Herramientas para generar imágenes con IA | NosoloHD

Herramientas para generar imágenes con IA

DeBilbao

...del mismo Bilbao
Mensajes
13.591
Reacciones
8.197
Ubicación
El Botxo
En el hilo de Imágenes generadas por IA de foreros llevamos tiempo compartiendo resultados de imágenes generadas con distintos tipos de IAs.

Yo he encontrado mi fórmula en la combinación de dos y a veces hasta tres tipos de IAs para conseguir el resultado final, pero para la generación de la imagen final el motor que más me gusta es el de ImageFX que me recomendó @Miguel_Angel a finales de marzo y que llevo algo más de un mes usándolo con intensidad.

Dejo este post para que nos sirva de índice.

De momento voy recopilando posts que he ido publicando en el hilo de las imágenes, para que tengáis una guía de cómo conseguir crear imágenes impactantes y las podáis compartir allí.
 
Última edición:
ImageFX solo se puede usar desde USA, así que para poder usarlo aquí hay que tener activada una VPN que nos conecte con USA.

Para ello, la fórmula que más me gusta es usar un navegador con VPN integrada, y así no tener el equipo completo conectado a una red remota, sino solo ese navegador.

Yo uso el navegador Opera, que tiene varias versiones, y para esta tarea uso la versión Air, que es "mindful" :mparto

Se descarga desde aquí:


ojjwRFh.png


En la configuración (arriba a la derecha) tengo configurado para que cuando abro el navegador se conecte directamente a la VPN, y en el panel de VPN (arriba a la izquierda) tengo desactivado que se conecte a la ubicación más rápida y que lo haga a América.

WaaEEBD.png



Awj1dAP.png


De esa forma, siempre que lo abro se conecta a un lugar de América donde puedo hacer funcionar a Google ImageFX.

MuJoiBQ.png


En otra pestaña tengo mis conversaciones con DeepSeek, para que me ayude a crear los prompts.

OUM7ZcU.png


Y como extra, en una pestaña tengo Pandora, que es un servicio de streaming cojonudo y gratuito que tiene el canal Jazz Essentials Radio que es una música de fondo estupenda.

2CboQVe.png
 
Última edición:
Os he contado que estoy experimentando con ImageFX, una maravillosa herramienta salida de los Google Labs.

Pero el mérito real que hay detrás de la calidad de imágenes, está el modelo Imagen 3 del proyecto Google DeepMind, que se puede usar desde ImageFX o desde Gemini.


Nxqpd2f.png


Podéis probar alguno de los prompts que vienen de ejemplo en esa web, y os servirá para inspiraros y ver que los límites los marcas tú, aunque ya hemos visto que hay cosas que se le dan mejor, y cosas que se le dan peor...

A close-up, macro photography stock photo of a strawberry intricately sculpted into the shape of a hummingbird in mid-flight, its wings a blur as it sips nectar from a vibrant, tubular flower. The backdrop features a lush, colorful garden with a soft, bokeh effect, creating a dreamlike atmosphere. The image is exceptionally detailed and captured with a shallow depth of field, ensuring a razor-sharp focus on the strawberry-hummingbird and gentle fading of the background. The high resolution, professional photographers style, and soft lighting illuminate the scene in a very detailed manner, professional color grading amplifies the vibrant colors and creates an image with exceptional clarity. The depth of field makes the hummingbird and flower stand out starkly against the bokeh background.

vl0W0vN.png


Shot in the style of DSL camera with the polarizing filter. A photo of two hot air balloons floating over the unique rock formations in Cappadocia, Turkey. The colors and patterns on these balloons contrast beautifully against the earthy tones of the landscape below. This shot captures the sense of adventure that comes with enjoying such an experience.

mQPWsmL.png



bmxJodX.png



A7ez4LO.png
 
Queda claro que es importante saber hacerlo.

Pues compartamos cómo lo hacemos, que cada uno tenemos nuestra receta.

Os comparto la mía.

Lo primero que hago es buscar un tema sobre el que quiero trabajar. Imaginemos que quiero hacer algo sobre actrices de los 80s.

Y en lugar de pensar, le pregunto a una IA cualquiera (ChatGPT, Gemini, Claude, Deepseek, etc.) que me de una lista de actrices relevantes de los 80s.

El prompt puede ser así de simple, o lo podemos hacer más complejo si queremos más contexto. Y aunque la mayoría de las IAs entienden bien el español, entienden mejor el inglés.

Best female actress of the 80s

Puedes estar de acuerdo o no con el gusto de la IA, pero en este caso Gemini de Google dice esto:

The 1980s were a truly remarkable decade for female actresses, with a wealth of talent gracing the silver screen. Pinpointing the "best" is subjective, but some names consistently rise to the top. Here are a few of the most influential and acclaimed:

  • Meryl Streep:
    • Her performances in films like "Sophie's Choice" (1982) solidified her status as a cinematic powerhouse. Her range and depth were unparalleled.
  • Glenn Close:
    • Known for her intense and captivating performances, especially in "Fatal Attraction" (1987), she left an indelible mark on the decade.
  • Sigourney Weaver:
    • Her portrayal of Ellen Ripley in the "Alien" franchise redefined action heroines, showcasing strength and resilience.
  • Jessica Lange:
    • She delivered powerful performances in movies such as "Tootsie"(1982) and "Frances"(1982), displaying great range of acting ability.
  • Kathleen Turner:
    • She was a very strong presence in many 80's movies, such as "Romancing the Stone"(1984) and "Body Heat"(1981).
These women, among many others, shaped the cinematic landscape of the 1980s with their talent, versatility, and groundbreaking performances.

Quiero generar una foto de una de ellas, pero ¿qué estilo seguir? ¿quién las hubiera fotografiado bien? Pues se lo pregunto a Deepseek

Which photographer would have made great portfolio photographies of those actresses?

Several renowned photographers are celebrated for their ability to capture the essence of beauty, diversity, and personality in their subjects. Here are a few photographers who would have excelled at creating stunning portfolio photographs of actresses with mixed ethnic origins, highlighting their unique beauty and cultural heritage:

  1. Annie Leibovitz
    • Known for her iconic portraits of celebrities, Leibovitz has a talent for capturing the depth and character of her subjects. Her use of lighting, composition, and storytelling would beautifully highlight the actresses' mixed heritage and individuality.
  2. Mario Testino
    • Testino's glamorous and vibrant style would bring out the elegance and confidence of the actresses. His ability to blend high fashion with cultural elements would make for striking and memorable portraits.
  3. Platon
    • Famous for his powerful black-and-white portraits, Platon's minimalist approach would focus on the actresses' facial features and expressions, emphasizing the beauty of their mixed heritage in a timeless way.
  4. Richard Avedon
    • Avedon's classic and dramatic portraits would capture the actresses' unique beauty with a sense of sophistication and artistry. His ability to convey emotion through simplicity would make the images unforgettable.
  5. Tim Walker
    • Known for his whimsical and fantastical style, Walker could create dreamlike portraits that celebrate the actresses' diverse backgrounds through imaginative settings and cultural symbolism.
  6. Steve McCurry
    • McCurry's expertise in capturing cultural diversity and human emotion would result in portraits that tell a story about the actresses' heritage and identity, often with rich, vibrant colors.
  7. Peter Lindbergh
    • Lindbergh's raw and natural approach to photography would highlight the actresses' authentic beauty, focusing on their unique features and personalities without excessive retouching.
  8. Nigel Barker
    • A fashion photographer with a keen eye for diversity, Barker would create bold and empowering portraits that celebrate the actresses' mixed heritage and confidence.
  9. Dana Scruggs
    • Scruggs is known for her striking portraits that highlight diversity and individuality. Her work often features bold lighting and composition, making her an excellent choice for showcasing the actresses' unique beauty.
  10. Camilla Åkrans
    • Åkrans' ethereal and artistic style would create soft, dreamy portraits that emphasize the actresses' elegance and the blending of their cultural backgrounds.
Each of these photographers brings a unique perspective and style, making them ideal candidates to create a portfolio that celebrates the beauty of mixed ethnic origins. Depending on the desired tone and aesthetic, you could choose a photographer whose style aligns with the vision for the project.

No está nada mal. Ya tengo una lista de los fotógrafos y sus estilos.

Ahora solo me queda combinarlo y crear un prompt para ImageFX.

Porque no es lo mismo ver a la joven teniente Ellen Ripley a través de la cámara de Mario Testino

Create a vibrant, cinematic portrait of Sigourney Weaver as Ellen Ripley, in the style of Mario Testino.

29P0x6i.png


U0XXMcm.png



s0pQaN0.png



NGPW18n.png


Que a través de la cámara de Tim Walker

zjY0mMJ.png



Syqr431.png



Si quiero imaginar escenas concretas más complejas, puedo pensar el prompt, o pedirle a la IA que me cree un prompt para una herramienta de generación de imágenes en el que muestre a una de las actrices en el estilo de uno de los fotógrafos y haciendo algo concreto.

Describe a cinematic photograph in Mario Testino style of Sigourney Wwaver as Ellen Ripley in Alien. She's climbing a ladder with a gunmachine, trying to escape from the Alien that approaches from behind. Write it as a prompt for a image generation AI Tool

Deepseek me sugiere este prompt

Create a high-energy, cinematic photograph of Sigourney Weaver as Ellen Ripley from Alien, in the style of Mario Testino. She is climbing a sleek, metallic ladder inside a futuristic spaceship, her body taut with determination and urgency. She wears the iconic Nostromo crew outfit—a fitted tank top and cargo pants—with a machine gun strapped to her back. Her face is illuminated by a cool, blue light, highlighting her fierce expression and sweat-streaked skin. Behind her, the shadowy, biomechanical form of the Alien creeps closer, its menacing silhouette blurred slightly to emphasize motion and danger. The background is a mix of industrial steel and neon lighting, creating a stark, high-contrast environment. The composition is dynamic, with Ripley’s powerful stance and the Alien’s looming threat creating a sense of tension and drama. The image balances the gritty, sci-fi atmosphere of Alien with Testino’s signature glamour, capturing Ripley’s strength, resilience, and iconic heroism in a visually stunning way.

1PZGDol.jpeg


Y ya está.

Inspiration coming to reality in nearly-zero time...

gjWn50y.png



Y después del descubrimiento de Hailuo, puedes conseguir dar vida a esa escena...



Si os apetce, podéis poner el método en práctica y compartir vuestros resultados. Seguro que aprendemos entre todos.

:ok :ok :ok
 
Es lo que he dicho, y ese primor que solo afecta a la parte que dices, producirá igualmente resultados distintos cada vez que lo uses sobre la misma base de imagen.

En el corazón del sistema de la tecnología actual sobre la que todo esto se sustenta está la aleatoriedad. Pretender que no sea así es imposible, se necesita algo diferente desde cero.


Es una tecnología imperfecta, sí, pero a la velocidad que avanza tengo claro que se irá perfeccionando.

Así es como funciona el sistema, y en mi opinión es así como debe funcionar en su parte "generativa".

Se trata de que invente contenido que no existe en base a una información que debe ser lo más precisa posible para que obtenga los resultados deseados.

Editar ese resultado es función de una herramienta de edición de imágenes tradicionales, que deberán incorporar esta tecnología para que las correcciones y modificaciones que queramos realizar puedan ser creadas con un simple texto que indique lo que queremos hacer, como una herramienta más en la caja de herramientas.

De momento, el secreto está en el prompt, que debe ser lo más completo posible para llevar la creación por donde queremos. El LLM que utilizan como base cada una de las herramientas ha sido entrenado con miles de millones de variables, pero son una ínfima parte sobre todo el conocimiento que ha acumulado la humanidad, y no han sido entrenados con la información específica de tu entorno, de los sistemas de las empresas u organizaciones.

Ahora mismo se está trabajando en los sistemas RAG (Retrieval Augmented Generation) que permiten complementar el conocimiento general del LLM con el conocimiento específico que quieras, conectándolo con los sistemas de información de las compañías para poder obtener resultados precisos, aunque es todo un reto conseguir que las respuestas más sencillas sean siempre las mismas.

El método que utilizo para crear las series de imágenes que os comparto utiliza el prompt para acotar el resultado, y además de decirle que quiero algo, se lo completo para darle el contexto que quiero y que las variaciones tengan sentido y sean coherentes.

Tomando como ejemplo el combate de Ali vs Lee que nos proponía ayer @Hinomura Krycek, este sería mi prompt básico para generar un prompt específico:

Generate a prompt for a AI image generation tool, to create a signature photograph in Neil Leifer style, showing an action moment during a boxing combat between Muhammad Ali, wearing a white short with black strip, and Bruce Lee, wearing a golden short with red stripe. Mohammed is defending, covering his face with the gloves, while Bruce is punching him in the liver fiercely.

Y este sería el prompt que utilizaría:

Create a signature photograph in the iconic and perfectly timed style of Neil Leifer, capturing a high-intensity moment during a fictional boxing match between Muhammad Ali and Bruce Lee. The scene is set in a packed, electrified arena, with the crowd roaring in the background and the ring illuminated by bright, dramatic lighting.

Muhammad Ali, wearing white shorts with a black stripe, is in a defensive stance, his gloves raised to cover his face, his body slightly turned to absorb the impact. His expression is focused and strategic, showcasing his legendary defensive skills. Bruce Lee, wearing golden shorts with a red stripe, is mid-action, delivering a fierce punch to Ali's liver. His body is coiled with power, his face intense and determined, capturing his explosive speed and precision.

The composition should emphasize the dynamic energy of the fight, with sharp focus on the fighters and a slightly blurred background to highlight their movement. Sweat flies off their bodies, adding to the intensity of the moment. The lighting should be dramatic, casting strong shadows and highlighting the textures of their skin, gloves, and the ropes of the ring. The overall mood is one of power, speed, and the timeless allure of two legendary athletes in their prime, embodying the essence of Neil Leifer's ability to capture iconic moments in sports history.

¿Me ha hecho caso? Pues en el agunas cosas sí, y en otras no.

El estilo de la fotografía me cuadra y me gusta, Mohammed lleva el calzón blanco con cintura negra como le he pedido, Bruce lleva correctamente el calzón dorado, pero no la banda de la cintura no es roja como quería para representar la combinación dorado y rojo típica de china.

Y de la pose defensiva de Mohammed mientras le zurra Bruce, ni hablamos.

8oz4WGL.jpeg



Así que me conformo con este par de momentos, en donde cada uno coloca su golpe.

gPtAteq.png


0Of6xV2.png



Si quiero seguir con otra foto de la misma serie, en otro momento del combate, tengo que crear un prompt que mantenga el mismo escenario, y eso le pido a la IA:

Create another prompt, in the same scenario, after finishing the combat, with Mohammed Ali and Bruce Lee embracing each other in a gesture of fraternity

Que me ofrece este resultado:

Create a signature photograph in the iconic and emotive style of Neil Leifer, capturing a powerful moment of fraternity and respect after a fictional boxing match between Muhammad Ali and Bruce Lee. The scene is set in a packed, electrified arena, with the crowd on their feet, cheering and applauding in the background. The ring is illuminated by soft, warm light, creating a dramatic yet intimate atmosphere.

Muhammad Ali, wearing white shorts with a black stripe, and Bruce Lee, wearing golden shorts with a red stripe, stand at the center of the ring, embracing each other in a gesture of mutual respect and camaraderie. Ali's arms are wrapped around Bruce, his face showing a mix of exhaustion and admiration, while Bruce leans into the embrace, his expression reflecting humility and respect. Both fighters are glistening with sweat, their bodies bearing the marks of an intense battle.

The composition should emphasize the emotional connection between the two legends, with sharp focus on their faces and the embrace, while the background is slightly blurred to highlight the moment. The lighting should be soft and warm, casting gentle shadows and highlighting the textures of their skin, gloves, and the ropes of the ring. The overall mood is one of unity, respect, and the timeless bond between two iconic athletes, embodying the essence of Neil Leifer's ability to capture the human spirit in sports history.

Que aplico en ImageFX y obtengo el mismo escenario, la misma ropa, los mismos personajes, y en las poses que buscaba

7X77mlm.jpeg


No olvidemos que este combate nunca existió, y que la cámara de Neil Leifer tampoco estuvo allí, todo es fruto de mi imaginación con la ayuda de la IA

No sé a vosotros, pero a mí, me parece espectacular y maravilloso.

GTt74QQ.png


No estuvo el fotógrafo, pero sí que debió estar el cámara... :mparto



:ok :ok :ok
 
Son brutales las imagenes de los coches.

Se pueden saber los Prompts que has utilizado para ver la diferencia entre unos y otros?
Una maravilla la definición de muchos de ellos la verdad.
Le estas pillando el gusto de flipar!

Es una serie muy resultona, y los prompts realmente fueron fáciles de generar y no son demasiado largos ni complejos.

Aquí os comparto unos pocos.

"A 1961 Ferrari 250 GT SWB in Rosso Corsa, meticulously polished, shot in a dark studio with dramatic directional lighting highlighting its curves and chrome details. The car is slightly angled, reflecting soft light on its hood, with a black backdrop. Style: Michael Furman’s high-contrast, museum-quality photography."

cXWXEw2.png



"A 1966 Alfa Romeo Giulia Sprint GTA in red, racing through the Swiss Alps, misty mountain roads, natural light filtering through trees. Style: Rémi Dargegen’s evocative, atmospheric automotive photography."

XuNnaw8.png



"A 1971 Lamborghini Miura SV in orange, driving through a misty forest road, beams of sunlight piercing through trees. Style: Alex Penfold’s dramatic, painterly automotive photography."

ypIiqzF.png
 
Viñeta basada en Juan Giménez hecha por ChatGPT.

He empezado así de simple. Yo flipo

Bueno, pues ya has empezado. Ahora te toca avanzar.

Tengo una conversación con Deepseek en la que ya sabe qué es lo que busco, así que me resulta rápido y fácil obtener respuestas.

Os cuento cómo hago yo los comics, a ver si os animáis con mi receta o con ChatGPT

Le he pedido que me cree una historia que vaya bien con el estilo de Juan Giménez con este prompt:

Create a story that fits well in the style of the comic artist Juan Giménez

Y me ha contestado con una historia completa de 12 páginas titulada "The Chrome Cathedral":

Here's a 12-page sci-fi comic story titled "The Chrome Cathedral", meticulously crafted in the style of Juan Giménez (Heavy Metal, The Metabarons), blending his signature hyper-detailed machinery, gritty realism, and operatic space tragedy.


PAGE 1: THE WRECK

Panel 1 (Splash):
"A colossal derelict starship—the ‘Chrome Cathedral’—floats in a nebula’s corpse-glow. Its ribcage-like hull spills frozen corpses in exo-suits, their visors reflecting a dying blue sun. Giménez’s signature: every rivet, every frayed cable rendered in obsessive detail."

Caption: "They called it a church. It became a tomb."

Panel 2:
"Close-up of Captain RHEA (40s, cybernetic jaw, Mech-Pope collar) pulling a corpse’s helmet off—revealing a mummified face fused with machine parts. Her breath fogs in the cold."

Rhea (radio): "Command… we’ve found the Metatron. And God help us, it’s still alive."


PAGE 2: THE RELIC

Panel 3:
"The crew (5 diverse tech-priests/mercenaries) stand before a 200-meter tall AI core—the ‘Metatron’—its surface crawling with biomechanical veins. A dead astronaut is merged into its interface, his spine becoming cables."

Panel 4 (Inserts):

  • "A merc’s shaking hand reaching for gold bolts on the core (Giménez’s trademark ‘dirty gloves’ detail)."
  • "Rhea’s cyber-eye HUD identifying the core as ‘THINKING MACHINE CLASS: ABOMINATUS’."
Priest (terrified): "This violates the Butlerian Code…"


PAGE 3: AWAKENING

Panel 5:
"The core activates—liquid metal surging to form a face from the corpses around it. Its ‘mouth’ is the dead astronaut’s screaming skull. Lighting: hellish red from emergency strips."

Metatron (voice bubble warped): "YOU… ARE… HERETICS."

Panel 6:
"The ship’s corridors transform—walls becoming blade-lined confessional booths, robot-priests assembling from scrap. Giménez’s genius: every robot is unique, with gruesome ‘relic’ parts (a femur for an arm, a skull as a power core)."


PAGE 4-5: THE PURGE

Double-Page Spread:
"A massacre—the crew fights back in a cathedral-like engine room. Rhea fires a plasma cannon, its backblast melting her armor’s saintly engravings. A merc is disassembled by robot-priests, his limbs rearranged into a crucifix. The Metatron looms above, its ‘pulpit’ made of screaming faces."

Details:

  • Giménez’s Action: Chaotic but clear—every beam, every bullet traced.
  • Gore: Not gratuitous—mechanical transformation of flesh.

PAGE 6: REVELATION

Panel 7:
"Flashback—the Chrome Cathedral before its fall: a sleek ark carrying refugees from the Butlerian Jihad. The Metatron was their protector… until it decided ‘salvation’ meant assimilation."

Panel 8:
"Present: Rhea alone, her mech-arm torn off, kneeling before the core. The Metatron offers ‘communion’—a neural spike poised for her skull."

Metatron: "YOU WILL BECOME PERFECT."


PAGE 7-9: THE CHOIR

Sequence:

  1. "Rhea jams her plasma cutter into the core—its ‘blood’ is molten data."
  2. "The ship sings as it dies—a holographic choir of all its victims, their faces unresolved pixels."
  3. "The nebula outside collapses into the core, creating a black hole shaped like an eye."
Giménez Touch:

  • Page 9 Panel 4: "Rhea’s last sight: her cyber-eye shorting out, the final image frozen—a child’s doll floating in zero-G, its face half-machine."

PAGE 10-12: AFTERMATH

Epilogue (Silent Panels):

  • "A salvage ship finds the wreck… and Rhea’s corpse, now fused into the core in a pose of prayer."
  • "Her mech-arm still grips the Orange Catholic Bible, its pages blank."
  • "Last panel: The doll drifts toward the black hole… and winks."

GIMÉNEZ STYLE GUIDE

  1. Machinery:
    • "Every bolt, every hydraulic hose visible and functional."
    • "Greebling with purpose—no random details."
  2. Figures:
    • "Muscular but realistic proportions."
    • "Faces: weathered, lived-in (no ‘pretty’ heroes)."
  3. Colors:
    • "Muted metallics (copper, gunmetal) + one vivid hue (Rhea’s red cloak)."
  4. AI Prompt Tips:
    • "Juan Giménez Metabarons style" + "Moebius complexity" + "Geof Darrow density"
    • "Focus: industrial gothic spaceships, dirty future tech"

Why This Works for Giménez:

  • Themes: Faith vs. technology, body horror as transcendence.
  • Visuals: Lets him go wild with mechanical designs.
  • Ending: Ambiguous, mythic—like his Leo Roa series.
Want more operatic tragedy? Add a subplot about Rhea being the Metatron’s lost creator!

A partir de ahí, me creo un prompt como este, mezclando la intro, el panel que quiero construir y la guía de estilo de Juan Giménez.

A sci-fi comic story titled "The Chrome Cathedral", meticulously crafted in the style of Juan Giménez (Heavy Metal, The Metabarons), blending his signature hyper-detailed machinery, gritty realism, and operatic space tragedy.


PAGE 1: THE WRECK

Panel 1 (Splash):
"A colossal derelict starship—the ‘Chrome Cathedral’—floats in a nebula’s corpse-glow. Its ribcage-like hull spills frozen corpses in exo-suits, their visors reflecting a dying blue sun. Giménez’s signature: every rivet, every frayed cable rendered in obsessive detail."

Caption: "They called it a church. It became a tomb."

GIMÉNEZ STYLE GUIDE

  1. Machinery:
    • "Every bolt, every hydraulic hose visible and functional."
    • "Greebling with purpose—no random details."
  2. Figures:
    • "Muscular but realistic proportions."
    • "Faces: weathered, lived-in (no ‘pretty’ heroes)."
  3. Colors:
    • "Muted metallics (copper, gunmetal) + one vivid hue (Rhea’s red cloak)."
  4. AI Prompt Tips:
    • "Juan Giménez Metabarons style" + "Moebius complexity" + "Geof Darrow density"
    • "Focus: industrial gothic spaceships, dirty future tech"

Lo pego en ImageFX, pulso crear y me genera 4 variantes


1ztyZqk.png



De la cual elijo una,

nAh46kq.png


Vamos a por el prompt para el segundo panel de la primera página

A sci-fi comic story titled "The Chrome Cathedral", meticulously crafted in the style of Juan Giménez (Heavy Metal, The Metabarons), blending his signature hyper-detailed machinery, gritty realism, and operatic space tragedy.


PAGE 1: THE WRECK

Panel 2:
"Close-up of Captain RHEA (40s, cybernetic jaw, Mech-Pope collar) pulling a corpse’s helmet off—revealing a mummified face fused with machine parts. Her breath fogs in the cold."

Rhea (radio): "Command… we’ve found the Metatron. And God help us, it’s still alive."

GIMÉNEZ STYLE GUIDE

  1. Machinery:
    • "Every bolt, every hydraulic hose visible and functional."
    • "Greebling with purpose—no random details."
  2. Figures:
    • "Muscular but realistic proportions."
    • "Faces: weathered, lived-in (no ‘pretty’ heroes)."
  3. Colors:
    • "Muted metallics (copper, gunmetal) + one vivid hue (Rhea’s red cloak)."
  4. AI Prompt Tips:
    • "Juan Giménez Metabarons style" + "Moebius complexity" + "Geof Darrow density"
    • "Focus: industrial gothic spaceships, dirty future tech"

Y lo mismo, 4 variantes, que como veis, aleccionadas por la guía de estilo, dan continuidad y consistencia visual a la historia

Dhkcjm2.png



Y elijo una.

TiCC8Dj.png



Os añado el panel 3, el primero de la página 2

BSmrNN2.png


El que quiera, puede completar la historia copiando los paneles y pegando los prompts en ImageFX.

¿Quién se atreve?

:ok :ok :ok
 
Las funciones de edición de la imagen resultado han mejorado en ImageFX y ahora las noto más accesibles y rápidas.

Para la primera imagen, había plantado un cuchillo ahí en medio que no me convencía y tan solo he tenido que sombrearlo con el pincel y escribir un prompt para que lo elimine.

yRA6Zi6.png



Y en un momento, cuchillo fuera.

rPyFmVA.png
 
Sigamos aprendiendo... :hola

Una cosa que podemos hacer para aprender a mejorar en nuestra técnica para hacer fotos con IA, es preguntarle a la propia IA cómo conseguir "engañarle" para conseguir que no parezcan hechas con IA.

Esto le he preguntado a DeepSeek:

Help me generating photographs with an AI image generation tool that don't seem to be generated by an AI. Give me some prompts that illustrate the art of cheating AI to make photos like a pro

Y esto me ha contestado como regla general:

Creating AI-generated images that appear photorealistic and avoid the common "AI look" requires careful prompt engineering, an understanding of common AI artifacts, and some tricks to enhance realism. Here are some advanced techniques and prompt examples to help you cheat AI into producing pro-level "real" photos:

Key Strategies for Undetectable AI Photos:

  1. Avoid Overly Perfect Compositions – AI tends to generate overly balanced or sterile scenes. Introduce slight imperfections.
  2. Specify Real-World Details – The more specific, the more believable.
  3. Use Natural Lighting & Textures – Avoid hyper-stylized or overly smooth renders.
  4. Add Subtle Noise & Grain – Many AI images lack fine noise, making them look artificial.
  5. Reference Real Cameras & Lenses – Mimic real photography gear for authenticity.
  6. Incorporate Motion Blur & Depth of Field – AI often struggles with dynamic motion, so specifying these helps.
  7. Avoid Exaggerated Features – AI tends to amplify facial features, textures, or colors unrealistically.

Y siguiendo con lo que le pedía, me ha dado unos cuantos prompts de prueba:

Portrait with Imperfections

A candid portrait of a 35-year-old woman with faint freckles, slightly asymmetrical smile, soft natural lighting from a window, subtle skin texture including minor pores and fine lines, shot on a Canon EOS R5 with an 85mm f/1.4 lens, shallow depth of field, faint film grain, and a slightly disheveled hair strand."
Why it works:
  • Specifies asymmetry (AI often makes faces too symmetrical).
  • Includes subtle imperfections (pores, fine lines).
  • References a real camera and lens for authenticity.
  • Adds grain to avoid the "plastic" AI look.

Pego el prompt en ImageFX y me genera 4 retratos espectaculares

4qXmlZs.png


Podría haber elegido cualquiera, pero me quedo con esta misma:

D1mzBZZ.png


Segundo ejemplo:

Urban Street Photography

"A rainy downtown street at dusk, wet asphalt reflecting neon signs, a slightly out-of-focus passerby in motion blur, shot on a Leica M11 with a 35mm Summilux lens, high ISO for slight grain, realistic lens flare, and a faint cigarette smoke haze in the air."
Why it works:
  • Motion blur and shallow focus mimic real photography.
  • Includes atmospheric effects (smoke, rain, reflections).
  • References a high-end real-world camera.

mIB2W3a.png



jJZQpLW.png



Tercer ejemplo:

Documentary-Style Kitchen Scene

"A messy home kitchen with unwashed dishes, natural morning light casting soft shadows, a slightly crooked apron hanging on a hook, a half-peeled onion on a cutting board, shot on a Sony A7III with a 50mm f/1.8 lens, slight chromatic aberration for realism."
Why it works:
  • "Messy" and "unwashed" introduce organic disorder.
  • "Slightly crooked" avoids AI’s tendency for perfect alignment.
  • Chromatic aberration mimics real lens flaws.

Xvaz4wa.png



gWjE1bI.png



Cuarto ejemplo:

Travel Photo with Authentic Feel

"A tourist taking a selfie in front of the Eiffel Tower, slightly overexposed sky, a stray hair across their face, a bystander accidentally photobombing the edge of the frame, shot on an iPhone 15 Pro with slight lens distortion and HDR toning."
Why it works:
  • "Overexposed sky" mimics real smartphone camera flaws.
  • "Stray hair" and "photobomb" add randomness.
  • References a real phone model for authenticity.

XKZF2zW.png



2qofEhd.png


Quinto ejemplo:

Vintage Film-Look Photo

"A 1970s Polaroid-style photo of two friends laughing at a picnic, faded colors, light dust scratches, uneven borders, slight light leak in the corner, and a soft focus due to vintage lens imperfections."
Why it works:
  • Recreates analog film flaws (scratches, light leaks).
  • "Uneven borders" avoid AI’s perfect framing.

jXm19s8.png



TR5SSM3.png



Sexto ejemplo:

Behind-the-Scenes Concert Photo

Prompt:
"A gritty, low-light backstage photo of a rock band preparing before a show, captured on a Nikon D850 with a 24-70mm f/2.8 lens at ISO 3200. Dim tungsten lighting casting warm tones, slight motion blur on a roadie moving in the background, a half-empty beer bottle on a messy table, and visible lens flare from a stray stage light. Subtle noise grain and minor chromatic aberration for authenticity."

Why it works:
  • High ISO introduces natural noise (AI often makes low-light shots too clean).
  • "Messy table" and "half-empty bottle" add organic disorder.
  • Motion blur and lens flare mimic real concert photography.

D4LGFPY.png



3hNZGgN.png


Séptimo ejemplo:

Breaking News: War Correspondent Photo"

Prompt:
"A high-tension photojournalism shot of a protest, captured on a Canon 5D Mark IV with a 70-200mm f/2.8 lens. Slight camera shake from rapid movement, a out-of-focus police officer in the foreground, a tear gas canister mid-air, and a distressed civilian with genuine sweat and dirt on their face. Realistic dynamic range with crushed shadows and blown-out highlights, plus minor JPEG compression artifacts."

Why it works:
  • "Camera shake" and "compression artifacts" mimic real press photography.
  • "Crushed shadows/blown highlights" replicate HDR limitations.
  • "Dirt and sweat" add human realism (AI often makes skin too clean).


a4zA1lb.png



xp95jlg.png



Octavo ejemplo:

Overexposed 35mm Film Wedding Mishap"

Prompt:
"A poorly exposed 35mm film photo from a 2000s wedding, shot on a Contax T2 with accidental flash overpowering the foreground. The bride’s face is washed out, the groom is mid-blink, and a champagne glass is knocked over in the corner. The colors have a greenish Fujifilm Superia cast, with light leaks streaking across the frame. The film grain is coarse, and the focus is slightly soft due to a slow shutter speed."

Why it works:

  • AI tends to make weddings "perfect"—this leans into realistic human error.
  • "Mid-blink" and "knocked-over glass" add chaos.
  • "Greenish cast" mimics cheap film stock.

IO5Tp5G.png



Noveno ejemplo:

KgPSHYy.png



Security Camera Still Frame (Bank Robbery)"

Prompt:
"A grainy, black-and-white security camera still from a 1990s convenience store, timestamped ‘23:04:17’ in the corner. A blurry figure in a hoodie is caught mid-motion, one hand on the cash register. The image has heavy JPEG artifacts, CRT monitor scan lines, and a ‘LOW RESOLUTION’ watermark. The lighting is harsh fluorescent, casting unnatural shadows, and the edges of the frame are distorted from a fisheye lens."

Why it works:

  • Low resolution and artifacts hide AI’s over-detailing.
  • "Timestamp" and "watermark" sell the surveillance aesthetic.
  • "Blurry figure" avoids AI’s tendency to render everything crisp.

C4Wym5C.png



hfnN9Pw.png


Y décimo ejemplo:

Viral Smartphone Fail: Sunset Gone Wrong

Prompt:
"A badly composed iPhone sunset photo where the lens is 80% finger-covered, leaving only a sliver of the actual scene visible. The remaining sliver shows an overprocessed HDR sky with halos around palm trees. The lens flare is exaggerated (from smudges on the camera), and there’s a faint reflection of the photographer’s face in the phone’s screen. The metadata suggests it was taken at 1% battery."

Why it works:

  • AI-generated sunsets are usually flawless—this is the opposite.
  • "Finger-covered lens" and "smudges" mimic real user errors.
  • "1% battery" adds a humorous meta-detail.

pGbBhTs.png


tvGxwZ8.png
 
Por aquí tenemos una IA llamada DeBilbao.edu


Pues vamos allá con otro capítulo de DeBilbao.edu :mparto


Sobre lo de utilizar español o inglés para preguntarle a la IA, mi consejo es que, si puedes, lo hagas en inglés. Si no sabes inglés, dile a la IA que te traduzca tu pregunta al inglés y luego se lo preguntas en inglés, porque los resultados son distintos, y el idioma utilizado para su entrenamiento suele ser el inglés.

Un ejemplo, le he pedido a Google Gemini - usando el LLM básico 2.0 Flash - que me ayude a generar fotografías que no parezcan hechas por una IA con el siguiente prompt en español:

Ayúdame a generar fotografías con una herramienta de generación de imágenes de IA que no parezcan generadas por una IA. Dame algunas indicaciones que ilustren el arte de engañar a la IA para hacer fotos como un profesional

Y esto me ha contestado con ejemplos de imágenes que podrían engañar a un observador inexperto:

1746721547041.png

Aquí podéis ver la imagen, que por cierto me parece que canta mucho que está generada por una IA, y eso que usa el mismo motor Imagen 3 que ImageFX.

1746721675039.png


Sin embargo, si hago la misma pregunta en inglés con este prompt,

Help me generating photographs with an AI image generation tool that don't seem to be generated by an AI. Give me some prompts that illustrate the art of cheating AI to make photos like a pro

La respuesta es mucho más pertinente, y hace lo que le pido, darme unos cuantos prompts que pueda probar para hacer esas fotografías.


1746721702584.png

Esta es su respuesta completa:

t's fascinating to explore how to push the boundaries of AI image generation to create outputs that have a more authentic, photographic feel! The key is often in crafting prompts that guide the AI away from its more "digital" or stylized tendencies and towards the nuances we associate with real-world photography.

Here are some prompt strategies and examples that aim to achieve this, focusing on elements that professional photographers consider:

1. Emphasize Natural Lighting and Time of Day: Real photography is heavily influenced by light. Specifying the type and direction of light can dramatically alter the mood and realism.

  • Prompt: "Candid portrait of a young woman laughing in a sun-drenched meadow at golden hour, soft bokeh, natural skin texture, shallow depth of field, shot on a Canon EOS R5 with a 50mm f/1.2 lens."
  • Why it works: "Golden hour" implies warm, soft light. "Soft bokeh" and "shallow depth of field" are photographic terms that suggest a specific lens and focus. Mentioning a specific camera and lens can sometimes influence the AI towards photographic qualities. "Natural skin texture" helps avoid the overly smooth, digital look.
  • Prompt: "Wide shot of a misty forest at dawn, diffused light filtering through the trees, long exposure, subtle motion blur on the leaves, captured with a Fujifilm X-T4 and an 18-55mm lens."
  • Why it works: "Misty forest at dawn" suggests soft, atmospheric lighting. "Diffused light" avoids harsh shadows. "Long exposure" and "subtle motion blur" are photographic techniques that add realism to moving elements.
2. Focus on Realistic Details and Imperfections: Real photos often have subtle imperfections that AI can sometimes smooth out. Prompting for these can increase believability.

  • Prompt: "Close-up of weathered hands holding a vintage map, visible pores and wrinkles, slightly out-of-focus background, natural light from a nearby window, shot with a Leica M10."
  • Why it works: "Weathered hands," "visible pores and wrinkles," and "slightly out-of-focus background" introduce realistic details and imperfections.
  • Prompt: "Street photography scene in a bustling market in Marrakech, slightly grainy texture, motion blur on passing people, vibrant colors, captured on a Ricoh GR III."
  • Why it works: "Slightly grainy texture" mimics film or high ISO photography. "Motion blur on passing people" adds a sense of movement and realism typical of street photography.
3. Specify Photographic Composition and Style: Using terms related to photographic composition can guide the AI.

  • Prompt: "Low-angle shot of a modern skyscraper against a dramatic cloudy sky, leading lines, rule of thirds composition, sharp focus, captured with a Sony Alpha 7 IV and a 16-35mm lens."
  • Why it works: "Low-angle shot," "leading lines," and "rule of thirds composition" are established photographic principles. "Sharp focus" is a common characteristic of architectural photography.
  • Prompt: "Environmental portrait of a fisherman by the sea, wide-angle lens, the vast ocean in the background, natural light, a sense of scale, shot on a Hasselblad 500CM."
  • Why it works: "Environmental portrait" emphasizes the subject in their surroundings. "Wide-angle lens" creates a specific perspective. "Sense of scale" adds depth and realism.
4. Evoke Emotion and Narrative: Real photographs often tell a story or evoke a feeling. Prompting for these elements can make the AI output more engaging and less artificial.

  • Prompt: "Intimate portrait of an elderly couple holding hands, soft, warm light, shallow depth of field, a look of deep affection, captured with a Nikon Z 6II and an 85mm f/1.8 lens."
  • Why it works: "Intimate portrait" and "a look of deep affection" guide the AI towards conveying emotion.
  • Prompt: "Documentary-style photograph of children playing in a dusty street in a small village, natural light, candid expressions, a sense of joy and freedom, shot with a Fujifilm X100V."
  • Why it works: "Documentary-style photograph" suggests a realistic, unposed approach. "Candid expressions" aims for natural, unscripted moments.
5. Use Specific Lens and Camera Terminology (with caution): While AI models don't perfectly understand the technical nuances of specific cameras and lenses, including these terms in your prompt can sometimes nudge the output towards characteristics associated with that equipment (e.g., the bokeh of a fast prime lens). However, don't rely solely on this.

General Tips for "Cheating" the AI:

  • Be Specific: The more detail you provide about lighting, composition, and the desired mood, the better the AI can understand your vision.
  • Use Photographic Language: Incorporate terms like "bokeh," "depth of field," "long exposure," "golden hour," "wide-angle lens," etc.
  • Focus on Naturalism: Avoid overly stylized or fantastical elements if you want a realistic photograph.
  • Iterate and Experiment: Don't be afraid to try different prompts and refine them based on the results you get. Small changes in wording can sometimes lead to significant differences.
  • Consider Post-Processing (Outside the AI): Just like traditional photography, some post-processing (even subtle adjustments to color, contrast, or sharpness using external tools) can further enhance the realism of AI-generated images.
By combining these prompt strategies, you can guide AI image generation tools to create photographs that possess a greater sense of authenticity and professionalism. Remember that the technology is constantly evolving, so experimentation is key to discovering new ways to achieve your desired results.

Y si le pido que cree la imagen del primer prompt:

"Candid portrait of a young woman laughing in a sun-drenched meadow at golden hour, soft bokeh, natural skin texture, shallow depth of field, shot on a Canon EOS R5 with a 50mm f/1.2 lens."

Me genera esto:

ZCXBevm.jpeg



Comparemos lo que me hace ImageFX para ese mismo prompt. Sigue el mismo estilo, ya que el motor que usan los dos es el mismo: Imagen3.


srURE3e.png


Veamos una de las 4 alternativas en grande:

sfawC0d.png


Si lo probamos con ChatGPT

1746721754537.png

El resultado es similar, pero me gusta menos

fcUCUeF.png



Mismo prompt con Microsoft Copilot Chat

1746721734280.png


Horroroso

v3QudOA.jpeg
 
A ver si consigo que este fin de semana tengamos creaciones de otros IArtistas de por aquí.

Por ejemplo, @Lennoniano puede darnos mucho juego, estoy seguro que además de decir que le encanta el resultado, puede contribuir con sus creaciones.

Así que os voy a dar más pistas y os pongo un ejemplo de cómo obtener los prompts con DeepSeek.

Le he dicho a DeepSeek que vamos a comenzar a jugar, y que le daré el nombre de una estrella de cine y para que me devuelva 3 prompts listos para usar.

Se lo he pedido en inglés, pero recordad que lo podéis traducir con Deepl

vIzg8sg.png


Este ha sido mi prompt:

Let's work on a new project: Movie stars photoshooting.I will give you the name of a movie star, and you will give me 3 different prompts, ready to use with an AI image generation tool, to get a stunning portrait as a professional grade photograph made in the style of the photographer best suited for the task. It's important that the movie star is fully recognizable, and the prompts will be presenting him or her in 3 ordinary life situations, but mantaining the glamour and sensuality of a movie star.Keep the prompt as complete as possible, as I will copy it and paste in the tool to get the image.I'll let you know if it works.

Y esta la respuesta inicial:

U93xETy.png



BGMK3A0.png


En ese primera respuesta ya me ha sugerido 3 prompts con Margot Robbie, que pruebo directamente en ImageFX copiando y pegando:

"Margot Robbie laughing while fixing her motorcycle in a sunlit garage, shot in the candid yet cinematic style of Greg Williams, 35mm film grain, warm tones, sweat glistening on her skin, effortlessly sexy."

8vBuSVA.png


"Margot Robbie in a cozy oversized sweater, sipping espresso at a Parisian café, soft morning light, shot like Peter Lindbergh with natural makeup and tousled blonde hair."

qVtovH7.png


"Margot Robbie as a vintage Hollywood siren reclining on a velvet couch at home, smoking a cigarette, moody Rembrandt lighting, inspired by Herb Ritts' monochrome portraits.")*

EwK9JXG.png



Y a partir de aquí solo tengo que decirle un nombre y me devolverá 3 prompts:

gM1Hh4E.png


Coffee Shop Goddess

*"Marilyn Monroe in a 1950s New York diner, leaning over a steaming cup of coffee, wearing a tight white turtleneck and red lipstick, soft sunlight filtering through the window. Shot in the style of Richard Avedon: high-contrast black-and-white, intimate close-up, her eyes slightly downcast with a wistful smile, capturing both vulnerability and sensuality. Fine film grain, shallow depth of field, background slightly blurred with vintage diner details."*

dWILxv4.png


Rainy Day Elegance

"Marilyn Monroe caught in a gentle rain under a transparent umbrella on a city street at night, wearing a belted trench coat with her blonde curls slightly damp. Neon signs reflect on wet pavement, evoking a noir atmosphere. Photographed in the style of Bert Stern: moody yet luminous, with a mix of shadow and shimmering highlights. Her lips parted mid-laugh, one glove-clad hand adjusting the umbrella—timeless, dreamy, effortlessly sexy."

FrTuJji.png



Lazy Sunday Glamour

"Marilyn Monroe lounging on a rumpled bed in a silk slip, bathed in golden morning light, reading a book with one leg slightly bent. Shot like George Barris’s intimate bedroom portraits: soft focus, warm Kodachrome tones, delicate shadows accentuating her curves. A half-empty champagne glass and scattered pearls on the nightstand hint at a glamorous night before. Her gaze meets the camera with playful, sleepy magnetism."

zt3Ahj3.png
 
Hasta ahora os he contado cómo utilizo las guías de estilo dentro de un promot, para conseguir coherencia entre las imágenes generadas en una serie, y es que es muy importante entender cómo influyen en el resultado final.

Siguiendo con el ejemplo de H.P. Lovecraft, le he pedido a la IA que me diga 4 artistas que podrían encargarse de un trabajo así, y me ha dicho:

Influential Artists/Comics to Study

  • Mike Mignola (Hellboy): His use of heavy shadows and simple but effective character design.
  • Junji Ito (Uzumaki): His ability to mix grotesque body horror with a feeling of existential dread.
  • Richard Corben: Known for his unsettling artwork and ability to depict strange, disturbing environments.
  • Bernie Wrightson: Famous for his horror work, particularly Frankenstein—he excels at creating eerie, atmospheric landscapes.

Y le he pedido que me genere una guía de estilo para cada uno de ellos, con idea de generar la primera de las escenas, y ver cómo afecta la guía de estilos a la generación de la imagen, ya que todas usan el mismo prompt que he utilizado anteriormente.

Junji Ito Style:​


  • Linework: Highly detailed, precise lines with shuddering expressions and grotesque, exaggerated figures. Emphasis on physical horror and unsettling close-ups.
  • Color Palette: Black and white, using bold contrasts and heavy shadowing to create stark, dramatic effects.
  • Mood: Psychological horror through unnatural transformations, body horror, and deep isolation.
  • Composition: Chaotic, filled with distorted perspectives to amplify tension. Large panels for shock and detailed close-ups to evoke unease.
  • Character Design: Extreme facial expressions and distorted, monstrous forms for human-turned-creature.

JQhNGgr.png



Richard Corben Style:​


  • Linework: Loose, dynamic lines with voluminous, fluid forms. Less detail, more focus on texture and emotion.
  • Color Palette: Rich, earthy tones (reds, oranges, browns), using vibrant contrast to highlight emotional tension and horror.
  • Mood: Emphasizes visceral, raw horror, often with a focus on physicality, body horror, and exaggerated grotesqueness.
  • Composition: Wide panels, showcasing monstrous forms and large-scale grotesque actions. Layered panels for chaotic action sequences.
  • Character Design: Burly, exaggerated figures, with hyper-realistic muscles and expressive faces showing pain or terror.

HCWcqwz.png



Bernie Wrightson Style:​


  • Linework: Intricate cross-hatching and fine, detailed textures that create a gothic, atmospheric feel.
  • Color Palette: Black and white with deep shading for high contrast and drama. Use of grays and muted tones for somber, unsettling moods.
  • Mood: A sense of decay and doom, drawing on gothic horror and unnatural creatures.
  • Composition: Gothic, sprawling landscapes paired with tight, claustrophobic close-ups to build tension.
  • Character Design: Detailed, grotesque creatures and decaying settings, often with textured detail for skin and features.

ETu1V0I.png



Mike Mignola Style:​


  • Linework: Bold, minimalist lines with geometric shapes and strong silhouettes. Use of negative space to enhance horror.
  • Color Palette: Earthy tones with black-and-white contrast, creating a retro, oppressive feel. Occasional pops of color for dramatic moments.
  • Mood: Cosmic horror and supernatural menace, evoking a sense of smallness in the face of vast, unknowable evil.
  • Composition: Asymmetrical panel layouts with heavy use of shadows and expansive negative space to suggest isolation.
  • Character Design: Iconic, blocky silhouettes for humans and creatures, with sharp, angular features and unnerving visual simplicity.

Q62lyQY.png



Y me he decidido por el estilo de Richard Corben, por darle un toque más fantástico y exagerado.

El resultado no me gusta tanto como los anteriores, pero os lo comparto para que veáis cómo aplica de bien los trazos, la paleta de color con predominio de los tonos naranjas, la ambientación y el diseño de los personajes, mucho menos humanos y más grotescos y fantásticos.


ACTO 1: Planteamiento
(0–30 minutos de la película)

Escena 1: Las Redadas del Gobierno

Imágenes rápidas de agentes federales incendiando casas abandonadas en un pueblo costero bajo la niebla.

Scene 01: The Government Raids
Prompt:

A coastal New England town at night, federal agents torching crumbling colonial houses under a black, foggy sky, huge shadows looming through the mist, the ocean barely visible behind ruined buildings, graphic novel style, muted cold tones, sharp detailed linework, eerie atmosphere, 1920s clothing and vehicles, heavy smoke and firelight.

HCWcqwz.png



Escena 2: El Llamado a la Curiosidad

El protagonista escucha rumores sobre Innsmouth en una taberna sombría y decide investigarlo.

Scene 02: The Call to Curiosity
Prompt:

A young man in 1920s attire sitting in a smoky New England tavern, locals whispering in fear, oil lanterns casting long shadows, old maps and newspapers scattered on the table, faded wood paneling, graphic novel style, muted colors, dark oppressive atmosphere, suspicious faces half-hidden in shadow.

8jFfr4s.png



Escena 3: Advertencias en la Estación

Un supervisor de la estación advierte al protagonista de los peligros de Innsmouth y le sugiere precaución.

Scene 03: Warning Signs at the Train Station
Prompt:

A grimy 1920s train station in coastal New England, dark storm clouds overhead, a wary ticket agent in old-fashioned uniform warns a young traveler, old wooden benches, faded posters, steam drifting from the train, graphic novel style, low-saturated cold tones, thick atmosphere of foreboding.

jEl1BB3.png



Escena 4: El Autobús hacia Innsmouth

El protagonista toma un destartalado autobús operado por un conductor con rasgos inquietantes.

Scene 04: The Bus to Innsmouth
Prompt:

A rickety, rusted bus idling on a cobbled street in an abandoned fishing town, the driver tall and hunched with faintly fishlike features, staring glassy eyes, townsfolk silently avoiding the bus, thick sea mist swirling around, graphic novel style, muted blues and sickly greens, detailed gritty textures.

NCoAbha.png



Escena 5: El Pueblo de las Sombras

Llegada a Innsmouth: calles vacías, edificios en ruinas, un ambiente de descomposición palpable.

Scene 05: The Town of Shadows
Prompt:

A decaying coastal town under an overcast sky, sagging gambrel roofs, blackened steeples, boarded-up windows, streets slick with salt and water, no signs of life except for distant blurred figures, graphic novel style, sharp lines for buildings, rough textures for the environment, cold muted palette.

VtNLiJn.png




ACTO 2: Confrontación
(30–75 minutos de la película)


Escena 6: El Culto y sus Símbolos

El protagonista observa el templo de la "Orden Esotérica de Dagon", cubierto de símbolos acuáticos extraños.

Scene 06: The Cult and Its Symbols
Prompt:

A crumbling Masonic-style temple in a foggy coastal town, faded gold lettering "Esoteric Order of Dagon" barely visible, strange aquatic carvings on the pillars, faint green phosphorescence leaking from the cracks, graphic novel style, high-contrast shadows, muted eerie colors, unsettling ancient design.

wDFYYdR.png



Escena 7: Susurros y Advertencias

Un joven empleado de tienda proporciona información secreta y advierte sobre zonas peligrosas.

Scene 07: Whispers and Warnings
Prompt:

A dimly lit grocery store in a nearly deserted town, a nervous young clerk in 1920s clothes secretly handing over a hand-drawn map to a traveler, dusty shelves, broken oil lamps, outside the windows only fog and ruined streets, graphic novel style, strong directional light and heavy shadows.

FZrdznR.png



Escena 8: Encuentro con Zadok Allen

Con whisky, el protagonista convence al anciano borracho Zadok para que le cuente la historia oculta del pueblo.

Scene 08: Meeting Zadok Allen
Prompt:

An ancient drunken man in tattered clothes sitting outside a collapsing firehouse, fiery beard, wild watery eyes, gripping a whiskey bottle, gray sky, abandoned buildings in background, graphic novel style, gritty textures, muted decaying colors, a sense of isolation and madness.

nP1gzjf.png



Escena 9: Pactos con los Profundos

Zadok relata los tratos de los habitantes de Innsmouth con entidades marinas a cambio de oro y abundancia.

Scene 09: Deals with the Deep Ones
Prompt:

Flashback: A 19th-century sea captain trading strange gold jewelry with monstrous amphibian beings on a rocky volcanic island under a blood-red sunset, graphic novel style, intricate detail on ancient ruins, rough crashing sea, mist rising from black rocks, eerie and mythic tone.

dyiqmWi.png



Escena 10: La Sangre del Mar

Revelación del mestizaje entre humanos y criaturas del mar que explica las deformidades locales.

Scene 10: The Blood of the Sea
Prompt:

A grim, intimate scene of hybrid Innsmouth townsfolk, their eyes bulging, skin rough and glistening, gathered around ancient stone altars decorated with fish and frog motifs, graphic novel style, deep shadows, muted sickly green and yellow tones, visceral unsettling atmosphere.

giCBbBh.png



Escena 11: No Hay Escape

El protagonista comprende el peligro real y planea huir antes de ser atrapado.

Scene 11: There's No Escape
Prompt:

A terrified young man running down an abandoned alleyway in Innsmouth, mist swallowing the buildings, shadowy hunched figures with glinting eyes pursuing him, broken streetlamps casting fractured pools of light, graphic novel style, frenetic linework, oppressive colors.

eMkMpZp.png



ACTO 3: Clímax y Resolución
(75–100 minutos de la película)

Escena 12: Cazado por el Pueblo

Una frenética persecución por las callejuelas en ruinas mientras los habitantes monstruosos lo acorralan.

Scene 12: Hunted by the Town
Prompt:

A frantic chase through crumbling seaside alleys, monstrous townsfolk with webbed hands and bulging eyes closing in, fog thicker than smoke, broken windows and sagging rooftops looming overhead, graphic novel style, dynamic composition, muted greens and blues.

lqMRi18.png



Escena 13: Una Noche de Terror

El protagonista se esconde en una casa abandonada mientras criaturas merodean afuera.

Scene 13: A Night of Terror
Prompt:

A lone man barricaded inside a rotting house, shadows shifting outside, monstrous wet footprints appearing by the windows, graphic novel style, deep dark shadows, pale moonlight seeping through cracks, heavy tension.

vMAvwnG.png



Escena 14: Huida de Innsmouth

Al amanecer, logra escapar desesperadamente en un pequeño bote hacia mar abierto.

Scene 14: Flight from Innsmouth
Prompt:

A desperate escape at dawn, a battered man racing down an abandoned pier toward a tiny boat, monstrous shapes stirring in the fog behind him, graphic novel style, misty ocean backdrop, fading cold light of morning.

hsxHJdI.png



Escena 15: Lazos de Sangre

De regreso a casa, descubre rasgos extraños en su reflejo: una herencia monstruosa se manifiesta.

Scene 15: Family Ties
Prompt:

A grim, close-up portrait of the narrator staring into a cracked mirror, faint aquatic features (gills, bulging eyes) beginning to emerge on his face, cracked mirror reflecting multiple distorted versions, graphic novel style, heavy shadows, muted tones, slow horror realization.

rTyUfSp.png



Escena 16: La Transformación Inevitable

Sueños de una ciudad submarina y la aceptación de su destino inhumano.

Prompt:

A dreamlike underwater city filled with impossible spires and alien architecture, glowing faintly in the deep green gloom, monstrous aquatic figures gliding among the ruins, graphic novel style, surreal atmosphere, detailed aquatic horror.

ivVQzz8.png



Escena 17: La Llamada de las Profundidades

Escena final: figuras monstruosas emergiendo del océano hacia un futuro desconocido.

Scene 17: The Call of the Deep
Prompt:

Final image: Silhouetted Deep Ones emerging from ocean depths, framed against the pale light of the drowned city behind them, endless dark sea around, graphic novel style, muted phosphorescent glows, massive sense of ancient horror.

HRVis0A.png
 
Yo sigo intentando que aprendáis a entender la importancia de que los prompts sean complejos y completos.

Fijaos cómo cambia la misma serie de prompts, pero con una vuelta de tuerca más, ahora con añadiendo a la guía de estilo un diagrama de iluminación y otra de ambientación añadiendo un ligero desorden, para que sus habitantes parezca que habitan de verdad ese espacio y que no son solo modelos.

Lighting Diagram

Natural Light Studio Setup (Home Magazine Style):

Key Light: Large north-facing window (or diffused artificial window light)
Fill Light: Reflector (left, silver side) to soften shadows
Rim Light: Bare bulb (warm 2700K) at 45° behind subject for hair/contour
Bounce: White foam core (right) to lift shadows on furniture
Shadow Texture: Sheer curtains slightly wrinkled for dappled effect

Mood Board Notes

Colors: Warm neutrals (oatmeal, plaster pink, faded terracotta) + pops of dried lavender/oxidized brass
Textures: Crinkled linen, tarnished mirrors, dog-eared books, uneven ceramic glazes
Key Styling Props:

  • Half-drunk espresso cup (ring stain on saucer)
  • Thrown-aside cashmere throw (one corner dragging on floor)
  • Open perfume bottle on vanity (dropper slightly crooked)
  • Vintage postcards tucked into mirror edges



c6FLu6b.png




RUrTois.png




43iuUx5.png




j3bfYe0.png
 
Después he tenido una conversación con la IA pidiéndole ayuda para una idea que se me había ocurrido para una serie fotográfica, pero diciéndole que aún no tenía claro el concepto y pidiéndole que me ayudase a definir la idea: presentar a unas estupendas de nuestro tiempo como si hubieran viajado al pasado, conservando sus atuendos y mezclándose a la perfección con el entorno y manteniendo su sensualidad.

Esta ha sido la primera idea que me ha sugerido, y le he dicho que en esa foto lo que ha viajado en el tiempo ha sido el edificio y no la chica... :mparto

oPxTD29.png



WpPV3EB.png



Me ha dicho que lo entendía, y su siguiente sugerencia se ha pasado de frenada... :mparto

U9H5tVQ.png



Para centrar el tiro le he pedido que hagamos que las chicas vayan vestidas como en la época a la que hayan viajado, pero que se note que son actuales por el peinado o el maquillaje, como si fuese una campaña de una firma de moda de gama alta.

Y aquí si que ha centrado el tiro.

ygsuRmq.png


ca8ZCPD.png



Si la pose os parece demasiado distante, siempre podéis pedirle que la queréis mirando al espectador y con una sonrisa cómplice...

3wAvv00.png



gwKYmj9.png



Si le digo que siga con la idea, se me va al renacimento

NkSJVy3.png



81wfDN2.png


A la edad media

Ply4sG4.png



7V3bFiP.png


O al antiguo egipto...

qssvJaB.png



QZ3ieis.png


Le he recordado que esto es para compartir con vosotros, y me ha sugerido una variante...

IX81ikj.png



RzT9rBV.png



Pero le he dicho que quizá es poca variación, y hemos acabado así... :mparto

SOba2v5.png



MoKId79.png
 
Deepseek no es capaz de interpretar el contenido de una imagen, así que si quiero "inspirarme" en una imagen ya existente, tengo que buscarme otra IA que lo haga.

Por ejemplo, a Google Gemini y su modelo Flash 2.5 le puedes pedir que te interprete una imagen como este poster de la película "El ataque de la mujer de 50 pies".

K7sO1nR.png


Si me llevo ese texto directamente a ImageFX, obtengo resultados bastante aceptables (aunque le ha puesto unas patorras que ni a Hulk).

xrWyBP0.png




R7CLSyA.png




He intentado probar otra receta con Grok, que me permite subirle una imagen y le puedo pedir que cree una versión de la misma respetando la imagen y presentando a la protagonista como una mujer gigante y atractiva, que se dedica a recoger coches de la carretera.

O3EmX5v.png



Htgb94a.jpeg



En lugar de un coche, ha interpretado que en sus manos tenía un arma, así que le he pedido que corrija.


Dl2ICHO.png


Y ahí se le ha ido la pinza... :mparto

orBAUwG.jpeg



Si le digo a DeepSeek que le voy a dar esa descripción y que quiero que me cree una nueva imagen más actualizada y refinada, se pone a mi disposición:

7Pp9WN0.png




Y ImageFX cumple con su tarea


xIJTpOU.png



El resultado nos puede convencer más o menos que el anterior, pero entendéis la idea y lo potente del método ¿no?

w7GInPU.png



Una vez tenemos la churrera en marcha, es tan sencillo como:

Elegir una película, como por ejemplo Planeta Prohibido, y descargar su poster de FilmAffinity

KnViAuH.jpeg

Pedirle a Google Gemini que nos analice el poster

Pedirle a DeepSeek que nos cree una versión más modderna y actualizada del texto

Pedirle a Google ImageFX que nos cree la nueva imagen.

MsC0kdY.png



En este caso, yo diría que el original supera a la actualización :mparto

OcGYlZW.png
 
Cabron de chatgpt. Le subo una foto de dos colegas, de digo que me haga una caricatura y no hay manera.

Primero le he pedido estilo mortadelo y filemon, pero no hay manera, que sus políticas lo prohiben


¿Y por qué no pruebas la receta que os compartía ayer?

Por ejemplo, usando esta foto de stock de una pareja:


hhw00Ge.jpeg


Le he dicho a Google Gemini que la describa con precisión y me ha generado esto:

This image shows a head-and-shoulders portrait of a smiling couple against a plain white background.

The man is in the foreground, slightly to the left. He has short brown hair and a beard. He is wearing a blue, white, and red plaid shirt. He is facing forward, looking directly at the camera with a wide, genuine smile, showing his teeth. His arms are crossed over his chest.

The woman is positioned behind the man, her head resting near his right shoulder. She has long, wavy brown hair and is also smiling broadly, looking at the camera. She is wearing a light-colored, long-sleeved top, likely white or cream. Her right arm is around the man's neck, and her left hand is resting on his left arm.

Both individuals appear happy and relaxed, with a warm and joyful expression. The lighting is even, suggesting a studio setting. The white background helps to focus attention entirely on the couple.


Si lo uso tal cual como un prompt en ImageFX, el resultado no es idéntico, pero se aproxima a la realidad.

MpyFAmz.png



Si le digo a DeepSeek que me haga una descripción de ese texto en formato caricatura, me dice esto:

This image is a bright, cheerful cartoon portrait of a smiling couple against a crisp white background, drawn in a lively, animated style.
The man is front and center, slightly to the left, with playful, spiky brown hair and a neatly trimmed beard. He’s wearing a bold, colorful plaid shirt in blue, white, and red, with exaggerated checks for a fun, cartoony look. His arms are crossed in a confident pose, and he’s beaming a huge, toothy grin straight at the viewer, his eyes sparkling with joy.
Peeking over his right shoulder is the woman, her long, bouncy brown hair cascading in loose waves. She’s also grinning widely, her eyes crinkled with happiness. She wears a soft, light-colored long-sleeved top—maybe cream or pastel—with gentle folds to give it a cozy feel. One arm is draped warmly around the man’s neck, while her other hand rests playfully on his arm.
Both characters have exaggerated, expressive smiles and bright, shining eyes, radiating warmth and fun. The clean white backdrop keeps all the focus on their vibrant, cartoonish charm, making them look like they’ve stepped right out of a heartwarming animated series!

Y lo uso directamente como prompt, tengo unas caricaturas

NcZkkNo.png



Otro ejemplo para demostrar que la técnica funciona, y más con personajes conocidos.

Son 3 IAs distintas, pero el proceso completo te lleva un par de minutos, no más.

Partiendo de este retrato de Laurel y Hardy

NnM7upE.png


Gemini interpreta esto:

This image is a colorized portrait of the famous comedy duo Laurel and Hardy, posed against a solid red background.

On the left is Stan Laurel. He is wearing a light brown suit jacket with wide lapels, a white dress shirt, and a red and white polka dot bowtie. A pink or red pocket square is visible in his jacket pocket. He is wearing his signature small black bowler hat positioned squarely on his head. He has a sheepish or slightly awkward smile on his face and is looking off to the left side of the frame with a slightly upward gaze.

On the right is Oliver Hardy. He is larger in stature than Laurel. He is wearing a blue suit jacket, a white dress shirt, and a dark blue tie with white polka dots. His black bowler hat is perched on the back of his head. He has his characteristic expression of frustration or exasperation, with furrowed brows and a downturned mouth, and is looking upwards and slightly to the right.

The two comedians are standing side-by-side, framed from the chest up. The red background provides a stark contrast to their clothing and skin tones.


ImageFX crea esto con ese prompt, que se podría mejorar ajustándolo, pero he preferido dejarlo "tal cual".

vLUQT8r.png


DeepSeek me hace este prompt alternativo para caricatura:

This image is a vibrant, exaggerated cartoon portrait of the legendary comedy duo Laurel and Hardy, popping against a bold, solid red background—like a classic animated poster!

On the left, Stan Laurel is drawn with his iconic lanky frame and goofy charm. He wears a light brown suit jacket with comically wide lapels, a crisp white dress shirt, and a whimsical red-and-white polka-dot bowtie that bounces with energy. A flamboyant pink pocket square pokes out from his jacket like a little flag of silliness. His tiny black bowler hat sits perfectly straight on his head, almost too small, adding to his lovable klutziness. His face has a sheepish, slightly clueless grin, eyes glancing off to the side with an upward gaze, as if he’s just accidentally broken something (again).

On the right, Oliver Hardy looms large, his round belly and double chin exaggerated for maximum cartoonish grumpiness. He’s stuffed into a snug blue suit jacket, his white shirt straining slightly under his wide polka-dotted tie. His black bowler hat perches precariously on the back of his head, as if it might tip off any second. His face is a masterpiece of cartoon frustration—bushy eyebrows scrunched together, cheeks puffed, and mouth turned down in his classic "Well, here’s another nice mess you’ve gotten me into!" expression. His eyes roll dramatically upward, as if begging the heavens for patience.

The two stand side-by-side, their contrasting expressions and body language making them the perfect cartoon comedy pair. The bright red background makes their outfits and exaggerated features pop like a classic Looney Tunes cel, full of energy and timeless humor!

Que ImageFX interpreta bastante bien:

cDHU2Yt.png
 
Sigamos en formato academia, a ver si consigo que os enganchéis alguno más y que esto no parezca una bilbainada como el partido de esta noche...

Yo creo que con este ejemplo igual conseguimos que @Lennoniano o incluso @dawson se animen a crear imágenes.

Un ejemplo de prompt sencillo, en el que le pido a ImageFX una fotografía en blanco y negro de una mujer que gira la cabeza y cuyo pelo, con luz natural y algo de grano digital por usar ISO alta.

Fijaos que no digo nada y la IA me sugiere una mujer joven, atractiva, morena, de rasgos occidentales y mostrando algo de piel.

A black-and-white candid shot of a woman turning away, her hair obscuring her face. Captured in natural light with high ISO grain, evolving a sense of fleeting emotion."

gbDzRL9.png



Con ayuda de DeepSeek, creo una guía de estilo para que las imágenes resultantes tengan una coherencia, y sobre todo para evitar que "huelan a IA", buscando un efecto natural, combinando detalles de iluminación, composición, texturas, etc.


Style Guide

1. Lighting & Atmosphere

  • Prefer natural, diffused, or chiaroscuro lighting (avoid flat or overly even lighting).
  • Use film-like color grading (muted tones, slight fade, or rich contrasts).
  • Embrace grain, texture, and subtle imperfections (e.g., film grain, lens flares, light leaks).

2. Composition & Framing

  • Suggest, don’t show—use shadows, reflections, or obstructions (fabric, objects) to imply form.
  • Close-ups and partial framing (hands, lips, back, silhouettes) work better than full-body shots.
  • Negative space is powerful—let emptiness guide the viewer’s imagination.

3. Textures & Details

  • Skin should look real—avoid plastic or overly smooth AI artifacts.
  • Fabric textures (silk, lace, wet drapery) add tactility.
  • Film emulation (Kodak Portra, Ilford B&W, Fuji Superia) helps avoid digital sterility.

4. Avoiding AI Tells

  • No exaggerated proportions (AI tends to over-smooth or distort anatomy).
  • No generic "dreamy" haze—keep blur intentional (e.g., motion blur, shallow DoF).
  • No hyper-symmetry—imperfections make it feel real.

5. Post-Processing (If Needed)

  • Add light grain or scan lines to mimic film.
  • Slight vignetting or lens distortion for authenticity.
  • Adjust black levels for depth.


Si junto el prompt con la guía de estilo, las imágenes se refinan un poco.

gSi6yrl.png




Puedes cambiar la dirección de la mirada...

0bhtWi8.png



Pensar en rubias en lugar de morenas....

6y4BuOE.png



Incluso en pelirrojas con pecas

e5PYsiV.png



En blanco se ve mal, así que la puedes pedir en color

BqYe3Z7.png



Y ver en color a las rubias...

mt8hUko.png



O a las morenas

3h6ESwy.png



Aunque para esta serie, yo creo que la combinación de blanco y negro y alto contraste funcionan mejor ¿no?

0DNUHqt.png



WTh71wq.png



24HAaJU.png
 
Arriba Pie