Live Portrait vs. Hedra: AI Facial Animation Showdown & Tutorial

So many facial animation solutions, so little time!

A solution for facial animation is making the rounds right now, called “Live Portrait.” This can be run within ComfyUI to bring any still portrait image to life by being driven by another video.

The output is reminiscent of what you get from Hedra, which interprets audio files and creates appropriate facial animation showing emotion appropriate to the words.

So even though they operate totally differently, the output of each is worth comparing, and that’s what we’re doing here.

The question is, which solution gives you the results you most want?

We also provide a step-by-step tutorial on creating animations, from preparing audio tracks in Adobe Audition to generating and fine-tuning facial animations using both Comfy UI and Prome AI.

We conclude with tips on achieving better results and showcases how to bring diverse styles and images to life using Live Portrait.

For Installation instructions for LivePortrait on your local system, see this video and consider subscribing to Sebastian’s channel if this is your area of interest:

Resources mentioned in video:

00:00 Introduction to Live Portrait
00:10 History of Facial Animation Apps
00:51 Comparing Hydra and Live Portrait
01:15 Creating Music for Animation
02:05 Voice Conversion Process
04:21 Generating Singer Images
04:43 Animating with Hedra
10:14 Preparing Videos for Animation
11:09 Installing Live Portrait
11:54 Pre-Testing and Initial Setup
12:13 Workflow Overview and Adjustments
13:11 Common Mistakes and Fixes
14:12 Re-Recording and Fine-Tuning
15:58 Final Adjustments and Results

👍 LIKE If you found this video valuable. 🙂
🥰 SHARE If you know someone who might enjoy this video.
⏬ DOWNLOAD or ADD This video to your PLAYLIST for easy access later.
💬 COMMENT Your thoughts and questions are welcome!

That's what keeps me going!

🔔 ​​And make sure to hit the NOTIFICATION BELL to stay updated! 🔔

🌎EXPLORE

❓ASK ME

FOLLOW ME

[OsakaHarker]

You at the end were reducing the retargeting eyes and mouth values, trying to lower them but you had the option to retarget both at 'off' you must turn it on. A great video and result in the end.

[artvsmachine]

Hedra looks cool. I'm going to try it out. I have your "Live Portrait" song stuck in my head now. Over and over and over. Somebody help me.

[Avalon]

It tears too much, Hedra far outclasses this

[SamuraiG]

Good video, been working on a movie for a while but options for character face talk animations it was no good but now with things like hedra I can finally do dialogue conversations correctly.

[TashaCaufield]

*Fantastic* comparison video. Thank you! 💫

[mickymcl]

Amazing video and tech, thank you for demonstrating how it works, there is so much stuff coming out and just imagine what it will be like in another year. Wicked.

[darrenpkruse]

Hendra is clearly easier and took far less time to achieve great results. Even after tweakig the hair is still partially stuck w Live Portrait. Im sold - Hendra.

[freealitytv]

Been using video leap since 2013, love the cap cut too

[axistiltproductions]

So do not waste time remixing or re-generating the static image to make the model look in certain directions, etc... Hedra still considers your prompt in final video

[Metaroids]

Bob I use the LivePortrait via the Pinokio app. It doesn't seem to have the options to adjust lip and mouth movement. Do you think I just didn't find it? Or do you really the comfyui setup to tinker with facial expressions?

[menamariano]

Do you know any software that allows to realtime webcam to photo-face animation? Is MugL the only one ?

[Macatho]

Cool video.
Do you have the comfyUI workflow somewhere?

[theno.idea]

Ok, here goes.. I've been a professional 3D modeler since just about the beginning - It's really a grind and always has been - I've tried all these AI tools - and as cool as they are, and as much as I wish they'd deliver pro standard results - they don't. In 3D model land, Blender, Maya, unbranded tools and others you build or import a close "replica" and then apply the "skin material" that can be manipulated to the "N'th" degree. It really doesn't take much longer than fine tuning these AI tools.. (And the work you do live on your hard-drive, no sharing - EVER. Yes the learning curve is slow - but it's getting better. It's the POST editing that fails in AI production - you cant fiddle with the end render.. I'll keep hoping for AI - but it has a long ways to go.. Thank you - for showing us all this new tech - it really does help us see how to get things done more efficiently..

[REALVIBESTV]

What I don't like about it is that it doesn't provide a 1080p output, and it appears to reduce the overall image quality.

[focehrh]

Hi, i am searching for a model for comfy ui or others ai local pc software that i give image + audio file and it give me a video that is avatar speaker.
do you know any model of this? can you make a tutorial for that?

[GManWrites]

I think you would make a very good Max Headroom.😉

[aivideos]

Increase the scale of the cropper for live portrait. you can capture full hair, hats, animal ears... etc.

[Sean]

Hi, thanks for video. Which of them give higher resolution?

[SUPRMASSIVE]

Live portrait is on Pinokio. That's how I installed it.

[davidking]

I see the headphones in your video and the green screen bleed is affecting the hair... have you noticed a difference in the quality of the eye movement if you record yourself without glasses on...?