What AI Image Generator Should YOU Be Using??

The video offers a comprehensive review and comparison of various AI image generators, including Mid Journey, Dolly 3, Firefly Image 2, Stable Diffusion XL, Google's generative search experience, and Idiogram. The evaluation is based on criteria such as accuracy, creativity, realism, illustration quality, logo and vector creation, texture and background tiling, text within images, censorship, usability, and pricing. The analysis includes testing each tool's adherence to prompts, customization options, and output quality. Mid Journey excels in creativity and realism, Dolly 3 in accuracy, and Google in logo creation. Leonardo stands out for its low censorship and versatility, making it the best overall value. The video concludes with recommendations on which tool to use for specific use cases and acknowledges that while all generators have their strengths, Leonardo offers a well-rounded performance across categories.

Takeaways

• 🤖 There are numerous AI image generators available, each with its strengths in specific use cases such as Midjourney, Dolly 3, Firefly Image 2, Stable Diffusion XL, Google's generative search, and Idiogram.
• 🎨 The evaluation of these tools is based on criteria like accuracy, creativity, realism, ability to create illustrations, logos, vectors, textures, background tiling, text within images, censorship, usability, and pricing.
• 📉 Midjourney scored high in creativity and realism but had usability drawbacks and was not free to use.
• 🏆 Dolly 3 was accurate but had limitations like censorship and inability to perform certain tasks like texture background creation, making it less versatile.
• 📈 Firefly Image 2 was good for illustrations and logos but struggled with realism and was not as customizable.
• 🔍 Google's image generator excelled in logo creation and was free to use, but it had usability quirks and potential censorship issues.
• 🌐 Idiogram was free, uncensored, and capable with text in images, but it was not as accurate or creative as some competitors.
• 💯 Leonardo offered a high degree of customization, was the least censored, and performed well across most categories except for text in images, making it a good overall value.
• 📋 The ability to generate tilable textures was a significant differentiator, with Midjourney, Stable Diffusion (via Leonardo), and Firefly 2 successfully creating seamless tiles.
• 💬 Text within images is a newer feature, with Dolly 3, Google, and Idiogram performing well in this area.
• 🚫 Censorship varied among the tools, with some like Firefly 2 and Dolly 3 inside Bing's Image Creator showing more restrictions on generating certain IP or celebrity images.
• 💰 Pricing played a role in the overall value, with free options like Idiogram, Dolly 3 in Bing's Image Creator, and Google being more accessible, while paid options like Midjourney and Dolly 3 in chat GPT were less so.

Q & A

What are the key factors considered in evaluating AI image generators?

The key factors considered include accuracy, creativity, realism, ability to create illustrations, logos, vectors, textures, background tiling, handling of text within images, censorship, usability of the interface, and pricing.

Which AI image generator was found to be the most accurate according to the video?

Dolly 3, particularly when used within Bing's image Creator, was found to be the most accurate AI image generator in adhering to the given prompts.

How does the video evaluate the creativity of AI image generators?

The video evaluates creativity by providing minimal context in the prompts and subjectively judging which tools produce more creative and less boring images.

What was the general consensus on the realism of the images generated by the AI tools?

Mid Journey Raw and Firefly 2 were considered the best in terms of realism, while Google's generative search experience and idiogram were at the lower end of the spectrum.

How does the video address the issue of censorship in AI image generators?

The video tests the AI tools' responses to prompts involving celebrities and well-known IP characters to see if they can generate images without censorship or restrictions.

Which AI image generator was rated as the most user-friendly?

Leonardo was rated as the most user-friendly due to its extensive customization options and intuitive interface.

What is the main advantage of using Dolly 3 within chat GPT?

The main advantage is its ability to easily alter the image through natural language interactions, making it highly customizable and user-friendly.

How does the video handle the evaluation of text within images generated by AI tools?

The video uses prompts that include text, such as 'a penguin holding a wooden sign that says subscribe to Matt wolf', and evaluates the AI's ability to accurately render the text.

What is the significance of testing tiling in the context of textures and backgrounds?

Tiling is significant as it tests the AI's ability to create images that can seamlessly repeat, which is important for creating patterns or textures that can be used as backgrounds.

Which AI image generator is currently free to use and does not censor content?

Idiogram is mentioned as being free to use and not appearing to censor content, making it a good option for users looking for these specific features.

Based on the video, which AI image generator would you recommend for someone on a budget?

For someone on a budget, the video recommends Dolly 3 in Bing's image Creator, Google's generative search experience, and idiogram, as they are currently free to use.

Outlines

00:00

🤖 Overview of AI Image Generators

The video discusses various AI image generators, including Mid Journey, Dolly 3, Firefly Image 2, Stable Diffusion XL, Google's generative search, and Idiogram. It aims to determine the best tool for specific use cases by evaluating them on criteria like accuracy, creativity, realism, usability, and pricing.

05:02

🎨 Testing Accuracy and Creativity

The video script details a test of accuracy and creativity for different AI image generators. It uses specific prompts to assess how well each tool adheres to the given instructions and rates them accordingly. Mid Journey and Dolly 3 perform well, with Dolly 3 showing a slight edge in accuracy.

10:03

🌟 Realism in AI-Generated Images

The paragraph focuses on the realism of images generated by the AI tools. It uses a specific prompt of a couple holding hands in front of the Eiffel Tower. Mid Journey, particularly in raw style, and Firefly 2 are noted for their realistic outputs, while others like Google struggle with realism.

15:05

🎭 Evaluating Illustration Styles

The script evaluates the illustration capabilities of the AI image generators. It tests each tool's ability to create anime-style images. Mid Journey, especially in NII mode, is praised for its contrast and quality, while other tools like Leonardo and Firefly 2 also perform well.

20:06

🏢 Logos and Vector Graphics

The video examines the effectiveness of each tool in generating logos and vector graphics. It uses a prompt for a simple flat vector image of a wolf. Google stands out for logo creation, followed by Mid Journey and Firefly 2. Leonardo, while capable, does not specialize in simple logos.

25:06

🧩 Tiling and Background Textures

The script tests the ability of the tools to create tiling background textures. Mid Journey excels in this category, with both regular and raw styles capable of seamless tiling. Other tools like Dolly 3 and Bing Image Creator struggle with proper tiling.

30:07

📝 Text within Images

The paragraph explores the newer feature of including text within AI-generated images. Dolly 3, Google, and Idiogram are highlighted for their ability to incorporate text effectively, while Mid Journey and Leonardo struggle with text accuracy.

35:10

🚫 Censorship and IP Restrictions

The video investigates censorship and the use of intellectual property in AI image generation. It finds that while some tools like Mid Journey and Idiogram generate images with celebrity faces and IP without restrictions, others like Dolly 3 and Firefly 2 are more restrictive.

40:12

🛠️ Usability and Customizability

The script discusses the usability of the different AI image generators. It praises Leonardo for its customizability and Firefly 2 for its intuitive interface. Google's generative search experience is noted for its usability quirks, and Mid Journey is criticized for being less user-friendly due to its Discord-based interface.

45:13

💰 Pricing and Value Assessment

The final paragraph compares the pricing of the AI image generators. It notes that while some tools like Dolly 3 (Bing Image Creator) and Idiogram are free, others like Mid Journey and Leonardo offer a mix of free and paid options. Google's service is completely free, but with potential usability drawbacks.

🏆 Conclusion and Recommendations

The video concludes with a summary of the findings, identifying Leonardo as the best overall value, followed by a tie between Mid Journey and Idiogram. It provides recommendations on which tool to use based on specific needs, such as creativity, realism, or cost.

Keywords

AI Image Generators

AI Image Generators are software applications that use artificial intelligence algorithms to create images based on textual prompts or other data inputs. They are a core focus of the video, as the host compares different generators on various criteria such as accuracy, creativity, and realism. Examples from the script include Mid Journey, Dolly 3, and Google's generative search experience.

Prompt Adherence

Prompt adherence refers to how closely an AI image generator follows the instructions given in a textual prompt to create an image. It is a critical aspect evaluated in the video, with the host testing each tool's ability to generate images that match specific and complex prompts. For instance, the host assesses if the generated images accurately depict a 'green bus floating in space' or a 'sitting artist with a bucket hat painting a canvas of a three-headed monster'.

Creativity

In the context of the video, creativity pertains to the originality and uniqueness of the images produced by the AI generators when given vague or broad prompts. The host examines how each tool interprets simple prompts like 'beautiful, creative epic RGB image' and 'beauty' to judge their creative output. Mid Journey and Leonardo are highlighted for their particularly creative responses.

Realism

Realism is evaluated by how lifelike and true-to-life the AI-generated images appear. The video tests this by using prompts that aim to produce images resembling real-world scenarios, such as 'a couple holding hands in front of the Eiffel Tower.' The host grades each tool on its ability to create convincing and detailed images that could potentially be mistaken for photographs.

Illustrations

Illustrations refer to the AI-generated images that have a drawn or painted style, often used for artistic expression or to convey a concept in a more stylized manner. The video discusses the effectiveness of each tool in creating illustration-style images, particularly noting the performance of Mid Journey, Dolly 3, and Firefly Image 2 with prompts like 'anime girl with braids in the neon streets of Tokyo'.

Logos and Vectors

Logos and vectors are specific types of graphic design where logos are symbols representing a brand or company, and vectors are images composed of points, lines, and curves that can be resized without losing quality. The video assesses how well each AI tool can generate simple, flat vector images of logos, as exemplified by the prompt 'simple flat vector image logo of a wolf on a white background.'

Textures and Backgrounds

Textures and backgrounds involve the AI's ability to create images that can be used as repeating patterns or seamless backgrounds. The host tests this by asking the tools to generate 'colorful circuitry' images that should tile without visible seams. Mid Journey, Stable Diffusion XL, and Firefly Image 2 successfully create tilable images, whereas others like Dolly 3 and Google struggle with this task.

Text in Images

The ability to include accurate text within the generated images is a feature assessed in the video. The host uses prompts that require the inclusion of specific words or phrases in the images, such as 'a penguin holding a wooden sign that says subscribe to Matt wolf.' Dolly 3, Google, and Idiogram are noted for their capability to handle text in images, albeit with varying levels of accuracy.

Censorship

Censorship in the context of AI image generators refers to the limitations or restrictions placed on the content that can be generated, often due to copyright, trademark, or content policy restrictions. The video discusses how some tools like Dolly 3 and Firefly Image 2 may refuse to generate certain prompts involving celebrities or well-known characters, while others like Idiogram and Stable Diffusion XL appear to have fewer restrictions.

Usability

Usability focuses on how easy it is to use each AI image generator, including the interface, the complexity of commands, and the available features for customizing the generated images. The video compares the user experience across different platforms, with Leonardo receiving high marks for its customizable options and simple prompting, while Google's generative search experience is critiqued for its confusing interface.

Pricing

Pricing refers to the cost associated with using the AI image generators. The video evaluates each tool's pricing model, from free options with limitations to paid subscriptions that offer more features or higher usage limits. The host discusses the affordability and value for money, with tools like Dolly 3 (within Bing's Image Creator) and Idiogram being praised for their free tiers.

Highlights

Mid Journey is considered by many as the best AI image generator, but there are several competitors like Dolly 3 and Firefly Image 2.

Dolly 3 is sometimes referred to as the 'Mid Journey killer' due to its capabilities.

Firefly Image 2 is gaining popularity for being as good as Mid Journey and Dolly 3.

Stable Diffusion XL is praised for its high level of customizability.

Google has integrated its own image generator into the generative search experience.

Idiogram was at the top of the AI art world a month ago for generating text inside images.

The video aims to determine the best AI image generation tool for specific use cases.

Tools are graded on accuracy, creativity, realism, and other factors like usability and pricing.

Mid Journey's raw style adheres more closely to the prompt with less of its own style influence.

Dolly 3 achieved a high score of 9 out of 10 for accuracy, closely following the given prompts.

Stable Diffusion XL, used through Leonardo, scored a 6.5 for accuracy, performing better than Mid Journey in some aspects.

Firefly Image 2 and Google's generative search experience both scored 6.5 and 7.2 respectively for accuracy.

Idiogram scored a 6.7 for accuracy, showing potential but requiring more complex prompts for better results.

Mid Journey excels in creativity, especially with its raw style, which scored a 9 out of 10.

Dolly 3 and Leonardo also showed strong creativity, but Firefly Image 2 and Google lagged behind.

In terms of realism, Mid Journey raw was the most convincing, scoring an 8.5.

For illustration style images, Mid Journey, Leonardo, and Firefly 2 performed well, but Google and Dolly 3 were less effective.

When creating logos and vectors, Google surprisingly outperformed other tools, followed by Mid Journey and Adobe Firefly 2.

Mid Journey and Stable Diffusion XL were able to create tilable textures, while others faced challenges.

Dolly 3, Google, and Idiogram were capable of generating text within images effectively.

Censorship varied among the tools, with Idiogram and Stable Diffusion XL being the least censored.

Usability was highest with Leonardo and Firefly 2, while Mid Journey and Google had more usability issues.

In terms of pricing, free options like Dolly 3 (Bing Image Creator), Google, and Idiogram offer cost-effective alternatives.

Leonardo is considered the best value overall, followed by a tie between Mid Journey and Idiogram.