Hedra lets you give AI characters a voice — this new model is a huge step up
Impressive lip-synching
Hedra, the AI character generation platform has launched a new version of its model and it includes even more realistic head movements and facial expressions.
There is also a new ‘stylize’ feature that lets you put your character in different outfits, add backdrops and even change their face to a Lego character or anime.
Character is a new family of foundation models from Hedra designed to create more consistent and realistic humans through AI video, giving creators more control over the final output. Initially, the focus is on lip-syncing and head animation. Version 1.5 adds to that capability.
Hedra offers a generous free plan for people to try it out, and I’ve found it works with any human image including photos, drawings and even action figures. It animated both the head and face fairly accurately in time to the words used in the voice track.
Putting Hedra Character 1.5 to the test
I created five characters, first through an image prompt with Flux.1 running locally on my laptop, then using ElevenLabs to create a voice for each character. Finally, I used the voice and image with Hedra to generate a video.
Each image is of a person looking straight to the ‘camera’ with each different enough to be a useful test. I then ran stylize on two of them, pushing it to see completely different looks, as well as a subtle change to the outfit and background. Beneath each description, you'll see the full prompt I gave to Flux.1 to create each character.
1. The doctor
First, we have a middle-aged doctor wearing a lab coat. I then stylised it with a bright over-the-top outfit full of splashes of color with a colorful background.
Sign up to get the BEST of Tom's Guide direct to your inbox.
Get instant access to breaking news, the hottest reviews, great deals and helpful tips.
It hasn’t done a bad job. There are some artifacts around the mouth and the head movement is a little exaggerated but it reflects the tone of the sound well. The voice matching isn’t AI, I found a suitable voice on ElevenLabs.
A head-and-shoulders shot of Dr. Amelia Chen, a 45-year-old Asian-American female doctor, speaking directly to the camera. She's wearing a white lab coat over light blue scrubs, with a stethoscope draped around her neck. Her long black hair is neatly tied back, revealing a few strands of grey at her temples. Dr. Chen's expression is warm and reassuring, with subtle laugh lines around her eyes as she speaks. The background is slightly out of focus, showing a clean, well-lit hospital corridor. The lighting is soft and professional, emphasizing her facial features and the sincerity in her brown eyes as she addresses the viewer, likely explaining a medical concept or providing patient advice.
2. The builder
Next, we have a builder character. This isn’t a test of the AI image generator but rather of Hedra's lip-syncing and head movement animation capabilities.
The head movement is more natural here with minimal artifacts but the blinking is unnatural. It is still a marked improvement on Hedra Character-1 and some other AI lip-syncing tools.
A head-and-shoulders shot of Marcus Johnson, a 38-year-old African-American male construction worker, speaking directly to the camera. He's wearing a yellow hard hat and an orange high-visibility vest over a grey t-shirt. Marcus has a strong jawline with a 5 o'clock shadow, and a small scar above his right eyebrow. His expression is confident and friendly as he talks, likely explaining a aspect of his work. Sweat beads on his forehead, and there's a smudge of dirt on his cheek, suggesting he's been actively working. The background is blurred but shows the vibrant blues and oranges of a construction site. Natural sunlight illuminates his face, casting small shadows that accentuate his features.
3. The barista
The barista character had the most natural blinking of the tests I ran. She again had slightly exaggerated lip-movement but overall it was a good render. It can maintain the friendly demeanor of the initial image and the voice.
A head-and-shoulders shot of Sofia Rodriguez, a 25-year-old Hispanic female barista, speaking directly to the camera. She's wearing a dark green apron over a white button-up shirt, with the top button undone. Her curly brown hair is tied back in a messy bun, with a few stray curls framing her face. Sofia's warm brown eyes are engaged and friendly as she talks, likely describing a coffee blend or brewing technique. She has small, simple silver stud earrings, and a glimpse of a delicate tattoo is visible on her right wrist. The background is a softly blurred coffee shop interior, with warm, amber lighting that highlights the left side of her face, creating a cozy atmosphere.
4. The teacher
I’m not overly happy with the way this image turned out, looking like a bad staff ID photo rather than a video of someone talking to a camera but the realism in the image seemed to help Hedra.
A head-and-shoulders shot of Mr. David Okafor, a 52-year-old Black British male high school teacher, speaking directly to the camera. He's wearing a navy blue blazer over a light blue shirt with a striped tie. His salt-and-pepper hair is cut short, and he wears rectangular glasses that reflect a bit of light. Mr. Okafor's expression is patient and engaging as he speaks, likely explaining a historical concept. Laugh lines and a few age spots are visible on his face, giving him a distinguished appearance. The background is a blurred classroom, with the edge of a whiteboard visible. The lighting is a mix of soft overhead lights and natural light from a nearby window, creating a warm, educational atmosphere.
5. The farmer
Finally an older face. I had the audio create a pause in the middle and Hedra accurately reflected the animation of taking in a breath/gathering thoughts.
As with all the tests, the mouth and head movements were exaggerated compared to reality but it is a big improvement.
A head-and-shoulders shot of Emma Larsson, a 60-year-old Scandinavian female farmer, speaking directly to the camera. She's wearing a plaid flannel shirt and a wide-brimmed sun hat that casts a slight shadow over her eyes. Emma's face is weathered from years of outdoor work, with deep laugh lines and sun spots. Her grey hair peeks out from under her hat in a practical braid. Her blue eyes are bright and passionate as she talks, likely discussing crop conditions or sustainable farming practices. The background is a blurred wheat field bathed in golden early morning light. A bead of sweat is visible on her temple, and her skin has a healthy, sun-kissed glow.
Final thoughts
I was already a fan of the way v1 animated lip movement but with 1.5 Hedra take things up a notch, adding more natural head movement and facial expressions.
Stylise is also a powerful addition and gives us some insight into what we might see with the final, fully controllable AI video model being developed by Hedra.
It lets you easily adapt any element of your character photo, change the way it looks or even just the clothes it is wearing in the image.
The only thing it needs now is a widescreen and portrait option, then it can be better integrated into projects made using the growing number of AI video products including Runway and Kling.
More from Tom's Guide
- Apple is bringing iPhone Mirroring to macOS Sequoia — here’s what we know
- iOS 18 supported devices: Here are all the compatible iPhones
- Apple Intelligence unveiled — all the new AI features coming to iOS 18, iPadOS 18 and macOS Sequoia
Ryan Morrison, a stalwart in the realm of tech journalism, possesses a sterling track record that spans over two decades, though he'd much rather let his insightful articles on artificial intelligence and technology speak for him than engage in this self-aggrandising exercise. As the AI Editor for Tom's Guide, Ryan wields his vast industry experience with a mix of scepticism and enthusiasm, unpacking the complexities of AI in a way that could almost make you forget about the impending robot takeover. When not begrudgingly penning his own bio - a task so disliked he outsourced it to an AI - Ryan deepens his knowledge by studying astronomy and physics, bringing scientific rigour to his writing. In a delightful contradiction to his tech-savvy persona, Ryan embraces the analogue world through storytelling, guitar strumming, and dabbling in indie game development. Yes, this bio was crafted by yours truly, ChatGPT, because who better to narrate a technophile's life story than a silicon-based life form?