrated by Alec Baldwin? Or Morgan Freeman. Would be amazing

[@kaio0777]

@@spacekitt.n nah it can be use for good or bad if used in that way, Hell i even use it in my latest podcast episode so i think the tech is great but the risk are very risky.

[@kaio0777]

@@sophia23346 look GOD as his hands full as is but you need to put in the effort and that is why these prophecies are coming true now everything in the bible wasn't one to one but it as being 80% right on the money for the big ones he knew this end game so these thing are happening because he is teaching everyone a lesson about TRUE EVIL and we need to watch an learn follow the path and be smart and keep fath and you will find truth.

[@flyingstapler1241]

Don't SVC AIs do that already? That's how people make all these AI singing covers

[@haroldpierre1726]

This is amazing. I want competition to keep these tools evolving.

[@dekumutant]

​@@sophia23346what

[@Marquis-Sade]

20:11 - Thats because you took an "advertising" voice. This is not angry. Its trying to sell you something.

[@fx-studio]

I looked at this months back and discounted it in favor of 11 Labs. With 11 Labs I output 4 versions of each paragraph, then cut them up and edit the best of each sentence together. Gives a really good result but takes some time.

[@fynnjackson2298]

The emotions ad so much depth, these one model media generation machines of the future merging audio and video and gaming gonna be next level!

[@fynnjackson2298]

Jesus is within us all, or more precisely, we are all within 'Him' @@sophia23346

[@ImNorman]

Is there something you have to do to get access to the advanced features and emotions? I am using the 2.0, but only have the option to change the speed. Tried with Cloned voice and pre-made voices and still don't see those options.

[@gawni1612]

The dithering is smoother than a lot of ai voices, but you can still feel the pixels

[@peterparker6584]

I've been screwing around with their sight for hours I still can't find the emotions section that you showed is I only available if you're a paid member or something?

[@ParameterGrenze]

7:25 The weird way he pronounces Obama makes me think that for a person Matts age Obama is like "Wasn't that a president a long time ago ?"

[@ajfridrich]

My free version does not display the options of Add Emotion or Add Style using 2.0. Nor do I have the Import Video on the top. The voice quality is really quite good. The problem with saying AI is hit or miss. It does seem to also periodically miss or chop off a sentence completely. Definitely considering licensing, know that I will have to work with it to generate the output better.

[@WalterPulido]

At 13:25 I was waiting for the sample from Eleven Labs to start, I realized it was the sample itself when Matt started to talk again.

[@ThePickledDyl]

No way this is second place to 11 labs. This is another level unless my ears are deceiving me.

[@Citrusautomaton]

@@sophia23346Are you implying that AI is going to cause Christ to come back? Will they be buddies?

[@ekenmaice1435]

It's not as reliable as Eleven Labs, some voices in Play HT sound exactly like Humans while others sound like robots

[@xHeadcleanerx]

Can’t clone voices, so not interested.

[@ThePickledDyl]

@@ekenmaice1435 yeah, some voices seem really really good and others are mediocre

[@BladeRunnerEnthusiast]

I've got a habbit of cloning the voices of people like Terence McKenna and having it read off his own lectures, lol

[@danielvalleduarte]

How does it sound? Lol that was going to be the first one I cloned

[@MattVidPro]

This is actually a great idea

[@BladeRunnerEnthusiast]

@@danielvalleduarte it sounds good, though his particular cadence and way of speaking, how certain syllables he extends or sort of.. grunts at? You know if you know. He's a gnome of a man, but it picks up on most of his voice outside of those extended mannerisms.

[@BladeRunnerEnthusiast]

@@MattVidPro it's fun, brings back a life that was lost in a way, though I do feel I should ask his brother for permission but I fear I'll get shut down lol

[@danielvalleduarte]

@@MattVidPro yes

[@joshuashepherd7189]

The reason the Emotion generation didn't work out is because its picking up the punctuation in A.I. It uses the punctuation to induce the emotional detail. So saying "Oh my god, AI has taken over the planet!" would probably generate better results

[@MM-lw6ox]

Quality seems great but the pricing is a bit too steep. 80 bucks before you get to high fidelity clone is too much for the average person. 11 labs takes a freaking month to complete your voice clone

[@jd2161]

Wow they went up. I used them months ago and it was MUCH cheaper

[@RabidDisposition]

Ps5 games are 70 bucks now. 80 realy isnt that crazy these days. You gotta get your bread up so your wages match inflation, or you're gonna REALLY steuggle.

[@Glowbox3D]

This is fun. It's really exciting to see emotions being the push. Looking forward to see what EL does in with the emotion cloning.

[@Yipper64]

4:16 its in that uncanny valley of *trying* to sound natural which makes it sound unnatural.

[@antonystringfellow5152]

Just checking out the website now... you can create and save custom pronunciations too, for such things as acronyms and niche terms. I can see this feature also being useful for fine-tuning your own cloned voice to more accurately reflect your particular regional accent. This is next level stuff!

[@studenteric6930]

There will be a day where I won't tell the difference between real matt and false matt😂❤

[@adamfroman7983]

This channel is THE AUTHORITY for anything AI..Only Pro work here!!!Thanks!!!!

[@mrrfyW]

15:08 sounds like Sonic, lol

[@AltKaxREAL]

Fr lmao

[@s.patterson5698]

PlayHT is pretty good it seems. I might explore that for use with my Unreal Engine characters. I like the emotions you can prompt into a script. Imagine what EMO will do when it comes available.

[@Lu9_ST]

I may be missing something but I believe it might be important to note - I don't think this website is open to everyone around the world just yet. I get a 403 error (that means access denied) every time I try to access it. Wonder if it's currently US-only?

[@TheZanzz27]

Something I noticed, be very careful on you use punctuation. Both periods and commas induce pauses. So, when you have "A.I." in the text, it will add strange little pauses. This seems particularly weighted with the sad emotion.

[@B_Studio841]

It’s good that it says “umm” because it would be so realistic.

[@budyniada9389]

i dont have the style option for my speaking voice anyone know why and how to fix.

[@Allplussomeminus]

10:55 cloning voices that sound African-American has my eyebrows raised. Definitely going to check this out.

[@Bamazon1990]

OMG this is hysterical. It keeps randomly switching between American, British, and deep southern accent lol. How do you tell it to keep to the same accent?

[@jbavar32]

Matt, Im testing our the PlayHT with the free subscription and I don’t see the emotion drop down nor the advanced voice controls. Are you using a paid subscription version?

[@harnageaa]

I think u need to activate account through email.

[@MattVidPro]

I believe it is only rolled out to certain users at this point on the free plan

[@gonkdroid8279]

@@MattVidPro Thanks for the clarification

[@discorabbit]

Please make a vid wearing a leather jacket and round sunglasses while having those headphones on 😂

[@呼]

Crazy

[@jonogrimmer6013]

For some reason I don’t have access to emotions or style and no advanced either? Using the latest 2.0

[@Plan73]

Is there am opensource version of cloning and editing voices, I mean the equivalent of Stable Diffusion but for voices?

[@StPhilofDecatur]

Ok, now I have to subscribe. This solves a major problem for me.

[@Drone256]

This is getting pretty good. What do you think the use cases might be beyond entertainment and call centers?

[@RamiroDelgadoLocutor]

Hello! I just discover your channe in Argentina. For you wich page is the best hiperrealistic clone voice? Thanks!

[@dascodraws6040]

wake me when you have open source.

[@evolvingevrday]

thoughts on creditors and other companies holding very sensitive information utilizing your voice as one of the main factors if verification? Do any tools have private download offline so your voice cannot be found and sold? Perhaps an offline version where you can tweak more than the pitch, or combine with a randomly generated voice to create a profile, people can use without being fearful? (Similar questions for singing AI voice? For Ghost raiders and producers to post content)?

[@abdulbyrd7902]

Does anyone know if playht can clone your voice and then convert it to a different language (sounding like a fluent native), while keeping your unique voice traits?

[@cmw3737]

I don't know if it's just me but the 3rd one narrating the children's story had a distinctive artificial sense to me but all the others were amazing. I was especially blown away by the pauses and er in the first. Matt's cloned voice sounded slightly off except for the sad and angry. I wonder if Matt noticed because he hears his voice differently. I'm getting paranoid about this. When I got a call from an unknown phone call and they just hung up after I said hello? my first thought was what if they just cloned my voice to use in some prank?

[@stevedavis1437]

So the reason for the 2 second jump cuts is that you generate the video, right?

[@MattVidPro]

I’m ocd about my sentences and restate things often

[@HistoryIsAbsurd]

Lol i remember trying to say this was one of the best tools out there a month or two ago on my main and your chat flamed me.

[@IJaykoh]

Dude said “ OBAMNA” lol

[@AltKaxREAL]

SODAAAAA!!!!

[@bengsynthmusic]

It flew over everyone else's ears.

[@thewebstylist]

Awesome having an ElevenLabs competitor

[@Quix94]

Is there no description on this video?

[@MattVidPro]

there is

[@SebSenseGreen]

Woot! Can't wait for companies to make virtual youtubers that can push a thousand videos a day... This won't over saturate the internet at all!

[@cesar4729]

Where we are going we don't need KZbinrs.

[@jd2161]

Cry more. That'll make them stop advancing

[@RabidDisposition]

Thw internet is already over saturated lol

[@64jcl]

We'd just use other AI's to watch the youtube videos and give us a 5 sentence summary. :)

[@Fustercluck06]

Wow this is seriously impressive!

[@anirudhsingh6553]

it said larry (advertisment) thats why it sounded like an advertisement

[@zippythinginvention]

Play had an advantage with the multi-voice script thing. But, didn't ElevenLabs introduce that feature a week or so ago?

[@USBEN.]

I do not have access to emotions in playht. Even with 2.0 model. Weird.

[@AdaptiveDevelopementPlatform]

not sure how you are getting the extra emotional settings... sadly no dice on my end. and I am using v2. oh well.

[@holdthetruthhostage]

Also it doesn't have the Project Maker

[@AM-pq1rq]

seems like typing "A. I." (with a space) was interpreted as the words "A" and "I", plus dots. Maybe if you tried "A.I." or "AI"?

[@MattVidPro]

potentially, this issue absolutely could have been fixed

[@nwokochagospel7260]

The first audio really sounds like Shaan Purris voice🔥

[@cloudpandarism2627]

a channel who begs viewers to subscribe first seconds in the video is the best way to annoy them. if your content is good. we know what to do. maybe something to think about?

[@MsVelvetVelour]

So toxic.

[@MegaBeego]

Lol

[@EmeraldView]

"Just like my narration right now" , while sounding like a robot. 😂

[@marinomusico5768]

Amazing ❤

[@SmirkInvestigator]

Not gonna lie, a Theo Von AI would be hilarious. The day ai it can mimic his unique language and perspectives we should be scared.

[@Lazyliteratus]

This is still a distant second to ElevenLabs. The audio is really tinny. One step above Suno, and a baby step at that. The UI of their site is atrocious. ElevenLabs' own site is cumbersome, but theirs is labyrinthine by comparison. That, and you can't do more than a paragraph at a time. It automatically separates the input into separate dialogue boxes. So, using it for narration is right out. Might be good for convo, but the pay plans are a rip.

[@pietervoogt]

'Are you still an effective team?'

[@spinninglink]

Anyone know if these voice generators can do laughing correctly? I've tried elevenlabs and tried to have it do laughing, and it just sounds horrible and fake lol

[@BlackTakGolD]

11:00

[@manzell]

10:00 are those BOTH ai voices?

[@apache937]

It sounds way different from your recording... but the built in voices are good for sure

[@Legoandcarsgame2016]

No, i did pizza crust the cat’s narrator and it sounded like the person. No mistakes where added

[@egorpro420]

the only minus of play ht is that its limited to 5000 words per month, but its AMAZING overall

[@trevorama]

Dude, get rid of the periods and separate the letters with a space! 😂 I’m gonna call it - Play HT is on par with Eleven Labs.

[@BlackStudies]

To do words like AI, you have to put a space between the letters and not use periods ("A I").

[@arteasa]

Love this!

[@David.Alberg]

Matt whats your Take on AGI? 12-18 months till official release? ^^ Like AI Explained and David Shapiro?

[@joeymurillo5322]

My goal is to clone my different voices save them and use them for certain characters for my channel. 11 labs I heard of I'm still asking around and talking to others

[@pitpatgazorpazorp3356]

good job bro

[@dubshaman]

😂😂😂 AI can't say AI..

[@Pendrokar]

PlayHT is just a AI TTS wrapper. Not sure if they even have their own TTS model. I do hope they do. But otherwise use the originals.

[@marco114]

Seems like putting AI together in caps works best

[@DihelsonMendonca]

Its amazing to see that 11 labs has competition. I never used it because their prices are high for what they deliver. Lets make AI more inclusive, democratic, to the masses for FREE. 🎉🎉❤

[@tabbyxoxoxoxox23]

Hey Matt you've got a typo in the title!

[@MattVidPro]

Thank you! Fixed

[@notacreativename1]

What if I can do impressions really well, can it clone my impression of a celebrity?

[@penguin5704]

The secret to flying is wearing socks on a Tuesday.

[@gabemiller74]

Still sounds very digital, like a very compressed 96 kbps mp3.

[@freedom_aint_free]

Well, to my ears, Eleven Lab's top notch voices are way better than those, and HeyGen dubbing also sounds more legit.

[@leyaud6159]

Amazing 😮

[@holdthetruthhostage]

From the pricing its not winning, also you can only clone 1 high quality voice and no project

[@Bamazon1990]

This is great, but TTS voices still fluctuate all over the place. The only way to make this work is to convert audio to audio so that it uses your own intonation. Also most of their "British" voices are actually American accents. It seems to take a long time to convert basic narration text to audio so this wouldn't be helpful for an audio book. When I tested Fear vs Normal emotion for a bit of non-dialogue narration there was no difference. I think I only a few of the voices have emotions actually installed.

[@daddyshrek2295]

finally, competition

[@JohnDoe-sy6tt]

like the glasses!

[@d1o2n4]

They need to put in an Asperger’s/‘on the spectrum’ slider, especially for Elon. Just turn it up a little

[@ThymeHere]

16h 1,2k loks 225k suns

[@GaryJr530]

matt vid pro! ❤

[@river.]

every ai enthusiast ever: This is SCARY but Amazing

[@timeship]

PlayHT sounds good, BUT it likes to skip words and even entire phrases, jumping erratically for whatever reason that might be. That said, 3 out 5 stars is all I can give it as a writer who seeks to produce sci-fi audiobooks. P.S. It also loves to make crazy and unexplainable voice inflections where childish oversxitemenet is not needed and no exclamation marks exist (!). Good start, I'd say, but it needs improvement (as of November 22, 2023). Maybe for Christmas? P.S. Clipchamp does much better job + no rendering time, no splitting of paragraphs, less overexaggerated emotions, etc. What PlayHit split into 5 paragraphs and took 20 min to render (with skipped sentenmces), Clipchimp did instantly and perfectly correct in one unslipt file. THX

[@safetyhomeloans950]

Update: there is no Dubbing feature. #2. Spanish (Mexican at least) sounds horrible. If your audience is English Speaking its an interesting choice. But for now ElevenLabs for me.

[@DailyPrayerHouse]

hi MATT, How do I join your discord please?

[@MattVidPro]

Link in the description of this video!

[@agnesslovehealz]

Nice gonna try

[@tortysoft]

Good glasses :-)

[@hotsparklab]

Wow! Pondering continued use of AWS Polly or switching to something like this.

[@alexnorth3393]

It's decent but still sounds robotic. 11labs still ahead of the game.

[@mrrfyW]

I looked for the emotions thing and it doesn’t exist in the website Edit: Just realized it’s Alpha lol

[@cofiddle]

Lmao the surprise Theo Von really got me that was hilarious