15 Tips To Help You Create Better Prompts And Generate Higher Quality Images On Dall-e & Midjourney

 

Dall-e is a powerful generative model that can create realistic images from text prompts. But how can you make the most of it and get the best results? Here are 15 tips to help you craft better prompts and generate higher quality and more diverse images.

 

1. DESCRIPTIVE WORDS

Use descriptive and specific words. The more details you provide, the more likely Dall-e will produce an image that matches your vision. For example, instead of "a cat wearing a hat", try "a fluffy orange cat wearing a purple fedora".

2. SEPARATE ATTRIBUTES WITH COMMAS

Use commas to separate different attributes or concepts. This helps Dall-e understand the structure and hierarchy of your prompt. For example, instead of "a blue house with a red roof and a green door", try "a house, blue, with a roof, red, and a door, green".

3. USE MODIFIERS

Use modifiers to adjust the size, shape, position, orientation, or style of the elements in your prompt. For example, instead of "a bird on a branch", try "a small bird, perched on a curved branch, facing left".

4. NEGATIONS

Use negations to exclude unwanted elements or features from your prompt. For example, instead of "A house abandoned by humans", try "A house, abandoned".

5. USE SYNONYMS

Use synonyms or alternative words to explore different variations or interpretations of your prompt. For example, instead of "a flower vase", try "a floral arrangement" or "a bouquet".

6. USE HYPHENS

Use hyphens to combine words that describe a single concept or entity. This helps Dall-e treat them as one unit and avoid confusion. For example, instead of "lightning man ", try "a lightning-man".

7. USE PARENTHESES

Use parentheses to add clarifications or explanations to your prompt. This helps Dall-e understand the context or meaning of your prompt better. For example, instead of "Morgoth", try "Morgoth (Tolkien Middle Earth)".

8. USE COLONS

Use colons to introduce examples or specifications to your prompt. This helps Dall-e narrow down the scope or range of your prompt. For example, instead of "a house", try " a house: Pool, garage, balcony, apple tree with a swing, 3 floors ".

9. USE QUOTATION MARKS

Use quotation marks to indicate literal or exact words that you want Dall-e to include in the image. This helps Dall-e preserve the spelling or formatting of your prompt. For example, instead of "a logo for Road to Nowhere", try "a logo for “Road To NoWhere” ".

10. EMOJI'S

Use emojis to add emotions or expressions to your prompt. This helps Dall-e convey the mood or tone of your prompt better. For example, instead of "Two guys and a ghost", try "🧔👻🧔".

11. USE SPECIFIC NUMBERS

Use numbers to specify the quantity or order of the elements in your prompt. This helps Dall-e create images that are consistent with your expectations. For example, instead of "knights", try "three knights, left black knight, right grey knight, middle green knight on his knees".

12. USE PREPOSOTIONS 

Use prepositions to indicate the spatial relationships between the elements in your prompt. This helps Dall-e arrange the elements in a logical and realistic way. For example, instead of "A werewolf on a table and a terrified human", try "a werewolf on top of a table, a man hiding underneath the table".

13. USE CONJUNCTIONS 

Use conjunctions to connect or contrast different elements or features in your prompt. This helps Dall-e create images that are coherent and complex. For example, instead of "a dog and a cat", try "a dog and a cat that are friends" , or  "Frost and fire being enemies".

14. USE QUESTIONS TO CHALLENGE

Use questions to challenge Dall-e to generate images that are creative and surprising. This helps you discover new possibilities and ideas that you might not have thought of before. For example, instead of "a unicorn", try "what if unicorns had wings?" or "What if Mars was populated?”

15. USE FEEDBACK TO REFINE

Use feedback to refine or improve your prompt and image. This helps you iteratively adjust and optimize your prompt and image until you are satisfied with the result. For example, after generating an image from your initial prompt, you can add or remove words, change the order or structure of words, or use different modifiers, negations, synonyms, hyphens, parentheses, colons, quotation marks, emojis, numbers, prepositions, conjunctions, or questions to modify your prompt and generate a new image.