How do I generate a photo with ChatGPT?

1. Generation via description

You can enter a description like this:

Generate a photo in landscape format with a landscape image showing the flags of the USA, the UK and Germany. The flags themselves should be aligned so that they can be seen verti­cally in the middle 300 pixels.

Problems

As you can see, ChatGPT ignores or misin­ter­pretes the second sentence. The generated flags are not in the vertical middle of the generated picture.

A second problem are generated texts, because ChatGPT has 2 restric­tions with that:

• It only generates texts in englisch language
• The resolution of the text is not good — as you can see in examples in the media library.

Why did I ask for vertical middle?

As you can see I wanted to use the generated picture (unchanged) for the blog overview and the header picture in this single article. For the display of the header picture will be used the middle part on a dektop computer in this case.

And the flags should have been shown fully in the header — instead of showing them cut off at the top as it is now on wide monitors.

2. Generation after text generation

In case you have generated a text article in the first step, you can simple tell ChatGPT to generate a photo for the article like this:

Create landscape format photo for the article

Problem

You can enter short sentences. But you will see this can result in misun­derstan­dings, i.e.

Create landscape photo for the article

can create a photo of a landscape instead of a photo in landscape format. This is a language problem. In german language you do not have that problem in this case, because the used words (Querformat or Landschaft) are fully different.