​

So I want to drop the the "t" from AiTuts because it's cleaner. Do you like it or hate it?

eg.: Welcome to Aituts, the only AI newsletter for creative professionals.

What we've got today:

• The first real AI-generated TV show trailer?
• 5 more really cool things about Midjourney V6
• Artists are suing Midjourney

HEADLINE

The first real AI-generated TV show trailer?

The German production company PANTALEON Films has asked freelance studio Storybook to create an AI TV-show trailer.

Here's the link to the video:​

The question you're probably wondering is... does it look like AI?

Yes, yes it does.

But the creators say the AI trailer isn't meant as a substitute for a real trailer.

Instead it does a lot of other stuff:

• Proof of concept to communicate the vision
• Generates interest before the “real” trailer is ready
• Showcases the abilities of AI for potential use in the film

The cool part? It took 3 people 1 week of full time work and was created with open source software only.

The creators used Stable Diffusion to generate the images (Juggernaut XL model) which were fed into SVD (Stable Video Diffusion)

If there was a better time to get into AI film, it's now.

​SVD (Stable Video Diffusion) Beginner Tutorial​

TO-TRY

5 more crazy things about Midjourney V6

A lot of you enjoyed these last week, so here's a few more:

1/ Make your own font

Everyone is talking about generating text but what about making your own font?

2/ V6 understands punctuation well

So be careful where you put your commas.

3/ V6 can handle a lot of things

Midjourney only starts 'forgetting' after 7-8 things. That's a lot!

4/ Here's a prompt template that works

This one was suggested by a member of the Midjourney team.

Set up your subjects at the start, and refer to them by what position (left, right, above, under) they are in. This lets you stay fairly accurate with complicated prompts.

5/ Use --no

What you write after "--no" specifies what you don't want

It's a mixed bag with V6; it can be absurdly powerful or completely useless.

RESOURCE

Free Book: SDXL Magic

SDXL Magic is a quick and easy handbook on how to get better SDXL generations. With:

• My favorite SDXL ComfyUI workflow
• Recommendations for SDXL checkpoint models, LoRAs, upscalers
• Example prompts

ROUNDUP

1/ Midjourney and StabilityAI (creator of Stable Diffusion) hit by copyright lawsuit: Artists allege that these organizations have infringed the rights of “millions of artists”.

2/ The scary truth about AI copyright is that nobody knows what will happen next: Both critics and supporters have long wondered: is this actually legal?

3/ PixArt-α (pronounced PixArt Alpha) is a new diffusion model that is comparable in quality to Midjourney & SDXL, but only takes 12% of Stable Diffusion v1.5's training time. This could be huge for people trying to train custom models. [Test it out]

4/ NVIDIA TensorRT Update will accerate the performance of SDXL, SDXL Turbo, and LCM-LoRAon RTX GPUs; SDXL Turbo will be able to generate up to 4 images per second. If you're running Stable Diffusion on an NVIDIA RTX GPU you'll want to check this out [Install for AUTOMATIC1111]

5/ The best ComfyUI workflows: The results of OpenArt's workflow contest: discover the best workflows for upscaling, face restoration, and video.

That's all folks!

You can reply directly to this email with feedback or comments. We read every reply.

Till next time,

Yubin & Crew

1223 Cleveland Ave #200, San Diego, CA 92103
​Unsubscribe · Preferences​