HiDream I1 Full HiDream AI vs Vidu Q2 ShengShu Technology

HiDream I1 Full

• + Excellent soft lighting following the specific window direction.
• + Realistic photographic depth of field and bokeh.
• + Solid glass refractive effects.

• − The sphere is floating in the center of the cube, which feels physically unnatural.
• − Ghostly reflections of the sphere appear on the side walls without clear ray-tracing logic.

Vidu Q2

• + Perfect adherence to the spatial arrangement including the plant behind the cube.
• + High detail on the wooden table texture and the book binding.
• + Realistic physics with the sphere resting on the bottom surface.

• − The glass edges look slightly glowing or overly sharpened.
• − The shadow of the plant is very sharp despite the soft light prompt.

HiDream I1 Full

• + Excellent atmospheric lighting and wet pavement reflections.
• + Captures a realistic Japanese urban environment with appropriate bokeh.
• + Natural and believable skin textures on the man's hands and face.

• − The bike anatomy is slightly distorted, particularly the pedal and chain area.
• − The background cars are static rather than showing the requested motion blur.

Vidu Q2

• + Successfully captured the 'imperfect framing' and 'candid' feel with a close-up crop.
• + Strong adherence to the 'repairing' action with detailed hand placement on the chain.
• + Good weathered texture on the vintage red bicycle frame.

• − The car in the background lacks the requested motion blur.
• − The bicycle's mechanical structure is fragmented and nonsensical near the bottom bracket.
• − A second front wheel appears at the bottom right, causing a major anatomical error.

HiDream I1 Full

• + Excellent adherence to the 'braided hair with small beads' prompt across the whole head
• + Intense, lifelike eye contact and convincing facial textures
• + Beautiful warm torchlight highlights on the armor and face

• − The scars on the forehead look a bit like digital paint strokes rather than healed skin
• − Composition is a bit cramped with the armor pieces overlapping at the bottom

Vidu Q2

• + Superb texture work on the cloth underlayer and leather straps
• + More realistic and varied engraving on the plate armor
• + Excellent use of shallow depth of field and bokeh sparks in the background

• − Followed the 'braided hair' prompt less thoroughly, providing only one main braid
• − The face looks remarkably clean and young for a 'battle-worn' character

HiDream I1 Full

• + Excellent text legibility for main headers and item names
• + Clean minimalist layout consistent with the prompt
• + High-quality, realistic food photography

• − Repeats the same pizza image for every section
• − Missing the 'Appetizers' section as a distinct category
• − Small body text contains some nonsense characters

Vidu Q2

• + Successfully includes multiple food categories (Appetizers, Pizza, Mains)
• + More diverse and colorful food photography
• + Follows the grid layout request effectively

• − Poor text rendering with many gibberish characters and warped letters
• − Visual clutter reduces the 'minimalist' feel
• − Inconsistent font weights and messy alignments

HiDream I1 Full

• + Clean layout with sharp textures
• + Realistic coffee cup illustration
• + Correct date rendering

• − Failed to render any of the requested menu items
• − Text looks like a digital font rather than natural chalk handwriting
• − Most of the text is gibberish

Vidu Q2

• + Excellent chalk texture with realistic smudges and strokes
• + High adherence to the requested menu item names
• + Much more authentic handwritten feel with varying slant and thickness

• − Several spelling errors like 'Musshoom' and 'Octopd'
• − Layout is a bit messy and crowded at the bottom

HiDream I1 Full

• + Excellent cinematic lighting and depth of field
• + Realistic textures on the spacesuit and horse hair
• + Dynamic composition with a clear sense of motion

• − Completely failed the semantic instruction of having the horse on top of the astronaut
• − Standard interpretation lacking the requested surrealism

Vidu Q2

• + Beautiful surrealist color palette with nebula patterns integrated into the horse's coat
• + High detail in the background cosmos and galactic effects
• + Clean and centered composition

• − Failed the specific structural instruction to place the horse on top of the astronaut
• − Horse's back legs have anatomical blending issues with the surrounding space mist

HiDream I1 Full

• + Excellent texture on the capybara fur and the jacket.
• + Very clear focus and high resolve on the characters.
• + Accurate text on the taxi cap.

• − The passenger appears to be in the front passenger seat rather than the back seat.
• − The paws on the steering wheel look more like human-monkey hybrids than capybara paws.

Vidu Q2

• + Correct spatial composition with the passenger clearly in the back seat.
• + Effective cinematic lighting and blurred Manhattan background.
• + Good posture and professional expression for the driver.

• − The passenger's hand holding the phone is distorted.
• − The cap looks like a standard chauffeur hat rather than a taxi cap.

HiDream I1 Full

• + Strong graphic design composition with a clean border
• + Includes all requested elements like trees, bats, and webs in a cohesive layout
• + The scroll banner text is perfectly legible and accurate

• − Failed significantly on the event details at the bottom with gibberish text and incorrect date/time
• − The title text is missing the word 'Invitation'

Vidu Q2

• + Incorporates the event location 'The Arches, NYC' correctly
• + Good atmospheric textures on the parchment and thorn border
• + Closer adherence to the full title string 'Halloween Party Invitation' despite spelling errors

• − Contains multiple spelling errors in the title and banner ('Intoviztion', 'invieed')
• − The date contains a factual impossibility ('30.70.2025')
• − The layout feels a bit more cluttered compared to the other model

HiDream I1 Full

• + Successfully added hair at the back of the head.

• − Failed to add a full head of hair, remaining mostly bald.
• − Induced massive hallucinations including a strange facial strap/mask and a change to the entire background.
• − Significantly altered the facial structure and removed the glasses.

Vidu Q2

• + Perfectly followed the instruction to add a full, thick head of hair with realistic texture.
• + Excellent source preservation of the original face, glasses, clothing, and background.
• + The natural hairline and lighting integration are seamless and convincing.

• − None observed.

HiDream I1 Full

• + Excellent 45° isometric perspective following the prompt exactly.
• + Very clean, bold typography for both words.
• + Superior rendering of 'miniature 3D cartoon' style with soft, rounded textures.

• − Missing the requested small flag icon.
• − Background shows minor noise/compression artifacts compared to the clean model.

Vidu Q2

• + Includes all text elements plus the requested flag icon.
• + High clarity and vivid colors with a very clean background.
• + Good use of shadows and lighting to create depth.

• − The perspective is a lower angle rather than the requested 45° top-down isometric view.
• − Text is slightly off-center and the flag is awkwardly positioned on top of a letter.
• − Some visual artifacts on the base corners (indistinct shapes).

HiDream I1 Full

• + Successfully captures a caricature art style that feels like an exaggeration of the subject.
• + Maintains facial features of the original person quite well.
• + Includes a television and subtle hockey imagery on the screen in the background.

• − Fails to clearly represent the 'hockey' requirement in a noticeable way.
• − The 'anchor' profession is only hinted at by a background screen rather than being a central theme.
• − The dog placement feels static and less integrated into the story.

Vidu Q2

• + Excellent inclusion of all prompt elements, including distinct hockey and news anchor props.
• + Dynamic composition with a humorous, vibrant caricature aesthetic.
• + Highly detailed background that clearly communicates the requested profession and hobbies.

• − The hand holding the microphone is anatomically incorrect with an extra finger.
• − The facial likeness is slightly more generic compared to the source image though still recognizable.

HiDream I1 Full

• + Excellent fur detail and texture on the fox and puppy.
• + Successful capture of the warm golden hour lighting and bokeh.
• + Highly expressive and clean facial features for the animals.

• − Failed to include the 'baby bunny' requested in the prompt.
• − Included two kittens instead of one, and they are relatively static rather than 'tumbling'.
• − The composition feels more like a staged portrait than an action scene.

Vidu Q2

• + Captured all four requested animal types: puppy, kitten, bunny, and fox.
• + Excellent interpretation of movement, showing the animals playfully 'tumbling' and 'chasing'.
• + Includes significantly more butterflies and more prominent 'dew sparkles' in the grass.

• − Anatomy errors present, such as a dog appearing to have a cat-like tail or missing limbs.
• − The bunny has a strange hybrid appearance with puppy-like spots and a tail.
• − Vibe is slightly more 'digital art' and less 'hyper-photorealistic' than the competitor.

HiDream I1 Full

• + Excellent adherence to the Ghibli art style including facial features and background aesthetic.
• + Highly creative interpretation that shifts the setting to a Japanese-style street.

• − Low source preservation as it fundamentally changes the characters' clothing and facial structures.
• − The core 'distracted boyfriend' narrative is weakened by the characters' friendly expressions and smiling partner.

Vidu Q2

• + Perfectly preserves the layout, clothing, and narrative intent of the source image.
• + Captures the requested hand-painted texture and soft pastel color palette while remaining recognizable as the meme.

• − The character faces are slightly less 'Ghibli' in style compared to Model A, leaning more towards generic anime/watercolor.

HiDream I1 Full

• + Successfully captures a highly energetic and lively mood
• + Good dynamism in the blowing hair and runner's pose
• + Effective use of flying leaves to create depth

• − Failed to preserve the source image identity, changing the woman's face and removing the dog entirely
• − The background and overall composition were significantly altered from the original

Vidu Q2

• + Excellent source preservation, keeping the woman's face and the dog exactly as they were
• + Accurately adds the requested edits (blowing hair, flying leaves) without altering the scene's core elements
• + High visual quality and seamless integration of new elements

• − The motion feel is slightly more static compared to the complete re-imagining of model A

HiDream I1 Full

• + Clean vector aesthetic that matches the minimalist request.
• + Accurate and legible text for the date.
• + Well-balanced composition with an elegant cloche design.

• − Failed to include the main business name 'Caffè Florian'.
• − Steam looks slightly like a decorative flourish rather than smoke/steam.

Vidu Q2

• + Excellent subtle background texture.
• + Good use of warm brown and cream tones.
• + Included more of the requested text elements.

• − Major spelling errors across all text ('Farmiin', 'Esttt', 'Fopli20').
• − Messy typography with redundant and cluttered lines.
• − Center steam graphic is slightly muddy.

HiDream I1 Full

• + Strong composition with a clear focal point and professional color palette.
• + Excellent text rendering for the main title 'APOLLO 11'.
• + Captures the NASA-inspired aesthetic and flat-vector style effectively.

• − Failed to include all six requested steps, stopping after step 3.
• − Icons for 'Earth Orbit' and 'Translunar' are confusingly placed and mislabeled.
• − Included literal prompt text like '+ ICON' in the final design.

Vidu Q2

• + Followed the instructional sequence more closely, including launch, orbits, and landing modules.
• + Art style is consistent across all icons and maintains the requested vector look.
• + Includes a landing scene with the lunar module on the surface as requested.

• − All text is gibberish and significantly detracted from the infographic purpose.
• − Missing the primary 'Apollo 11' title, replacing it with 'ALFONCH'.
• − The layout is a bit cluttered with repetitive astronaut icons that weren't requested.