Trist Profile picture
Mar 16, 2023 โ€ข 10 tweets โ€ข 8 min read โ€ข Read on X
Some #Midjourney V5 findings ๐Ÿ”ฅ
Finally! Full-blown cinematic aspect ratios!

2.4:1 widescreen cinema standard โœ…
2.76:1 ultra panavision 70 โœ…

you can go up to 14:1 (!) but: letterboxing seems to set in roughly at ultra widescreen (3.6:1)

#midjourney5 ๐Ÿงต
However, 3.6:1 (and higher ratios) seems to work better if you drop the cinematic prefixes (cinematic shot, film still, etc.) ๐Ÿคทโ€โ™‚๏ธ๐Ÿค”
Here it's just scene & style description. 3.6:1, no letterboxing.
despite the letterboxing, exploring 4:1 is fun ... ๐Ÿ„๐ŸŽ‰
#MidjourneyV5
Here's a 4:1 image in panorama view. this is raw #midjourneyv5, default quality (--q 1), not bad actually
I mean, that photorealism tendency is next level, right?
#midjourney5 with strong tendency to photorealism ๐Ÿ“ท

same prompt, left: V4, right: V5
Tile mode is back! ๐Ÿ”ฅ Here in 14:1. Max aspect ratio, checked.
#Midjourney5
The --chaos value seems to vary the styles much more within the initial grid #MidjourneyV5 #midjourney
#Midjourney V5 also gives much better results for homogenous textures, thanks to improved realism & "--tile"

Top left(V4)/right(V5): "grey sandpaper, highly detailed --tile"

Bottom: mossy forest soul & rusty metal (both #midjourneyv5)
photorealistic wide angles and homogenous textures from #Midjourney V5 put into WebVR API to animate in a local browser window
๐Ÿ™ thanks #gpt4

#midjourney #aicinema

โ€ข โ€ข โ€ข

Missing some Tweet in this thread? You can try to force a refresh
ใ€€

Keep Current with Trist

Trist Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @tristwolff

Mar 28, 2023
1/ ๐Ÿš€ Runway's groundbreaking #Gen1 is here!

@runway has the tech, the network, and the community to reshape the online video landscape.

Here's why we should get ready...

๐Ÿงต๐Ÿ‘€
2/ ๐Ÿ’ก @runway has been developing AI video tools since 2018 and garnered support from content creators and major movie/TV studios. ๐ŸŽฌ

And in 2021, they teamed up with @LMU_Muenchen to develop the first version of #stablediffusion ๐Ÿคฏ

technologyreview.com/2023/02/06/106โ€ฆ

#AICinema #AiArt
3/ ๐ŸŒ During #Gen1's development, @runwayml partnered with researchers, filmmakers, and an amazingly creative community ๐Ÿš€ #aicinema

Read 8 tweets
Mar 26, 2023
Spending Sunday with #Midjourney ... ๐Ÿงต
Read 8 tweets
Mar 26, 2023
#Midjourney does not reliably understand lenses.

However, at least in an #AICinema context, you can use shot sizes to approximate the lens.

๐Ÿงต Here's how ...

(prompts โžก๏ธ ALT)
#AIArtCommuity #AiArt film still, astronaut in the jungle, close up --ar 3:2 --see
Basically, you can choose between these lenses:

๐Ÿ” Wide-angle lenses (10mm to 35mm)
๐Ÿ•บ Standard lenses (35mm to 85mm)
๐Ÿ‘€ Telephoto lenses (85mm to 300mm)

They often correspond to these shots:

Wide angle (wide-angle)
Medium to medium close-up (standard)
Close-up (tele)
Prompting for "close-up", getting medium close-ups ... ๐Ÿ˜Š
We wanted tele, but are probably a bit off toward standard. However, given the depth of field and compression, closer to 80mm than to 50mm. Fair enough. film still, astronaut in the jungle, close up --ar 3:2 --see
Read 11 tweets
Mar 25, 2023
๐Ÿ˜ฎ So excited that a major TV production company approached me to help them develop prompts & workflows for story development using #GPT4. ๐Ÿš€

โฐ For #screenwriters & #filmmakers, this is a wake-up call

๐Ÿงต
๐Ÿ’ฐ๐Ÿ’ป๐Ÿ’ก The industry is eager to save money & resources /w AI efficiency + #GPT4 is leading the charge. I expect production processes to be changed within a year.

The infamous #OpenAI study has shown: Creative tasks are the most exposed to AI automation

The consequences of AI are real, and they're coming to town. Now, how can AI improve the creative process? By helping with structural groundwork, for example:
medium.com/design-bootcamโ€ฆ
#AI #storytelling
Read 6 tweets
Mar 21, 2023
New paper on #GPT4 & "Labor Market Impact" is out.

It's an early evaluation, but creatives should be paying attention to this ... ๐Ÿ‘€

๐Ÿงต 1/5 Image
All the jobs that show 0 exposure to being automated by LLMs are defined by

- blue-collar, manual labor
- outdoor or non-office
- lower educational requirements
- use of specialized tools or equipment

As we'll see: jobs /w higher educational requirements โžก๏ธ high exposure ๐Ÿšจ 2/5 Image
โ˜๏ธIn fact, if your job is defined by

- creative or intellectual labor
- office or indoor work
- high educational requirements

it's much more exposed to being replaced by #GPT (alpha), or #GPT-based apps (beta & theta):
3/5 Image
Read 8 tweets
Mar 21, 2023
AI text-to-video is now available as an open-source model ๐Ÿš€

It's still in its early stages & nowhere near as mature as #Midjourney, but it's definitely worth exploring.

Here's how it works & how you can experiment with it.

๐Ÿงต
#Modelscope #text2video #AIArt #AiArtCommunity
The model is simply called โ€œtext-to-video synthesisโ€. A brief summary:

- 1.7 billion parameters ๐Ÿ”ฅ
- training data includes public datasets like LAION5B (5,85 billion image-text pairs), ImageNet (14 million images) & WebVid (10 million video-caption pairs) ๐ŸŒŽ
- open source ๐Ÿ’ช
Text-to-video synthesis consists of three sub-networks that work together to produce short MP4 video clips:

- a text feature extraction,
- a text feature-to-video diffusion model,
- and a video-to-video diffusion model.

๐Ÿš€๐Ÿš€๐Ÿš€
Read 11 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(