How to use ChatGPT’s new image and voice options

ChatGPT ushered in a major upgrade: voice and image functions are coming! OpenAI's latest ChatGPT feature update allows this AI chatbot to have the ability to "see, listen, and speak". Say goodbye to plain text input and enable a richer interactive experience! (Currently, voice and image functions are only available to ChatGPT Plus paid users, but on November 21, voice chat function has been opened to all free users.)

[Related: ChatGPT now has visual, auditory and voice interaction capabilities]

OpenAI's official blog post states that the new feature allows ChatGPT to process voice and image input and generate voice and image output. This means you can use voice to talk to ChatGPT, such as having it tell bedtime stories or creating poetry, or upload images to have ChatGPT analyze the image content, or even generate new images based on your requirements.

Voice conversation:

To experience the voice conversation feature, you need to use the Android or iOS mobile app (not yet supported on the web version). After logging in, click the headphone icon in the lower right corner of the main interface to start the voice conversation. You can choose from five different voices and give a brief audition. After confirming the selection, you can start a voice conversation with ChatGPT.

The dialogue process is as simple and natural as talking to a mobile phone voice assistant. ChatGPT processes your voice input and generates voice replies, often asking questions at the end of the reply to continue the conversation. You can switch topics at any time or click the pause button in the lower left corner to start a new conversation.

If ChatGPT is difficult to recognize your voice or pause, you can try holding the screen to manually enter the voice, similar to the intercom mode, to ensure that the voice input is clear and accurate.

The advantage of voice replies is that you can get a more vivid experience, such as listening to ChatGPT telling bedtime stories or creating themed poems. You can precisely control the subject and tone like you would with text prompts. After ending the voice conversation, click the red and white cross icon to view the conversation record in text format.

Image input and output:

You can now use the web version or mobile app to upload images as a prompt for ChatGPT. The web version user clicks the paperclip icon on the left side of the input box to select the picture on the computer; the application user clicks the picture icon to select the picture from the gallery, or clicks the camera icon to take a new photo (if these icons are not visible, click the " ” button).

You can add prompt words in combination with images, such as asking for the content of the image, or uploading a photo of a leaky faucet and asking for repairs, or displaying ingredients in the refrigerator and seeking cooking suggestions.

In mobile apps, you can click on the image before adding a prompt word and circle it in a specific area of ​​the image with a pen, which can guide ChatGPT to focus on specific parts of the image, which is very useful for troubleshooting or obtaining specific information.

ChatGPT also integrates OpenAI's image generator DALL-E. This means you can ask ChatGPT to generate new images, or use your own image as a tip: You can ask it to generate rolling hilly landscapes, dirty street scenes at night, or cartoon-style renderings of interior scenes. You can also ask it to modify or extend the images you provide.

Same as text prompts, the more specific the description, the better the effect. You can describe the image content, style, colors, and shadows accurately. For example, you can ask for a cartoon field with a foreground well or a realistic portrait of CEO characters in black and white style. If you are not satisfied with the generated image, you can ask for ChatGPT to modify it with further prompts. Click or click on the generated image to find the download option.

The above is the detailed content of How to use ChatGPT’s new image and voice options. For more information, please follow other related articles on the PHP Chinese website!