Identity as Presence: Towards Appearance and Voice Personalized Joint Audio-Video Generation

Abstract

Recent advances in video synthesis have enabled realistic integration of real individuals, driving demand for identity-aware generation. While emerging methods support joint appearance and voice injection in audio-visual models, they primarily focus on single-subject settings. Multimodal identity integration across multiple subjects remains limited, and precise alignment between visual and vocal identities in multi-subject scenarios remains underexplored. We present Identity-as-Presence, a unified framework for joint personalized audio-video generation. An automated data curation pipeline constructs identity-labeled audio-visual pairs for single- and multi-subject scenes. A unified identity injection mechanism then binds paired appearance and voice through shared cross-modal identity binding and subject-anchored captions. A multi-stage training strategy further leverages large-scale unimodal data alongside scarce paired clips to mitigate modality imbalance. Experiments show superior audio quality, video fidelity, and audio-visual consistency, with stronger multi-subject binding than the compared methods.

Method

The overall dual-tower DiT architecture and training framework. The model encodes video, audio, paired appearance/voice references, and subject-anchored captions, binds them with shared identity embeddings, and injects them via reference positioning and decoupled asymmetric attention. Training proceeds in three stages: unimodal identity, joint multimodal alignment, and spatio-temporal multi-view fine-tuning.

Single-subject Personalized Generation Results

Here we show the results of single-subject personalized generation. Please click on the videos to play the audio.

Multi-subject Personalized Generation Results

Here we show the results of multi-subject personalized generation. Please click on the videos to play the audio.

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in a formal government office with soft frontal lighting and a prominent American flag on the left wall. On the left, <REF_1_S>a middle-aged man with neatly trimmed gray-black short hair, wearing a tailored navy blue suit and blue-brown striped tie<REF_1_E>, sits upright with calm authority. He speaks and says, <S1>Leadership requires vision and discipline.<E1>. On the right, <REF_2_S>a woman with platinum blonde short curly hair, wearing bright red sleeveless top<REF_2_E>, leans forward slightly with a confident smile. She speaks and says, <S2>But it also needs courage to break the mold.<E2>. Background sounds include the quiet hum of an air conditioning system and the occasional soft click of a mechanical clock.

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in a nostalgic film set with wooden props and soft golden light. On the left, <REF_1_S>a man with long silver hair and a gray mustache, wearing a thick brown tweed coat<REF_1_E>, writes on a chalkboard. He speaks and says, <S1>I don't need a formula to make magic.<E1>. On the right, <REF_2_S>a woman with platinum blonde short curly hair, wearing bright red sleeveless top<REF_2_E>, winks at the camera. She speaks and says, <S2>But even magic follows certain rules.<E2>. Background sounds include the scratch of chalk on slate and the distant hum of a vintage projector warming up.

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in a slightly dimly lit but lively event site or backstage, with a blurred background. On the left side of the screen is <REF_1_S>a young male with thin metal-rimmed glasses and a gentle, scholarly temperament, with short parted black hair<REF_1_E>, wearing a red round-neck knit shirt; on the right side is <REF_2_S>a young male with short black bangs and delicate features<REF_2_E>, wearing a white hooded sweatshirt over a gray printed T-shirt. He intimately rests his arm on the shoulder of the male on the left. They lean their heads together, showing relaxed and sincere smiles to the camera. The male on the left says gently: <S1>The atmosphere tonight is really good.<E1> The male on the right responds mischievously: <S2>With you here, of course it's lively.<E2>, showing the deep friendship between the two.

Prompt: This is a glamorous red carpet event scene, captured with a medium close-up fixed camera. The background features a blurred event backdrop and crowd, with bright, even lighting that exudes the atmosphere of a high-profile social scene. On the left is a young woman with side-parted brown wavy short hair and delicated makeup, wearing a dress with pink 3D floral shoulder straps, her eyes bright and a sweet smile on her face. On the right is a young man with voluminous brown curly hair, wearing a black suit with a black bow tie and a crisp white shirt underneath, his smile confident and radiant. They stand close together facing the media cameras, presenting their best public image. The woman whispers slightly to the side: <S1>The flashlights seem even more intense tonight than usual.<E1> The man keeps smiling and calmly responds: <S2>Don't worry, we just need to keep smiling.<E2>

---

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Prompt: This is a noble and elegant red carpet gala film scene, shot with a medium close-up fixed lens. The lighting is soft and bright, with a blurred background of crowds and camera equipment, creating a highly publicized, star-studded atmosphere. On the left is an elegant woman with a brown low ponytail, a few loose strands gently framing her face, wearing vintage pearl earrings, her eyes deep and captivating, a subtle smile on her lips. On the right is a man with neat side-parted short hair, a handsome and composed face, wearing a black suit with a black bow tie, his gaze firm and forward. They stand shoulder to shoulder, exuding confidence and poise as they face the cameras, showcasing star power. The woman softly says: <S1>This red carpet feels longer than I imagined.<E1> The man slightly turns his head and gently replies: <S2>It's okay, as long as you're by my side.<E2>, revealing deep affection between them.

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in a reflective evening office with city lights glowing outside and soft ambient desk lighting. On the left, <REF_1_S>a young woman with long wavy blonde hair and exquisite, atmospheric makeup<REF_1_E>, wearing a champagne satin off-shoulder gown, leans forward with a warm smile. She speaks and says, <S1>This moment feels like the beginning of something real.<E1>. On the right, <REF_2_S>a middle-aged male with short dark brown hair and a slightly weathered face<REF_2_E>, wearing a navy blue velvet blazer with a gray silk tie, nods slowly. He speaks and says, <S2>And the best beginnings are the ones we do not plan.<E2>

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in a moody fashion week afterparty, with colored lights pulsing and deep house music vibrating in the air. On the left, <REF_1_S>a young male with brown slightly curly medium-length hair<REF_1_E>, wearing a sleek black bomber jacket over a white shirt and a silver chain, leans in slightly. He speaks and says, <S1>I think we just became the best part of this night.<E1>. On the right, <REF_2_S>a young male with reddish-brown curly short hair<REF_2_E>, wearing a soft white knit shirt and layered pearls, laughs quietly. He speaks and says, <S2>Then let's make it last a little longer,<E2>

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set at a fan meet-and-greet after the event, with colorful banners and excited voices in the background. On the left, <REF_1_S>a young male with brown curly hair and a sunny and brilliant smile<REF_1_E>, wearing a black T-shirt with a logo, waves to the crowd. He speaks and says, <S1>They really love you.<E1>. On the right, <REF_2_S>a young female with dark brown center-parted long straight hair<REF_2_E>, wearing a pink denim jacket and hoop earrings, smiles warmly. She speaks and says, <S2>And they love you even more,<E2>. Background sounds include the joyful squeals of fans and the rhythmic clap of hands in the background.

---

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Prompt: This is a tense subway tunnel scene, shot with a tracking shot from a side rail-mounted camera. The lighting is flickering and unstable, with distant train lights sweeping across their faces. The background is a damp concrete tunnel with graffiti-covered walls. On the left side of the frame is <REF_1_S>a sharp, professional woman with neat black hair tied in a bun and large silver hoop earrings<REF_1_E>, wearing a dark raincoat and gloves, her expression focused and stern. On the right is <REF_2_S>a mature man with dark short, tousled hair and light stubble<REF_2_E>, dressed in a black overcoat and scarf, his gaze deep and piercing. They move cautiously down the tracks, flashlights in hand. The woman stops and whispers: <S1>Footprints end here.<E1> The man shines his light ahead and replies: <S2>Then he's still close.<E2>

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in a minimalist creative studio with soft cold-toned natural light from the side. On the left, <REF_1_S>a young man with short black messy hair, wearing a black branded hoodie<REF_1_E>, rests his chin on his hand. He speaks and says, <S1>等你好久。<E1>. On the right, <REF_2_S>a woman with brown long wavy hair in a low ponytail, wearing a pure white V-neck silk top<REF_2_E>, looks up with a gentle smile. She speaks and says, <S2>现在我们终于有时间好好聊聊了。<E2>

Prompt: A reflective pause on a quiet balcony, shot with a wide-angle close-up. The background is a city skyline under soft moonlight. On the left is <REF_1_S>a young woman with side-parted brown wavy short hair and exquisite makeup<REF_1_E>, wearing a dress with pink 3D floral shoulder straps, arms crossed. On the right is <REF_2_S>a man with neat side-parted short hair, a handsome and composed face<REF_2_E>, wearing a charcoal-gray double-breasted suit with a crisp white shirt and no tie, sleeves rolled slightly. The woman on the left says: <S1>Some moments feel like they were meant to be remembered.<E1> The man on the right replies: <S2>Especially the ones we don't rush through.<E2>

Prompt: A vibrant night market encounter, lit by colorful lanterns and warm stall lights. On the left, <REF_1_S>a young Asian man in a black leather jacket<REF_1_E> holds a snack and says: <S1>热闹的地方，心反而静。<E1> On the right, <REF_2_S>a woman with golden curls in a red sleeveless top<REF_2_E> smiles and replies: <S2>因为知道，有人在等你回家。<E2>

---

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Audio Ref 1 Audio Ref 2

Ref 1 Ref 2 Ref 3

Prompt: This is a dialogue presented in a static, eye-level medium shot from a front view. The scene is set in a dimly lit study room with bookshelves and a desk lamp. On the right, <REF_1_S>a young woman with black hair, wearing glasses and a white button-up shirt<REF_1_E>, sips tea. She says, <S1>Write down every idea, even the weak ones.<E1>. On the left, <REF_2_S>a young man with short black hair, wearing a dark jacket<REF_2_E>, flips through a notebook, saying, <S2>Because the weak ones sometimes lead to strong ones.<E2>

Prompt: This is a classic interview or group photo close-up, shot with a medium-close fixed lens at eye level. The scene is set against a minimalist indoor background, with beige textured wallpaper appearing simple and nostalgic, and soft frontal lighting clearly illuminating the faces. On the left side of the screen, <REF_1_S>a male with a very short buzz cut and a thin, cool face<REF_1_E> is wearing a dark gray turtleneck sweater; on the right side, <REF_2_S>a female with short black hair<REF_2_E> is wearing a black top with gold embroidery on the collar, looking gentle. They stand side by side, looking natural. The male looks straight ahead with deep eyes, his expression serious and reserved, while the female tilts her head slightly with a faint smile on her lips. The male says calmly: <S1>The story has been told.<E1> The female responds gently: <S2>But the touching part has just begun.<E2>

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in a hybrid space blending old books and digital screens. On the left, <REF_1_S>a man with long silver hair and a gray mustache, wearing a thick brown tweed coat<REF_1_E>, holds a fountain pen. He speaks and says, <S1>Wisdom comes from reflection.<E1>. On the right, <REF_2_S>a young Asian man with neat black short hair, wearing a glossy black leather jacket<REF_2_E>, turns on a tablet. He speaks and says, <S2>And progress comes from action.<E2>

Prompt: This is a short three-person conversation in a rainy-day cafe by a glass window, shot with a fixed eye-level frontal medium framing so all three faces face the camera clearly. The background shows a wooden table, rain streaks on the window, and soft indoor lighting. First, on the left, <REF_1_S>a middle-aged white man with short curly salt-and-pepper hair and a clean-shaven face, wearing a deep teal shirt jacket over a white henley<REF_1_E>, holds a cup with both hands and calmly says: <S1>Take your time.<E1>. While he speaks, the other two listen quietly with closed lips. Then, in the center, <REF_2_S>an older white man with short gray hair, deep forehead wrinkles, and thick dark brows, wearing a plum knit sweater over a collared shirt<REF_2_E>, looks steady and assured as he says: <S2>Patience pays off.<E2>. While he speaks, the others listen quietly with closed lips. Finally, on the right, <REF_3_S>a young white woman with long blonde hair and bangs, blue eyes, red lips, and defined eyeliner, wearing a cream cardigan over a mist-blue top<REF_3_E>, softly replies: <S3>Agreed.<E3>. The three speak in turn with brief, natural, coherent dialogue.

Cross-domain Personalized Generation Results

Here we show the results of cross-domain personalized generation. Please click on the videos to play the audio.

Text-reference Disentanglement Results

Here we show the results of text-reference disentanglement. Please click on the videos to play the audio.

Audio Reference

Init Generation

Edit: Init generation — fixed reference; expression, scene, and outfit follow the baseline prompt.
Prompt: This is an interview-style indoor talking-head clip with a stable eye-level frontal medium close-up framing the upper body. The scene is set in an indoor space with a white doorframe and glass windows under soft frontal light. In the frame, <REF_1_S>a middle-aged man with short grayish-white hair, facial wrinkles, and no beard<REF_1_E>, wearing a dark corduroy suit jacket over a blue-and-white checkered shirt, turns his body slightly toward the camera with his head facing forward. As he speaks, he raises his right hand to gesture near his chest, then lowers it naturally. He wears a mild smile with focused eyes, and says: <S>Consistency is the only way to turn good intentions into real results over time.<E>

Change Expression

Edit: Change expression only — angry face (lowered brows, flared nostrils, downturned mouth), while identity, scene, and outfit stay the same.
Prompt: This is an interview-style indoor talking-head clip with a stable eye-level frontal medium close-up framing the upper body. The scene is set in an indoor space with a white doorframe and glass windows under soft frontal light. In the frame, <REF_1_S>a middle-aged man with short grayish-white hair, facial wrinkles, and no beard<REF_1_E>, wearing a dark corduroy suit jacket over a blue-and-white checkered shirt, turns his body slightly toward the camera with his head facing forward. As he speaks, he raises his right hand to gesture near his chest, then lowers it naturally. He looks clearly angry, with lowered brows, flared nostrils, and a downturned mouth, staring ahead, and says: <S>Consistency is the only way to turn good intentions into real results over time.<E>

Change Scene

Edit: Change scene only — home study with bookshelves and warm lamps, while identity, expression, and outfit stay the same.
Prompt: This is an interview-style indoor talking-head clip with a stable eye-level frontal medium close-up framing the upper body. The scene is set in a home study with bookshelves and warm lamps under soft frontal light. In the frame, <REF_1_S>a middle-aged man with short grayish-white hair, facial wrinkles, and no beard<REF_1_E>, wearing a dark corduroy suit jacket over a blue-and-white checkered shirt, turns his body slightly toward the camera with his head facing forward. As he speaks, he raises his right hand to gesture near his chest, then lowers it naturally. He wears a mild smile with focused eyes, and says: <S>Consistency is the only way to turn good intentions into real results over time.<E>

Change Outfit

Edit: Change outfit only — light-gray casual suit over a light-blue shirt, while identity, expression, and scene stay the same.
Prompt: This is an interview-style indoor talking-head clip with a stable eye-level frontal medium close-up framing the upper body. The scene is set in an indoor space with a white doorframe and glass windows under soft frontal light. In the frame, <REF_1_S>a middle-aged man with short grayish-white hair, facial wrinkles, and no beard<REF_1_E>, wearing a light-gray casual suit over a light-blue shirt, turns his body slightly toward the camera with his head facing forward. As he speaks, he raises his right hand to gesture near his chest, then lowers it naturally. He wears a mild smile with focused eyes, and says: <S>Consistency is the only way to turn good intentions into real results over time.<E>

Spatiotemporal Multi-view Consistency Results

Here we show the results of spatiotemporal multi-view consistency. Please click on the videos to play the audio.

Spatial Packing

Prompt: A fixed medium close-up captures <REF_1_S>a young man with black curly hair, a nose ring, and dark skin<REF_1_E> standing behind a daytime glass-greenhouse planting table in casual gardening clothes. Hands on his hips, he leans slightly forward as he speaks. He first turns slowly from a frontal face to the left for a clear profile, then turns slowly to the right, and finally returns to a frontal face, saying: <S>Look at these thriving plants.<E> Soft natural light filters through the glass onto the leaves, with a slightly blurred background highlighting the subject.

Prompt: A fixed medium close-up captures <REF_1_S>a young man with black curly hair, a nose ring, and dark skin<REF_1_E> standing behind a daytime glass-greenhouse planting table in casual gardening clothes. Hands on his hips, he leans slightly forward as he speaks. He first turns slowly from a frontal face to the left for a clear profile, then turns slowly to the right, and finally returns to a frontal face, saying: <S>Look at these thriving plants.<E> Soft natural light filters through the glass onto the leaves, with a slightly blurred background highlighting the subject.

Prompt: A fixed medium close-up shows <REF_1_S>a young man with black curly hair, thin-framed glasses, and a dense full beard<REF_1_E> in a plain shirt beside a wooden kitchen counter. He raises one hand to scratch his hair while the other hangs naturally. He first turns slowly from a frontal face to the left for a clear profile, then turns slowly to the right, and finally returns to a frontal face, saying: <S>The coffee is ready now.<E> Soft everyday kitchen light falls across the scene, with blurred utensil outlines in the background.

Prompt: A fixed medium close-up shows <REF_1_S>a young man with black curly hair, thin-framed glasses, and a dense full beard<REF_1_E> in a plain shirt beside a wooden kitchen counter. He raises one hand to scratch his hair while the other hangs naturally. He first turns slowly from a frontal face to the left for a clear profile, then turns slowly to the right, and finally returns to a frontal face, saying: <S>The coffee is ready now.<E> Soft everyday kitchen light falls across the scene, with blurred utensil outlines in the background.

Temporal Expansion

Prompt: This is a fixed medium shot in a home study. <REF_1_S>A young man with short red hair and a beard<REF_1_E> sits facing the camera behind a desk piled with papers, wearing a dark-brown knit sweater. His brows are furrowed, blue eyes glare forward in strong anger, and he slams a pen onto the desk while angrily saying: <S>这种低级错误绝对不能容忍！<E> Warm yellow desk light hits his tense face, with blurred bookshelf books behind him.

Prompt: This is a fixed medium shot in a home study. <REF_1_S>A young man with short red hair and a beard<REF_1_E> sits facing the camera behind a desk piled with papers, wearing a dark-brown knit sweater. His brows are furrowed, blue eyes glare forward in strong anger, and he slams a pen onto the desk while angrily saying: <S>这种低级错误绝对不能容忍！<E> Warm yellow desk light hits his tense face, with blurred bookshelf books behind him.

Prompt: This is an outdoor medium close-up in soft overcast daylight, with a blurred gray stone wall behind the subject. In the frame, <REF_1_S>a young woman with long platinum-blonde wavy hair, a small side braid, fair skin, dark brows, and light blue eyes<REF_1_E>, wearing a deep navy-blue velvet high-neck garment with silver embroidered detailing on the shoulders. Strands of her hair blow lightly in the breeze as she starts from a calm soft smile, then breaks into a bright open-mouthed laugh with squinted eyes, saying: <S>Haha, this feels amazing!<E>

Prompt: This is an outdoor medium close-up in soft overcast daylight, with a blurred gray stone wall behind the subject. In the frame, <REF_1_S>a young woman with long platinum-blonde wavy hair, a small side braid, fair skin, dark brows, and light blue eyes<REF_1_E>, wearing a deep navy-blue velvet high-neck garment with silver embroidered detailing on the shoulders. Strands of her hair blow lightly in the breeze as she starts from a calm soft smile, then breaks into a bright open-mouthed laugh with squinted eyes, saying: <S>Haha, this feels amazing!<E>

Qualitative Comparison Results

Here we show qualitative comparisons with other methods. Please click on the videos to play the audio.

Single-subject Comparison

Audio Reference

Ours (OVI)

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

Ours (LTX)

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

DreamID-Omni

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

HuMo

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

HunyuanCustom

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

ID-LoRA

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

OmniCustom

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

Phantom

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

UniAVGen

Prompt: This is an interview or vlog-style video, presented as a static, eye-level medium shot from a front view. The scene is set inside a horse trailer or stable, with metal bars and dark walls in the background, under soft, even lighting. <REF_1_S>A young man with short blonde hair<REF_1_E>, wearing a dark blue polo shirt with black stripes, stands on the left side of the frame. He is petting a brown horse with a white marking on its forehead, which stands calmly on the right. The man speaks directly to the camera, gesturing with his hands as he talks, and says: <S>because I was kind of like not really sure how to feel and how I was going to feel. There was going to be a lot.<E>

Multi-subject Comparison

Audio Ref 1 Audio Ref 2

Ours (OVI)

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in front of a green chalkboard with a rustic wooden frame under soft warm lighting. On the left, <REF_1_S>an elderly man with wild silver-white hair and a thick gray mustache<REF_1_E>, wearing a brown textured blazer over a dark shirt and black tie, raises his hand and playfully points toward the woman beside him. He speaks and says: <S1>I don't need a formula to make magic.<E1> On the right, <REF_2_S>a woman with platinum-blonde short curly hair, blue eyes, and a beauty mark on her cheek<REF_2_E>, wearing a bright red sleeveless turtleneck, smiles brightly at the camera. She speaks and says: <S2>But even magic follows certain rules.<E2>

Ours (LTX)

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in front of a green chalkboard with a rustic wooden frame under soft warm lighting. On the left, <REF_1_S>an elderly man with wild silver-white hair and a thick gray mustache<REF_1_E>, wearing a brown textured blazer over a dark shirt and black tie, raises his hand and playfully points toward the woman beside him. He speaks and says: <S1>I don't need a formula to make magic.<E1> On the right, <REF_2_S>a woman with platinum-blonde short curly hair, blue eyes, and a beauty mark on her cheek<REF_2_E>, wearing a bright red sleeveless turtleneck, smiles brightly at the camera. She speaks and says: <S2>But even magic follows certain rules.<E2>

DreamID-Omni

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in front of a green chalkboard with a rustic wooden frame under soft warm lighting. On the left, <REF_1_S>an elderly man with wild silver-white hair and a thick gray mustache<REF_1_E>, wearing a brown textured blazer over a dark shirt and black tie, raises his hand and playfully points toward the woman beside him. He speaks and says: <S1>I don't need a formula to make magic.<E1> On the right, <REF_2_S>a woman with platinum-blonde short curly hair, blue eyes, and a beauty mark on her cheek<REF_2_E>, wearing a bright red sleeveless turtleneck, smiles brightly at the camera. She speaks and says: <S2>But even magic follows certain rules.<E2>

Phantom

Prompt: This is a conversation presented in a static, eye-level medium shot from a front view. The scene is set in front of a green chalkboard with a rustic wooden frame under soft warm lighting. On the left, <REF_1_S>an elderly man with wild silver-white hair and a thick gray mustache<REF_1_E>, wearing a brown textured blazer over a dark shirt and black tie, raises his hand and playfully points toward the woman beside him. He speaks and says: <S1>I don't need a formula to make magic.<E1> On the right, <REF_2_S>a woman with platinum-blonde short curly hair, blue eyes, and a beauty mark on her cheek<REF_2_E>, wearing a bright red sleeveless turtleneck, smiles brightly at the camera. She speaks and says: <S2>But even magic follows certain rules.<E2>