SadTalker: Bring Your Image to Life with Audio | Audio to Animation 

wow great video thanks for sharing. keep up the good work.

Thank you for such a comprehensive tutorial. Would also love to see the improved version update that you mentioned @13:10 And is there any other tool to transfer facial expressions from one talent to other, rather than just giving the audio to produce those facial expressions?

2 years ago i needed this for my video and i found it ... yeah

Thanks man love your channel never get disappointed ❤❤❤❤❤❤

Just love your videos. Always very useful, interesting, and informative

many thanks for this work. Wonderful discovery for me. thanks again

Can this be used to stream real time responses with a chatbot?

Great Video. Do you have a video on running it in your local machine? Also do you have a video on using an API liek you mentioned @ 9:53 in your video?

Interesting! Thanks. I had tried the Collab several times and failed. I wonder if that's because i didn't copy it to my Google drive? It wouldn't notice the images nor audio I added.

can we use this generated video on social media or personal youtube channel...? is this alternative of what you showed yoyo-nb/thin-plate-spline-motion-model

Thanks for the video. Have subbed and liked vid. I have a Q, for Google Colla free plan, what's the maximum amount of audio it can process. For example if I have a 1-hour audio, can I still use the free plan?

Please make video more projects like this, or any upcoming project। And project showed at last in video

Wonderfull video and content, thank you. I would be very interested in further detail regarding improving the lip-sync.

How does this work with a wider or taller image with a face, e.g. a talking character in a scene? Is it as good as D-ID to still detect the face?

Is there a limit to the length of the animation? Could you upload a 3min audio?

The licence says opensource and free for commercial use too. Why we should read all the content of the code searching for a disclaimer that say opposite? In my opinion when I use a code is my responsibility only to read the licence and don't violate it. It is not my responsibility to read all the content of the code. Why do you think that this disclaimer putted somewhere into the code package has some kind of legal value?

i tried on huggingface but it says error when generating

Is there a way to use Pose style, Batch sizes or 512 model in Colab? My face images are on an angle and I get some weird warping.

Where can I ask questions about sadtalker installation?

excellent can I incorporate this in a Streamlit python application if yes how to go about doing that ?

Please make a Video on the last part , where we can enhace the output of sadtalker. make a separate video.

I use it in Automatic1111. What do you suggest for the photo pose? Mouth closed or open? Thanks

Please share info about your malicious detection. It's one thing to yell DANGER but a better thing to explain why that's so.

Can the SadTalker software work in real time? That is to say, if the sadTalker were to receive a string of text, could it manipulate the face in real time, or does the text string need to be processed on-line by remote server? Please advise.

How long can a video be and how do you download it when it's done?

espectacular, justo lo que estuve buscando

Anyone knowing how much time it takes to generate that kind of video approximately (assuming you have a decent CPU and GPU, like i7 and RTX2070 for example) ? I can't find this anywhere, and as I'm testing several solutions, I'd rather find the information than trying all of them 😅

thanks for the kind video. may i ask the differnce between running it on google colab and hugging face?

Can we make it making the video faster so it will be suitable for live applications? Is there is any alternative model that can do the same in real time?

Is there a limit on the audio length?

Is there anyway to do that but with a facless/masked character for example deadpool or spider man ?

I’m not very tech aware and just tried this, however I can’t get the audio file to upload to sad talker. Any idea why?❤️🇦🇺

is there feature of text to audio still remain ? i am not able to see it, someone guide me.

How do you move the shoulders, only moving the face is very unnatural, very often I bounce when I see an avatar only taks with face without moving the body

I got error on Animation (colab) /bin/bash: line 1: python3.8: command not found

I created a copy for Drive and the first video I made was successful, about 1 and a half minutes long. However, even repeating all the processes, I was no longer able to generate even a 30-second video. The error that appears to me is when executing the fifth cell, when I want to see the preview and download it. It simply says the following: "IndexError: list index out of range". I don't know what else to do, if you can help me I would appreciate it in advance.

You used AI processing on your face in the intro yes?

sir how to get the full sized image? it's cropping my image

I can't find the text to voice

I get error messages in the final steps, when I try to run the code.

I keep getting this error. /bin/bash: line 1: python3.8: command not found when trying to run the animation section. Any ideas how to fix it?

Thank you for sharing this information sir..can you please suggest easier way ..😢

Any voice cloning model

keeps giving the error "This application is too busy. Keep trying!"

Mrci pro

how do you run this locally? or what are the implications of running it in collab?

Moist yet semi-creamy......

When I create the avatar video, at the end, I see two icons of a face and a comment at the top. How can I delete that? Thank you.

cool BUT the only runtime option a got was python 3, and after I selected the image I wanted to use, the animation script errored out with "/bin/bash: line 1: python3.8: command not found"

guys which is faster?

existe algún otro proyecto que permita el uso comercial?

It's an absolute bitch to install in A1111. Fucking nightmare!

Thank you for the instruction. Please make a video of upscaling ai video for free.

if the image is cartoon,how it work? like animal

yeah this is dumb how to add voice movement to a video from audio file

How to Run Sadtalker on an AMD GPU

not working

Just saw this tutorial is 2 months old... in AI terms that is 2 years old... Sadly outdated as most AI tutorials become if you blink haha... sure worked great a month ago or so... currently all local running methods are non-functional...

I don't know how long this is supposed to take but i have been waiting an hour for a 4 second video. Useless to me.

huh

try again

AI animation, like TTS and voice cloning, has yet to evolve. 14 years ago, we had better tools to achieve this.

Just ruined the 666 likes, sorry!

i got this error: examples/source_image/full3.png /bin/bash: line 1: python3.8: command not found

google colab how to solve this please mp4 name = glob.glob('./results/*.mp4')[0]

huh

huh