Nothing replaces, at this point in time, your lovely and natuarl presentation style. One day, maybe but not today. Thanks for the great video inspiration and information.
Bigger plot twist: This deepfake software existed before Jenn started her UA-cam channel & she started the channel using the 'Jenn' Ai Avatar that you see in all her videos. Jenn is actually a fat, bald, middle aged man named Todd. This deep fake video was intended to look less real than her actual Ai Avatar that she uses in all her videos so as to not expose her(Todd) as the fraud that she(he) is. You aint fooling me Todd.
What is cool is that both Synthesia and Descript Overdub is just the first generation of these. I am sure in a couple of years you can make these avatars at home with your phone as a camera for training and massively increased variation in the expression of your avatar. From what I hear Synthesia now supports some control over blinking, eye movement and other things as commands in the transcript used along the sound. Similarly many AI sound tools (at least the Google ones) support special tags to emphasize or put different emotions in to how words a spoken. Those can make it more varied and perhaps more believable too. We live in exciting times.
@@JennJager Excellent video, providing excellent content. I've been letting GPT-3 build a business plan which I'm carrying out in an experimental capacity, and your content has inspired me to dive far deeper into Synesthesia and Descript.
Great! Now I'll finally be able to create my avatar and let it watch UA-cam videos on my behalf. Now I understand why some people are saying that robots and AI can be liberating:)
@@medialex78 Pretty much, seeing it's more or less the only source of information for me at the moment. I haven't been watching TV or listening to radio for decades now, and unfortunately UA-cam is the best single source of information even considering all the flaws and censorship. And I'm not prepared to browse through several websites every day just to stay moderately informed. So I'd really be happy if I had some sort of AI "helper" which would create a list of relevant links. That would save some time for other things and projects I'm involved in.
@@johnsmith-000 I see. In this case you may have to feed the AI with some information about your personal preferences. Unfortunately I am not deep enough into this stuff to explain to you how this will work, but similar systems are already in use.
Yeah, even the _start_ of the audio sounds a bit wonky. It sounds like it _wanted_ to use your name's avatar to be referenced as a possessive, but the S seems to be missing from your name.
Thanks for doing these. I was looking at several brands of text-to-speech (Speechelo is one but there are many more.) As I browsed around, one of them had three tiers, like free, $10 and $30, sort of like that. The top tier, the "Our Best AI" tier, was WAY better for text-to-speech than the others. So the "best possible" current text to speech is now really impressive, with all sorts of very subtle pauses and intonations, and a lot of other lesser grades from different companies are just "kind of impressive" like in this video.
Bruce, would you please tell us what TTS "was WAY better for text-to-speech than the others". I unsuccessfully search to find which TTS you referred to. Novelai was the only one I thought that it might be. I have Speechelo, but not thrilled with it. Haven't tried Murf yet.
@@dennism5242 Again, I didn't keep notes but I did find one that had a much better top price tier. It may be Murf which has three prices, free, $13, $26. But I'm not sure. By my browser history I had looked at Speechelo, Murf, SpeechVid, Descript.
IT seems like every time I create a video with an avatar when I generate, it goes into "manually moderated" status, and its not controversial content at all. I can take hours even during business hours to get a video approved. I find it incredibly annoying.
it is going to replace everyone.. in few months.. and they are going to be super popular because its no point practising being perfect actor. its going to be smashing this is too early.. but in one year it's going to be much better than real you in every way
Everyone practicly already is an avatar, since everyone copies eachother on everything anyway. How people present content on youtube like this person for example look like a bot. How people edit things. How people dress, act, talk, eat, walk everyone is actually a poser.
Some companies are beginning to be given biometric data for things like this. Data breaches of even major companies that expect to have impregnable security have had customer's data stolen. Just sayin.
You’re more animated and have a warmer presence IRL.
Nothing replaces, at this point in time, your lovely and natuarl presentation style. One day, maybe but not today. Thanks for the great video inspiration and information.
in less that one year these avatars will be more interesting and charming than real you
Plot twist: Jenn made this whole video using her deep fake
And the avatar is in fact the real her
@@raizenman Or Jenn does not exist at all. Just like (supposedly) Bielefeld in Germany...
@@medialex78 made my day :-) only germans understand this. 😆😁
@@OliverFlossdorf Jawohl! Und seh´n wir uns nicht in dieser Welt, dann seh´n wir uns in Bielefeld...
Bigger plot twist: This deepfake software existed before Jenn started her UA-cam channel & she started the channel using the 'Jenn' Ai Avatar that you see in all her videos. Jenn is actually a fat, bald, middle aged man named Todd. This deep fake video was intended to look less real than her actual Ai Avatar that she uses in all her videos so as to not expose her(Todd) as the fraud that she(he) is. You aint fooling me Todd.
What is cool is that both Synthesia and Descript Overdub is just the first generation of these. I am sure in a couple of years you can make these avatars at home with your phone as a camera for training and massively increased variation in the expression of your avatar. From what I hear Synthesia now supports some control over blinking, eye movement and other things as commands in the transcript used along the sound. Similarly many AI sound tools (at least the Google ones) support special tags to emphasize or put different emotions in to how words a spoken. Those can make it more varied and perhaps more believable too. We live in exciting times.
Interesting. But the biggest issue is that the video does not sync the audio
Love it Jenn!
Nice! Are there any cheaper alternatives of custom talking avatars (personal body AND speech combined)?
Nice Jenn!
Oh, no. Ohhhh YES! I'm so glad we took a dive right into this.
What did you think?
@@JennJager Excellent video, providing excellent content. I've been letting GPT-3 build a business plan which I'm carrying out in an experimental capacity, and your content has inspired me to dive far deeper into Synesthesia and Descript.
Juts found your Chanel. Love it.
What program I can use to replace translate voice to an existing video?
I think the avatar is compelling. Having watched this and the Descript video, I still think both videos with Descript make you sound a bit robotic.
So can you dress your avatar in different outfits or change your hairstyle? :P
I thought the sync between the audio and video was way off.
Nice Jenn!👍🙏🌹🌹🌹🌹
Great! Now I'll finally be able to create my avatar and let it watch UA-cam videos on my behalf. Now I understand why some people are saying that robots and AI can be liberating:)
So you feel somehow locked into UA-cam content?!
@@medialex78 Pretty much, seeing it's more or less the only source of information for me at the moment. I haven't been watching TV or listening to radio for decades now, and unfortunately UA-cam is the best single source of information even considering all the flaws and censorship. And I'm not prepared to browse through several websites every day just to stay moderately informed. So I'd really be happy if I had some sort of AI "helper" which would create a list of relevant links. That would save some time for other things and projects I'm involved in.
@@johnsmith-000 I see. In this case you may have to feed the AI with some information about your personal preferences. Unfortunately I am not deep enough into this stuff to explain to you how this will work, but similar systems are already in use.
It’s not you. Jenn, we all know you can’t talk without waving your hands around.
Lmao, I can’t help myself
@@JennJager yup, she Moves Like Jager
You said it was expensive to create your own avatar, what was the cost?
It's 1000USD
@@angjelique thanks! 👍🏻
@@angjelique For a corporate training tool, that is not that expensive.
@@deathdisco89 After a year or two of wearing the same clothes in every video, folks will wanna change it up!
Yeah, even the _start_ of the audio sounds a bit wonky. It sounds like it _wanted_ to use your name's avatar to be referenced as a possessive, but the S seems to be missing from your name.
Thanks for doing these. I was looking at several brands of text-to-speech (Speechelo is one but there are many more.) As I browsed around, one of them had three tiers, like free, $10 and $30, sort of like that. The top tier, the "Our Best AI" tier, was WAY better for text-to-speech than the others.
So the "best possible" current text to speech is now really impressive, with all sorts of very subtle pauses and intonations, and a lot of other lesser grades from different companies are just "kind of impressive" like in this video.
Bruce, would you please tell us what TTS "was WAY better for text-to-speech than the others". I unsuccessfully search to find which TTS you referred to. Novelai was the only one I thought that it might be. I have Speechelo, but not thrilled with it. Haven't tried Murf yet.
@@dennism5242 Yeah, what I said was true, but I didn't keep any notes and I looked at 6 or more. So Sorry!
@@dennism5242 Again, I didn't keep notes but I did find one that had a much better top price tier. It may be Murf which has three prices, free, $13, $26. But I'm not sure. By my browser history I had looked at Speechelo, Murf, SpeechVid, Descript.
@@brucequinnplayground2114 Thanks, Bruce, especially for the extra effort. I'll compare those you mentioned.
Looking forward to best voice sampling!
Crazy but we are not there yet.
IT seems like every time I create a video with an avatar when I generate, it goes into "manually moderated" status, and its not controversial content at all. I can take hours even during business hours to get a video approved. I find it incredibly annoying.
When you speak, your voice goes up, just like the avatar's.
So You are sester of Mig Jager...??
Falls into the Uncanny Valley for me - en.wikipedia.org/wiki/Uncanny_valley
The last thing I need is another me.
Eventually this will replace newscasters
"If you have to ask how much it costs, you can't afford it..."
Brillant
it is going to replace everyone.. in few months.. and they are going to be super popular because its no point practising being perfect actor. its going to be smashing this is too early.. but in one year it's going to be much better than real you in every way
Everyone practicly already is an avatar, since everyone copies eachother on everything anyway. How people present content on youtube like this person for example look like a bot. How people edit things. How people dress, act, talk, eat, walk everyone is actually a poser.
Some companies are beginning to be given biometric data for things like this. Data breaches of even major companies that expect to have impregnable security have had customer's data stolen. Just sayin.
I didn't leave any negative comment here
Definitely uncanny valley territory - it really doesn't take much to get there. Humans are weird like that.
Somebody knowing you would notice, but others not
Still uncanny valley enough to notice that is isn’t real.
It doesn't seem real. It sounds robotic and if I had to listen to that voice over a long period of time I'd probably shut it off.
Nice try Jenn. But maybe don't use your avatar on this video. Almost fooled me.
It's you but it's not you. Your speech inflection is too different.
I don't like it... it's scary...