How To Make Photorealistic Images In Fooocus

In this tutorial, the presenter shares a straightforward method to create photorealistic images using the software Focus. The process involves ensuring the Advanced tab is selected and choosing between Focus V2, Focus enhance sharp, and photograph under Styles. The presenter uses the Juggernaut XL as the base model and Realistic Vision version 6 as the refiner, which is a stable diffusion 1.5 model. The video demonstrates the difference between using the base model with an SDXL refiner and using Realistic Vision as the refiner, highlighting a more photorealistic outcome with the latter. The presenter also discusses the benefits of this method, such as avoiding common issues like deformed limbs and double heads at higher resolutions. Additionally, the presenter provides information on where to obtain the necessary models, such as on Civit AI, and suggests experimenting with different base and refiner models to achieve desired results. The video concludes with a demonstration of the method's effectiveness on various subjects, including people and cars, and a brief mention of using face swap technology.

Takeaways

• 🖼️ To achieve photorealistic images in Focus, ensure the Advanced tab is checked for additional options.
• 🎨 Under Styles, select Focus V2, Focus enhance sharp, or photograph for different looks. You can also experiment with cinematic.
• 🤖 For the base model, Juggernaut XL is used, which is crucial for the photorealistic outcome.
• 🔍 The refiner should be set to Realistic Vision version 6, a stable diffusion 1.5 model, for better photorealism.
• 📏 Using SD 1.5 models as a refiner allows for higher resolution images without common issues like deformed limbs or double heads.
• 📦 Download and use additional models like Realistic Vision version 6 or Epic Photogasm from Civit AI to enhance your results.
• 🚀 Turbo models may not work well on Extreme Speed, so it's recommended to keep the speed setting as is until further optimization.
• ⚖️ Experiment with different base models and refiner models to find the best combination for your desired outcome.
• 🚫 Keep negative prompts minimal, focusing on avoiding ugly, deformed, text, and watermarks.
• 🌟 This method works well for half body and full body shots, producing coherent and realistic results.
• 🧍‍♂️ It can also be used for face swapping, getting closer to the user's likeness with fine-tuning of settings.
• 🌐 There are multiple combinations possible with different models, offering flexibility for various photorealistic applications.

Q & A

What is the easiest way to get photorealistic images in Focus?

The easiest way to get photorealistic images in Focus is by ensuring the Advanced tab is checked, selecting a suitable style under the Styles option, choosing the right model, and using a refiner such as Realistic Vision version 6 with a stable diffusion 1.5 model.

How does the method described in the transcript differ from using a stable diffusion 1.5 model as the base model?

Using the method described allows for higher resolution images while still benefiting from fine-tuned SD 1.5 models. Placing the stable diffusion 1.5 model directly as the base model doesn't work because the pipeline is an SDXL pipeline, which requires a different approach.

What are some of the styles available under the Styles option in Focus?

Some of the styles available include Focus V2, Focus enhance sharp, and photograph. Occasionally, cinematic style is also used, offering a slightly different look.

What is the role of the refiner in the process of generating photorealistic images?

The refiner, such as Realistic Vision version 6, is crucial as it enhances the image to achieve a more photorealistic look. It works in conjunction with the base model to produce the desired outcome.

How does the use of SDXL offset Laura with a weight of 0.5 contribute to the image generation process?

The SDXL offset Laura with a weight of 0.5 is used to fine-tune the image generation process, providing a balance that works best for achieving photorealistic results according to the transcript.

What are some of the advantages of using this method over generating a stable diffusion 1.5 model on a platform like Auto1111?

This method allows for higher resolution images without the risk of deformations such as double heads or limbs, which can occur when generating stable diffusion 1.5 models at higher resolutions on some platforms.

Where can one find and download the models needed for this process?

The models, such as Realistic Vision version 6, can be found and downloaded from Civit AI.

What are some other photorealistic models mentioned in the transcript?

Other photorealistic models mentioned include Epic Photo Gasm and Epic Realism, both developed by the same developer.

How does the Dream Shaper XL Turbo model compare to the Juggernaut XL in terms of performance?

The Dream Shaper XL Turbo model is not recommended to be used on Extreme Speed, and as of the knowledge in the transcript, it is not yet fully optimized, but it is expected to improve soon.

What are the negative prompts used in the process and why are they minimal?

The negative prompts used are 'ugly', 'deformed', 'text', and 'Watermark'. They are minimal because too many negative prompts can interfere with the generation of a high-quality, photorealistic image.

How effective is this method for generating images of people and objects?

The method is highly effective for generating images of people, as evidenced by the examples provided. It also works well for objects like cars, achieving a general photorealistic quality.

How close did the method get in replicating the presenter's likeness using face swap?

While not an exact match, the method provided a closer representation of the presenter's likeness, particularly resembling a younger version of the presenter, although not as muscular as depicted.

Outlines

00:00

🖼️ Photorealistic Image Generation with Focus

The video script introduces a method for creating photorealistic images using the Focus platform. The presenter begins by ensuring the Advanced tab is checked for options and suggests selecting 'Focus V2', 'Focus enhance sharp', or 'photograph' under Styles. The use of the 'Juggernaut XL' as a base model is highlighted, with 'realistic Vision version 6' as the refiner, which is a stable diffusion 1.5 model. The script explains that using this refiner allows for higher resolution images without the common issues associated with direct use of stable diffusion 1.5 models. The presenter also demonstrates the difference between using the base model with the refiner set to 8 versus using realistic Vision as the refiner for a more photorealistic look. The benefits of this method are showcased through examples, and the presenter provides guidance on where to obtain the necessary models from Civit AI. Additionally, the script touches on the use of negative prompts and the compatibility of the method with half-body and full-body shots.

05:01

🚗 Photorealism in Various Subjects Including Cars

The second paragraph discusses the versatility of the photorealistic method introduced, highlighting its effectiveness not only for human subjects but also for objects such as cars. The presenter shares their personal experience with face swapping using Focus and how the method has improved the results, making them closer to their actual appearance, albeit with some differences. The paragraph also mentions the potential for further refining the settings to achieve an even closer likeness. The presenter encourages viewers to watch a previous video on face swapping for additional context and concludes with a friendly sign-off, promising to see the audience in the next video.

Keywords

Photorealistic Images

Photorealistic images refer to digital or generated images that closely resemble real-life photographs in terms of lighting, texture, and detail. In the video, the speaker discusses a method to create such images using the software Focus, which is significant as it allows users to achieve high-quality visuals that are almost indistinguishable from actual photographs.

Focus

Focus is a software platform used for generating images, often with a focus on creating realistic and detailed visuals. The video's theme revolves around using Focus to produce photorealistic images, demonstrating the capabilities of the software and how users can leverage its features to enhance their creative work.

Advanced Tab

The Advanced Tab within the Focus software is a feature that provides users with more control over the image generation process. It is mentioned as the first step in the process of creating photorealistic images, indicating its importance in accessing the necessary tools and options for achieving the desired outcome.

Styles

In the context of the Focus software, Styles refer to different visual presets or settings that can be applied to the image generation process. The video mentions 'Focus V2', 'Focus enhance sharp', and 'photograph' as examples of styles, which can influence the final look of the generated images.

Juggernaut XL

Juggernaut XL is mentioned as the base model used in the video for generating images. It is a specific configuration or setting within the Focus software that serves as the foundation for creating the image, to which additional refinements and enhancements are applied.

Realistic Vision

Realistic Vision is a refiner model used within the Focus software to enhance the generated images. The video emphasizes its use in achieving a more photorealistic look, suggesting that it plays a crucial role in the image refinement process.

Stable Diffusion 1.5

Stable Diffusion 1.5 is a model mentioned in the context of being used as a refiner in the Focus software. It is noted for its ability to produce high-resolution images without common artifacts like deformed limbs or double heads, which is a significant advantage when generating detailed and realistic images.

SDXL Offset Laura

SDXL Offset Laura is an option within the Focus software that can be adjusted to fine-tune the image generation process. The video mentions a weight of 0.5 for this setting, indicating that it is a parameter that users can experiment with to achieve different visual effects.

Resolution

Resolution in the context of the video refers to the level of detail an image can display, measured in pixels. The speaker discusses the advantage of using the described method in Focus to achieve higher resolution images without the common defects associated with increased pixel density.

Civit AI

Civit AI is mentioned as a source for obtaining the models used in the Focus software. It is an online platform where users can download various models, such as Realistic Vision version 6, to enhance their image generation capabilities within Focus.

Negative Prompts

Negative prompts are instructions given to the Focus software to avoid certain unwanted features or artifacts in the generated images. The video mentions 'ugly', 'deformed', 'text', and 'Watermark' as examples of negative prompts, which help in guiding the software to produce images that are free from these undesirable elements.

Highlights

The easiest way to get photorealistic images every time in Focus is by using a specific method involving the Advanced tab and certain styles.

Under Styles, choose between Focus V2, Focus enhance sharp, and photograph for different looks.

Occasionally, switching between cinematic and photograph can yield slightly different results.

Using Juggernaut XL as the base model is key to achieving the desired photorealistic effect.

Realistic Vision version 6 is a stable diffusion 1.5 model used as the refiner for enhanced photorealism.

Focus is primarily an SDXL platform, but this method allows for the results of a stable diffusion 1.5 model.

SDXL offset Laura with a weight of 0.5 is recommended for optimal results, though other luras can be experimented with.

Unchecking random ensures the same seed is used for consistent results.

Comparing the results of the current setting with SDXL base and refiner at 8 reveals different styles.

The base SDXL and refiner tend to produce a hyper-realistic look, while Realistic Vision as the refiner offers a more photorealistic appearance.

Generating a stable diffusion 1.5 model on automatic UI may result in deformations, but this method avoids such issues.

Higher resolution images are possible with this method while still benefiting from fine-tuned SD 1.5 models.

Civit AI is a source for obtaining models like Realistic Vision version 6 and other photorealistic models.

Experimenting with different base models and refiner models can yield a variety of photorealistic results.

The method works well for half body and full body shots, producing great results for eyes and teeth.

Negative prompts should be minimal, focusing on avoiding ugly, deformed, text, and watermark.

The method can handle a wide range of subjects, including people and cars, with high photorealism.

Using face swap with this method can help achieve a closer likeness to the original subject.

The Dream Shaper XL Turbo is used as the base model for another example, with the refiner set to Epic Realism at 4.

Extreme Speed may not work well with the turbo models, so it's recommended to leave it on speed.