Ücretsiz Dall-e 3 Kullanarak Prompt Girmenin Mantığını Anlattım

Erhan Meydan

14 Oct 202312:18

Summary

TLDRThe video explains how to create visuals using AI tools like Bing Image Creator. It provides tips on writing effective prompts, starting with 3-7 words and adding descriptive adjectives. It advises using concrete words over abstract ones, avoiding allegories that AI may not grasp. The video then covers customizing images, like photograph settings, art styles and genres. It introduces helpful sites for art inspiration and notes that you own full rights to generated images. Overall, it aims to develop skills for prompt writing to get quality AI-generated visuals.

Takeaways

😊 You can generate visuals with AI tools like Bing Image Creator for free
🖼️ Start with 3-7 word prompts and add descriptive adjectives to get better results
🎨 Use concrete nouns instead of abstract concepts or allegories
🔍 Be as clear and direct as possible with your prompt wording
📸 You can specify photo settings like camera type, lens, lighting etc. in prompts
👩‍🎨 Apply different rendering techniques like ray tracing or art styles
🎥 Use movies, TV shows or art movements as prompts for desired styles
✏️ Generate sketches, drawings or paintings with details like pencil, charcoal etc.
🔎 Research and utilize aesthetic styles like Art Deco, Steam Punk etc.
💡 Experiment with different prompts and techniques to improve your skills

Q & A

What tool does the narrator use to explain generating images with AI?
-The narrator uses Bing Image Creator to explain generating images with AI because it is completely free and uses the DALL-E 3 AI.
Why does the narrator recommend entering prompts in English if possible?
-The narrator recommends entering prompts in English because the AI models have been trained mostly on English data, so prompts entered directly in English lead to better results compared to using translation.
What are some key tips the narrator offers for writing good prompts?
-Key prompt writing tips include using 3-7 words initially, adding descriptive adjectives, using concrete words instead of abstract ones when possible, avoiding metaphors the AI may not grasp, and clearly stating what you want to see.
How can you customize the look of generated images?
-You can add details like camera settings, time periods, art movements and styles, specific artists, or movie/TV visuals to make the generated image match a desired aesthetic.
What site does the narrator recommend for researching art styles and artists?
-The narrator recommends using Arts & Culture by Google to research different art movements, materials, and explore works by over 13,000 artists.
What file formats can you export the AI generated images as?
-The narrator states that you can download the AI generated images and edit them further in programs like Photoshop, indicating common image formats like JPG or PNG would work.
Does the video cover generating 3D images?
-Yes, the narrator gives the example prompt of visualizing a broken glass on a table rendered in the Ray Tracing style to generate a 3D-like image.
What key skills does the narrator say are important for prompt engineers?
-The narrator says prompt engineers need to be skilled at quickly entering effective prompts and getting good results without using up too much AI credit.
Can viewers freely use and share the images they generate?
-Yes, the narrator states there are no copyrights on the AI generated images, so viewers can use and share the images they create however they want.
Where can people ask the narrator follow-up questions?
-The narrator invites viewers to share their generated images and prompts in the video comments section if they have any follow-up questions or comments.

Outlines

00:00

😊 Introduction to writing prompts for AI image generation

The paragraph introduces prompt writing for AI image generation tools like DALL-E and Stable Diffusion. It explains why Bing Image Creator is used as an example, being completely free to use. The goal is to explain the logic behind writing effective prompts based on personal experience.

05:02

💡 Tips for writing better prompts with concrete words

The paragraph provides tips for writing better prompts, advising the use of concrete words to convey the intended image as clearly as possible. It notes that abstract concepts are harder for AI to interpret. The recommendation is to start with the most concrete description, then gradually add more details.

10:03

🎨 Customizing the image style and aesthetics

The paragraph explains how prompts can be customized to achieve different image styles and aesthetics. Examples are given for specifying the desired time period, art movement like impressionism, reference artists like Picasso, and visual media like The Matrix. Tools like Google Arts & Culture can provide inspiration.

Mindmap

Keywords

💡prompt

Prompt refers to the text instructions entered by the user to guide the AI image generation process. As explained in the video, effective prompts should be 3-7 words long initially, use concrete descriptive words, avoid metaphors, and clearly state what you want the AI to generate. Prompts are a core part of communicating desired outputs to the AI.

💡AI image creator

AI image creators like DALL-E, Midjourney and Stable Diffusion are the main tools used to generate images based on text prompts. The video focuses on the Bing Image Creator specifically, which uses the DALL-E 3 AI and provides free credits to experiment with.

💡Bing Image Creator

Bing Image Creator is the free AI image generation tool from Microsoft, built using DALL-E 3 AI. It allows users to create 24 images per day without any costs. The video uses this for demonstrating prompt creation techniques.

💡adjectives

Adjectives help describe the attributes of an object, which is useful for crafting more detailed prompts. Using color, texture, lighting and other descriptive terms allows users to better communicate what they want the AI to depict.

💡concreteness

Using concrete words helps the AI grasp key details rather than interpret abstract ideas differently than intended. Concrete words like "round table" lead to better results than vague terms like "happiness".

💡post-processing

The generated images can be downloaded and further edited in tools like Photoshop, which the video demonstrates to expand a scene and use AI to fill additional space. This post-processing and iteration can improve results.

💡art styles

Users can prompt the AI to generate images mimicking different art styles like impressionism, surrealism, or looking like specific artists like Picasso and Van Gogh. This leverages the AI's training on recognizing diverse artistic techniques.

💡image ownership

An important note is that images generated belong fully to the user, without any copyright issues to worry about. This enables unrestricted usage of the AI outputs.

💡prompt engineer

The video notes that skills in quickly ideating and trying different prompts to get good results with fewer credits is key. People good at this prompt engineering role are especially valuable.

💡iteration

Core message is that prompt writing involves an iterative process of trying different wording and styles, analyzing the results, and refining the instructions to get closer to the desired outcome.

Highlights

Explaining the logic of entering prompts to create visuals with AI

Using Bing Image Creator for examples because it's free and uses DALL-E 3 AI

Start with 3-7 words in prompts for best results

Use lots of descriptive adjectives when entering prompts

Avoid metaphors, be as concrete and literal as possible

Can specify photograph details like camera brand, lens specs, lighting

Lists of aesthetic styles and art movements to easily reference

Google Arts & Culture site to browse artists and art styles

Can create 3D renders and use different software rendering techniques

All generated images are royalty-free to use as desired