Midjourney Office Hours (23rd Oct 2024)
Check out Midjourney's new image editor and retexturing feature
Dear readers,
Here is the summary of the latest Midjourney Office Hours:
The image editor and retexturing feature will be released today. They will be available within hours of Midjourney Office Hours.
New features will be made available to users who meet one of these criteria:
Who have already generated 10K images
Already subscribed for the past 12 consecutive months
Who have yearly subscriptions
Retexturing works well for 3D shapes without texture and can be applied to architectural images. It works with photographs and 3D renders. It can also change the lighting.
The Editor is like a new product. It can mask everything except the face in order to change the person's appearance. It's like doing a photo shoot and looking professional.
Here are the things that were discussed during Office Hours: (Details about the new Editor and Retexturing features will only be known after the release, so please check out the official announcement):
(Update few hours after the Office Hours) : The feature has been released. Check out the Edit menu on the website.
Image editor works like inpainting. It should work if the change makes sense within the image. It will resist if you attempt to change an element in the image to something weird. If it works, it will work extremely well; if it does not, it can be quite frustrating. The developers plan to improve it over time after the launch.
Editor jobs will be set to private by default to avoid unnecessary photo sharing. It has a separate organizational area.
The Editor's next phase will include the development of additional selection tools for the depth map.
Following the launch of the web editor and retexturing, there will be multiple personalization profiles and faster personalization initiation.
The personalization model and personalized Explore page on the Midjourney website should be available next month.
Personalization aims to improve the core of the creative process.
Other features that will be released or developed include batch 8 generation mode, a feature that allows users to train the model on the fly, and Style Personalization (like a Super Sref/Sref Explore/Hardcore Style).
The video feature is progressing well. Currently, there is little technological risk. It is now on par with everyone else , but Midjourney wants to improve it further. It should be the ideal balance of quality, features, and speed.
The earliest possible timeline for the first Midjourney V7 model is about the next two months.
V7 completed the first phase of training. Going to the next phase soon. However, it still requires extensive testing.
There will be a quick follow-up to the V7 release. This means that V7.1 will arrive shortly after that to update its data.
In theory, V7 should outperform in all areas. However, this can only be confirmed once the model is ready.
Multiple consistent characters will be worked on after V7.
The Storytelling feature is an experimental and unpolished product. It may have a front end that can accommodate one or more characters.
There are several ongoing secret projects that may be revealed in the next six months. They are mind-blowingly cool. They tell the story of humanity's future, of which we want to be a part. Some of them will be usable soon, while others will take years, but Midjourney will share their long-term project development progress.
David said that in the future, we might not think about image or video models. It will be a real-time, open-world simulation. Everything that is being built today is the foundation for the future.
An example of retexture from Geeky Animals
Subscribe to the Geeky Curiosity newsletter for free by October 25th, 2024, and you will receive two high-quality Halloween theme Sref codes directly into your email inbox the following day!
Free Geeky Animals' Sref code: 198697906
This Sref code was used to generate the cover image for this article, along with multi-step process prompting tweaks.
