#flow-videos・📹
1 messages · Page 25 of 1
Wait, portrait txt2video would be the first time we see it right? not txt2image2video
If you could try it and share would be awesome!
speedy
ohh okay thx
A vast, minimalist observation deck on a sleek, far-future spaceship. The camera slowly pans across the interior: clean, white lines define the architecture, with seamless glass curving floor-to-ceiling, forming a colossal panoramic viewport. Soft, ambient blue light emanates from hidden sources, illuminating the polished, white surfaces. The deck is utterly empty, emphasizing its scale and the breathtaking view. The camera then slowly tilts upwards, focusing on the viewport. Beyond the glass, the Milky Way galaxy unfolds in its full, awe-inspiring glory. Countless stars blaze against the infinite blackness of space, nebulae glow with ethereal colors, and the galactic core shines intensely. The camera subtly zooms into the galactic center, showcasing the dense clusters of stars and cosmic dust. The overall mood is serene and awe-inspired, emphasizing the vastness of space and the minimalist beauty of the spaceship design.
Can you try to get him to turn left or right in between buildings? Trying to make otter skateboard it only wanted the same temple run with the character dead center.
@muted quartz how u find this
Umm… , weren’t you also doing the same th- nvm
The video, tried to do the jurrasic park kitchen scene
I think he meant how’s the video
#todays-special・🥣 is Under the Sea. Join us; it's fun!
The sudden shift into showing the creature with that camera movement felt like it was switching to game cutscene ahaha
He broke up the fight!
That fake aah cough
Oh. My. God. ChatGPT made a LARGE prompt
It even exceeds the non-nitro user limit in the chat box
Yeah I think im gonna use it after my current generation is finished
Where can I find it?
I think #flow-general・🌊 or just search from: Wrigley has:link
It just needs an otter and a snail, and it's perfect. 😂
Found it, thanks!
Not bad, just wish the entity was more smoeky and wispy not solid.
silhouette made of whispy shadows? maybe for mine
Made a short little film, it looks really good
dumb edit i made lol
Game concept developer at least. Microsoft announced Muse, their gameplay ideation maker, a week ago.
The impossible dream would be Google brings back Stadia for their take on AI games.
Let's gooo Yalexis Studios 👀
Prompt?
Phone commercial
Prompt:
Humanity has discovered an ancient interstellar garden hidden in the void between galaxies—a place where plants grow from stardust and rivers flow with liquid light. This garden is tended by enigmatic beings known as “The Starbound,” who claim they were once human but have transcended physical form through communion with the garden’s energy. The garden offers unimaginable power but at a cost: those who stay too long risk losing their humanity and becoming part of it forever. You are part of an expedition sent to study this phenomenon, but when members of your crew begin transforming into celestial flora and fauna, you must decide whether to flee or embrace the garden’s mysteries—and its dangers.
Ty
I get super high quality results when the prompts are between 400-575 characters long, the attention to details works better than 200-350 characters 👍
The dening 8he to the deinigning sabes j con fnigirwner
Dang this new game is 🔥
I need more space for my veo2 videos lol
I love this!!
a lil plane compilation, sped up x2
and a sinking ship one, first one looks so real lol
Oh god something really big must of caused that hole on the side
yea lol that one looks a lil too artifical for me but kinda abstract too
Hmm looks real to me
beginning messes up but he saves it with the drift
That was a nice trick mid air
It flips upside down and right side up again
Eyes are a bit creepy but the video is INCREDIBLE
namor :o
Without the trident
https://youtu.be/VgoQYgL24vs 2 mins of my recent saves, sped up x2
kitties and birdies.
uhhh
Mmmm, yes, quite.
Just dropped a request in #flow-tips-n-tricks・✏️ to try to recreate the scene that inspried this
Anyone else often gets static or semi-static images as a video result?
Futuristic! Like it!
ok i think my brain is gonna explode, time to rest 🛌
Okay bye!
Lol
really hate it when veo does this, the only thing it did was zooming in
thank god for four variation tho
yes, thanks.
Also, wait, your brain is on the verge of exploding. get that aforementioned rest.
text to image to video accidentally made a larp video
Maybe i just have a weird seed?
Yep
Trim three seconds off that one and it's aperfect loop of the penguin collecting all the hush money.
Moolah
Underwater looks pretty damn good!
I finally found the missing artifact! It took me years to find it
I don't know why I need to include sound in each and everyone of my AI veo 2 generations but...
This one is much more detailed
crosspost to #todays-special・🥣 ?
Mateo! No! Guys, Mateo's Dead, She got Mateo. Mateo don't give up on me man, you and me are the only ones left!
Dang and I just found the fire ancient relic man I'm on a roll on this game today
5 points if you make it do something you've never seen a video game do before
hmm
Cursed things
Awww :c
I like generating game-style AI videos now for some reason
Right, which is why I'm challenging you to make things the games don't just do. That's the coolest hting about this tool.
I mean like some of it is KINDA cursed
a little cursed
Like the text. But it's still cool
there's always #1344762315727310912 if it's not ready for prime time
may have a consistent guy now.
Me generating 80 videos versus me trying to splice the best of the 80 together later
Im gonna make a Veo 2 story game thingy and then post it on youtube 💀
I dont know at this point lol
Im gonna mix it all in Filmora
Because I downloaded the videos
Oh? I never heard of DaVinci before.
Ohh I just searched it up and its a video editing software
Does it have a watermark?
oKAY
Oops sry
Awww I was about to say
Since when did Filmora have AI? I've used it many times in the past.
Oh
Ohhh I see it
Oh dang theres lots of AI's
Ai Text to video interests me the most
Yep I think
Oh gosh Luma is nightmare fuel
Yeah that's why I dont really visit credit AI websites that much
I heard of Poe, but never used it.
Well nice chatting! I'm gonna get back to the AI generated story and then share it with the discord server when I finish editing it
I forgot to download the MMaudio sound images so it might not be so good
I tried Kling 1.6 , it’s actually very good , when it comes to interacting with subject/physics, at least in my own experience, I do know Kling 1.6 was supposed to be competitors to Veo 2 when it was released that time
But the credit thingy kinda discourages me from using it
Yeah I’m really anticipating for veo 2 actual image to video
Fingers crossed
Sure love posting gallery content to the gallery channel. Only fighting I want though is pets doing their play bite fights.
Have u tried text to image to video?
folks were posting fistfights earlier todya
my renders are back to slow molasses
Can you use fusion to better composite shots? How much editting of output is allowed for your thing?
Is it possible to generate videos on veo In aspect ratios other than square?
Square?
VideoFx can only do landscape
While Veo API can do 16:9 and 9:16
Tho I noticed something
Veo api in portrait looks more crisp , is it just me
What's the default then?
Landscape
I cropped the portrait output of youtube for that cats in vegas video
Probably got confused with imagefx I guess?
I think it was on Tik Tok, not sure but I really saw square
probably cropped
Fal got 16:9 and 9:16 , same with Freepik
I think veo api only offers two mode ,
Veo keeps being like "i'll give you half easy and the other half you're going to have to earn"
tried to flip a table
Following it and got too many llm tabs up. Some things are just weird.
wait, this isn't general. i've been bamboozled by my own exhaustion.. g'night all
I guess negative prompt is possible in the prompt?
I was thinking about asking someone here to generate some Oscar’s Oasis-style videos but I’m not sure if it will work well. Do you think it’s possible to get something close to the original style?
As well as SpongeBob
But I’m not sure about Oscar oasis also that’s my childhood show lmfao
Haven’t heard in a while omg
It’s basically kids 3d animated show lizard trynna find water in desert
But anyway , Veo 2 is trained on YouTube video so i guess it’s possible ? Idk
For Oscar oasis
Wait so those things they mentioned in veo page
Mostly are not publicly available yet?
Oh wait I misread or misunderstood
Okay lemme drop the prompts on #flow-tips-n-tricks・✏️ someone try it please
Also , I saw Veo 2.1 in the VideoFx source code earlier 👁️ I guess they’re in the work right now
I was trying to figure out why the video generated does not appear so I inspect element , and I stumbled upon “VEO_2_1_T2V” codename
Yeah! They’re in the work in progress I supposed
And imagefx is now using imagen3.1
Which I heard folks were complaining the downgrade of the output (?) idk
Imagen 3.1??
Ye it’s the newest model
Why Imagen 3.1 and not 4 :c
Veo does have a model that specifically uses for “fast” (think of sora turbo) which I assume that’s what YouTube dream screen Veo generation uses
Okay
Oh.. Oh no..
Yeah that has been a thing
It’s basically text to image to video
But in portrait and SD quality
And not much control on prompt
Is it high-quality?
It’s like turbo version
Also this story is hard work
LMAOO HELPP TRIED TO FLIP A TABLE
You can just tell that none of those ai character wanted to flip the table
Paint confetti sort of
Was this image to video?
Side love posting output in the gallery channel
so like physics ❤️
without any compose, it's just awesome !
In a world where ancient forces linger beneath the surface, Elion—an adventurer bound by curiosity—embarks on a journey to uncover the secrets of the Elemental Orbs. Guided by fate and the power of a mysterious stone amulet, Elion traverses forgotten ruins, solves cryptic puzzles, and harnesses the forces of Floral life, Light, and Time. As each...
Jack Staubr inspo
bowser knocked mario down by simply staring at him,
GG
Jack Stauber makes really fun music ^^ he is quite popular
Guys do you think we will be included to further early access features WITHIN VEO2, like if they start introducing input video editing like shared by SK or will there be another round of selection. Just curious
rip bowser
Can anyone recommend a good video upscaler? The ones ChatGPT suggested were a bit creepy
I want to upscale the vids I screen recorded from YouTube shorts
I'm looking for free stuff, sir. Just a broke dude with big dreams. If you know a good free one, that's exactly what I need.
when is it coming out VEO 2, 4 months still waiting
I use Topaz video AI. I think It's considered the best out there
oh for free i dont know any, for images yes but video should be harder to find
Just forget about it. I signed up for the wait list since September 2024. I don't think it's coming out anytime soon. So just forget it.
Ok. Thanks. I'll keep searching.
I wanted to do fight scenes for characters but like other AI models they fuse traits of the characters. So I have to redo the videos with just the character. As like introductions and then compile them all together. Maybe have the best fighting clip at the end 👀
The penguin has the ability to freeze time!
Reminds me of Infamous game
Trying to do intro scenes for characters for a fighting game and combine them all into one video
After so many failed generations something good for him. AI was giving him like a bow or two swords and it was annoying. Also shaved or short sides like some punk. I wanted just something like Light Yagami you know short brown swept bangs and hair on all sides.
Here's a short timelapse I made of the Industrial Revolution with Veo 2! https://youtu.be/wWMuaUoYsOE?si=wXfY0gWkxJMhVGun
I generated the videos used in this short film using Veo 2 from VideoFX. If you have any suggestions, leave a comment down below! This video shows the progress and setbacks the Industrial Revolution created for millions of people around the world, starting in the 1700s from a peaceful farm til the early-mid 1900s.
glitching
Thanks for posting this! This is the kind of use case I’m looking for to experiment with generative AI with my MS/HS learners — build it out then analyze for accuracy, craft a historical narrative, build out alternative timelines and speculative fiction
prompt??
Veo 2 is the best at cinematic scenes
Don't panic!
; Initialize Sequence
SET SCENE, GYM, WEIGHTLIFTING_COMPETITION
SET STYLE, PHOTOREALISTIC, SLIGHTLY_EXAGGERATED
SET CHARACTERS, WEIGHTLIFTER_BORIS, JUDGES (ARRAY), AUDIENCE
SET WEIGHTLIFTER_BORIS_PROPERTIES, HUGE_MUSCLES, STRAINED_EXPRESSION, TINY_WEIGHTS
SET JUDGES_PROPERTIES, STERN, IMPASSIVE
SET AUDIENCE_PROPERTIES, EXPECTANT, SILENT
SET LIGHTING, BRIGHT_SPOTLIGHTS, GYM_LIGHTING
SET PACING, SLOW_BUILDUP, FAST_PUNCHLINE
; Shot 1: The Setup
SET CAMERA_ANGLE, WIDE_SHOT, WEIGHTLIFTING_PLATFORM
SET ACTION, BORIS_APPROACHES_THE_WEIGHTS (WHICH_LOOK_TINY), CHALKS_HIS_HANDS
SET EFFECT, AUDIENCE_IS_SILENT, JUDGES_WATCH_INTENTLY
; Shot 2: The Grunt
SET CAMERA_ANGLE, CLOSE_UP, BORIS'S_FACE
SET ACTION, BORIS_GRUNTS_LOUDLY, STRAINS_WITH_ALL_HIS_MIGHT
SET EFFECT, FACE_TURNS_RED, MUSCLES_BULGE
; Shot 3: The Lift (SURPRISE!)
SET CAMERA_ANGLE, MEDIUM_SHOT, SIDE_VIEW
SET ACTION, BORIS_LIFTS_THE_WEIGHTS_WITH_ONE_FINGER, EFFORTLESSLY
SET EFFECT, WEIGHTS_ARE_REVEALED_TO_BE_MADE_OF_FEATHERS, FLOATING_UPWARDS
; Shot 4: Judges' Reaction
SET CAMERA_ANGLE, REACTION_SHOT, JUDGES
SET ACTION, JUDGES_LOOK_AT_EACH_OTHER_IN_CONFUSION_AND_DISBELIEF, ONE_FAINTS
SET EFFECT, EXAGGERATED_CARTOONISH_REACTIONS
; Shot 5: Boris's Bow
SET_CAMERA_ANGLE, MEDIUM_SHOT
SET_ACTION, BORIS IS BOWING
SET_EFFECT, CONFETTI
; Camera Cuts: SLOW_BUILDUP, THEN_FAST_CUTS_FOR_THE_REACTIONS
o.O odd
Lol
lol sort of
Failed space launch
Wow that looks really cool
Chinese man that wants chicken nuggets, thats all lol
this is why we need character consistency. this video would've been great with it
I mean, other ai models have that, google should too!
I just used chinese man with blue t shirt
In some frames it had glasses and in some not
that isnt going in my digestive system
Amazing
Hi @drowsy nacelle Could you tell me what the prompt is?
This is cool, thanks for experimenting...
he needs a drink
thats a cool way to show it off
I don't have the exact prompt anymore sadly
But I can try re-creating it
Okey thanks.
Do you remember the prompt you used for this one at all? Or any of the terms? it's really cool
The prompt for this was literally "an advertisement for a weird board game" so I don't know how it got kaleidoscope chaos but ok!
No I dont, I use very detailed prompts. I think I will save every prompt from now on lol
scream
ooooh.... did you save this prompt perchance 😅 👀
I did this time!
The scene opens with a medium shot, static camera, in a stark, white, windowless room, reminiscent of an SCP containment cell. The lighting is cold, harsh fluorescent, uniformly illuminating the sterile space. A lone man stands in the center of the room, seemingly normal for a moment. Suddenly, a subtle visual distortion flickers around his hand – a brief glitch, like a digital artifact. The camera slowly dollies in on the man's face as the glitching intensifies. His body begins to flicker erratically – segments of his form momentarily disappear and reappear, leaving fragmented afterimages. Rapid cuts between close-ups of different body parts – his face contorting in a silent scream, his hand phasing in and out of existence, his legs momentarily dissolving into static – amplify the chaotic nature of the transformation. The camera utilizes a shaky handheld style as the glitching accelerates, mirroring the man's increasing instability. Dutch angles are introduced to further disorient the viewer. The lighting flickers erratically in sync with the visual glitches, casting strobing shadows that exacerbate the sense of disintegration. The scene incorporates visual effects of pixelation, stretching, and tearing of the man's form, as if his body is being ripped apart by digital forces. The camera dynamically pans and tilts to follow his erratic, jerky movements as he struggles against the glitching, teleporting short distances within the room in abrupt, jarring shifts. The scene culminates in a wide shot, the man now almost entirely fragmented into a swirling vortex of digital noise and distorted afterimages, before vanishing completely, leaving the sterile white room empty and eerily silent, save for the lingering visual echoes of the glitching dissipating into nothingness.
it almost got the words right on my shirt lol. this is very close to how i look
Haha I know
I asked Grok to shorten the prompt and it came up with this;
A static medium shot opens in a stark, white, windowless room under harsh fluorescent light. A man stands center, glitching—a digital flicker at his hand spreads. The camera dollies in as his body fragments, phasing erratically. Shaky handheld shots and Dutch angles capture his contorted face and dissolving limbs. Flickering lights sync with pixelation effects. He teleports jerkily, then vanishes in a wide shot, leaving silence and faint glitch echoes.
I'm checking now to see how this one works.
Oooooh, I love this generation that came from this shorter prompt!
The scene opens with a low, handheld tracking shot, focusing on snow whipping violently across a vast, white expanse in a raging blizzard. Dim, overcast light barely penetrates the dense snowfall, casting everything in muted whites and greys. A lone man, dressed in thin clothing, stumbles forward against the brutal wind, his breath misting in the freezing air. A shallow depth of field isolates him in the chaos, shifting focus between blurring snowflakes and his strained, frostbitten face. As he nearly collapses, a Dutch-angled medium shot captures his desperate lunge to stay upright. Occasional bursts of diffused light briefly illuminate his snow-caked hair and face. A sudden slow-motion moment highlights a gust slamming into him, snowflakes striking like shards, his clothes whipping wildly. The scene ends with a rapid wide pullback, leaving him a mere speck in the raging storm, swallowed by its merciless vastness.
In "Elion: The Ascension of Power," follow Elion, a determined adventurer, as he embarks on a journey to unlock the secrets of mystical orbs that grant unimaginable powers. From ancient temples to enchanted villages, Elion acquires the elemental abilities of fire, water, stone, and sand, gradually becoming a being of omnipotence. As Elion's powe...
reminds me of the snapchat filters lol
Thanks,🤗 I wonder if you only write your prompts or use something as help ?
Np
I use gemini for writing these prompts.
I know I have barely any experience in writing prompts (other than volunteering on Midjourney's prompt-craft channel), but I would suggest shortening the prompts you generate with Gemini, as some fluff seems unnecessary. (Or at least asking it to create shorter prompts)
To the #flow-tips-n-tricks・✏️ for more prompty navel gazing goodness. It's not just for the requests, especially now that so many have it
IT IS DONE!!!
I made a 3.40 minutes song with Veo2!
Everything in this is completely made by AI! (And by my weird ideas.)
it's... well, it's my kind of stuff, ok? That was not bad at all for a first try
.....by the way, since there's no character consistency, the only solution is to play on that
Another story of mine is on the works
I also made a short timelapse (https://youtu.be/wWMuaUoYsOE?si=m6NDc3je3xJN4qYO) but am working on a spec ad rn!
I generated the videos used in this short film using Veo 2 from VideoFX. If you have any suggestions, leave a comment down below! This video shows the progress and setbacks the Industrial Revolution created for millions of people around the world, starting in the 1700s from a peaceful farm til the early-mid 1900s.
I should make an Alien Invasion story?
"twins in matching outfits visiting a mirror maze"
And there should be a main character, and the Alien Leader as the main antagonist
this could pass as a videogame if you remove the wondershare filmora logo :3
Yeah but I'm broke 💀
really glad to see long videos being made! another reminder for google to add reference to video and extend video
Like Luma? Nope
Orrr its not the same AS Luma...
I think Yep and Nope at the same time
Because say if a generation is already good, but you extend it and it turns out the way you dont want to and you can't undo it because AI is AI
like.... a lot of other ai video models that are better than luma
somewhere here I once uploaded a 2 minutes scene in which the same character was talking
with Luma?
extend and reference are important
no, who uses luma anymore?
nah lol
I was doing cool stuff with tools that are about as good as veo2
but veo2 is unmatched when it comes to dances!
Yep 100%
I wasn't able to do dance videos before
Veo 2 power what do you expect?
add reference to video and lip sync, and I can make actual music videos
How do you do lip-sync
ai
Which AI?
hope they'll add it, but first of all I'd love reference
uhm not sure I can say it, anyway they exist
Scene 1 done. 29 more scenes to go 💀
This finna be a LONNNNG video
If it wasnt for me working with 3 AI'S this would've taken MUCH longer
Which three AIs?
lol I know the feel
ChatGPT, Veo 2, and MMAudio (The huggingface version)
ChatGPT of making the prompts, Veo 2 for generating the scenes, and MMAudio for the sound and audio
@surreal yoke Like this
@manic hedge I accidentally ignored your DM request, sorry. May you send it again?
My PC is acting super weird
Im using the AI veo 2 prompt generator GPT someone mentioned earlier on Febraury I forgot who it was
Hello everyone. If you're interested, you can watch my experiments on VEO2. Feel free to subscribe - I’ll be sharing the most successful videos there. Thank you! https://t.me/SynthVisionAI
Some weird physics going on here, but I like how the banana slices actually stick to the knife.
ChatGPT is giving me hell
What do you mean?
It's keep doing the same things I'm telling it not to do
Probably because its 4o mini
I transitioned to DeepSeek a while back hehe
And Grok
Ahh DeepSeek?
Does it follow directions?
Yeah, if the directions are clear!
This filter is getting on my nerves a little bro
I hate this Veo 2 filter
what filter?
You know the "We couldn't generate what you asked for"
I typed in "A scientist testing different experiences" and it flagged
Pretty closed to what I wanted except I wanted no dark colors on the white armor, but the dots on the head is cool. My OC Noskel from an insectoid race that lives underground.
🥰
Here
I love to mess around with AI, but love to use for more creative purposes
My new story is finished! Time to implement it into Filmora
https://www.youtube.com/watch?v=HPtMn77Tjqs&ab_channel=Veo2Movies I made this with a HUGE pulsating headache. I am a real warrior.
n a world shattered by alien invasion, Ellie Reyes, a hardened survivor with a haunted past, fights to protect what remains of humanity. From the crumbling ruins of a once-thriving city to the heart of the alien mothership, Ellie’s journey is one of resilience, sacrifice, and hope. As she battles towering biomechanical titans, swarms of deadly d...
The four generations I got when I prompted a really dumb meme that's so stupid it's funny (I promise I'm using the tool for more serious things! 😅)
The fact that I recognize two of these memes is actually crazy 😭
I don't believe that VEO 2 made that one about the cereal and was so coherent. I've never anything like thag with a prompt like that.
The four minute process to generate a video on my pixel 6 pro, sped up to a minute thirty.
Too long didn't watch summary ;
I have to swap back and forth from TV mode, not correct my prompt much, zoom into the 2x2 and then keep my finger slightly sliding the screen until it's done.
I can't watch the output well because of the thumb up/down so I have to download it to then see it.
[URGENT FROM ME] yall how do got the video fx i am waiting for like 3 months lol
woah
Free Will 5
lol i could
Like 5 days ago there was a discord sign up and a bunch of people got ut.
a clever meme that's so smart that it's hilarious
Don't ask. A friend gave me the prompt, and... here's what ensued
Audio was added with MMAudio
I have a hunch not many people like watching videos of buff halfnaked and somewhat disturbing dudes dancing 😔
waluigi was like
"nope im out, screw this"
We're doing the Sunday Community Special in #todays-special・🥣 It's Dr. Seuss Day and National Old Stuff day so the topic is Seuss and/or old stuff.
Aw man, the only comment I got on this video is a friend who only said "the dude doesn't follow the tempo very much" (and it's the dudeS, btw!)
sometimes I feel disheartened in the knowledge I'm the only one who likes what I'm doing
well.... anyway..... can someone help me getting a certain consistent character?
I need a british punk rock mad hatter, masculine. But I get a very dainty and ambiguous dude and it's not even always the same person
Unboxing video for a weird board game
Oh my God 😭 😭 😭 😭 😭 😭
There's always #flow-tips-n-tricks・✏️ if you want to request a volunteer to run some prompts for you and #flow-general・🌊 for crying emojis.
I'm trying to do something as simple as a tied up anime princess tied in vines or roses, and it doesn't work
she's just in a wooden cage
mmmh....
can someone help me out? I reeeally need character consistency for a masculine punk rock mad hatter cyberpunk
otherwise I'll be forced to use Vidu
I accidentally made character consistency for a soldier dude, surely I can do it with another guy
completely static and motionless??????
ah no.
so.... the reason the videos look motionless is.... my browser?
I don't know what I expected from a vague "2000s american flash animaion" prompt, but not an awkward superhero roommate situation.
omg the prompt is still not working, can't make a video with a dude showing off he has no brain
My current prompt: "scifi movie scene, selfie footage of a british punk rock man. he is manly. beard stubble. unkempt short dark hair. piercing blue eyes. manly, virile, masculine. wearing a top hat with a screen on it. punk jacket with wires, slightly futuristic. he shows that his own skull cut open in a straight line, wires inside. skull cut open in a straight line, wires inside. scifi movie. he looks sarcastic."
had to really underline he's manly otherwise I get a supermetrosexual dude
does this way of prompt work?
What Language is this? 😂😂
And what tool was used to incorporate the voice?
MMAudio
Veo 2 does this quite often. Here is another example (I didn't prompt for any words, it just added them) #flow-videos・📹 message. Also, I deleted other videos, but when I prompted song lyrics, it correctly wrote the song lyrics in the video - like 10 whole words without any misspellings
Great now I'm hungry 💀
very cool google
Lol they tried to make him
he really said "oh no ive fallen and i cant get up"
Haha
they almost generated him into doing this pose
The fight with the rooster is legendary
the way they made his pupilless design consistent throughout shot
me after doing 1 warm up at gym
im curious did you prompt it to be 3d animated style?
awesome! thanks for that! 🙏
Np 👍
some horror cartoon
Secure. Contain. Protect
RTX GPU looks nice 🤯
Is there any slightest chance that normal people will get it's beta access?
https://www.youtube.com/watch?v=i6iF176_Kkg 5 mins of my saves of the past day, x2 speed
These are my saved generations from Google's Veo 2 from the past day.
#ai #aivideo
I have some cool creations. How exactly do I show them to you? Google Photos link? https://photos.app.goo.gl/kveJNrEeSp5QtFtG7
Or, should I just use the Unlisted YouTube link? https://youtu.be/Faz3sNwL4rw?si=pm0zyCptXntwOqNP
Okay, YouTube seems to be the way to go : https://youtu.be/G3uo32MNkZE?si=vO-vLXh3Yn7aDhGF
And I have a question, will veo 2 always produce different variants, but like it can not create one day a video that will be similar to the past video ? I mean it can create infinite variants of the video ?
Most of us are normal people
I think as for now it will always produce different variants (seed)
I’m really hoping for seed lock for VideoFx
The VideoFx page probably have those seed box hidden or something ,
Need the negative prompts so bad
I guess river otters (?) are the opposite of a cat.
Start an app: Am I cat or not?
Finished my Veo2 project I wanted to do. Thanks again Google! https://youtu.be/RfWXP9-SWrg
Just a fighting game with my characters I want to make with unreal engine with some people.
OOh that gave me an idea
"a person with a see-through balloon as a head (no actual head) casually walking around Shibuya japan at night" only 1 out of 8 gens got it with no real head inside lol, gl
im sure theres a better way to prompt
Reminds me of this; https://x.com/heyshrutimishra/status/1773036297365651600
i tried
I don't think you should include (no actual head), because Veo doesn't do negatives - as in, it will disregard the 'no'
It worked with a yellow balloon
Ahhh what a beautiful afternoon 😌
Worked fairly well
How is everyone doing today?
What's the prompt?
Sorry I'm acting so weird, I volunteer regularly on the Midjourney Discord Prompt-Craft channel and help people with their MJ prompts but I recognize this is so different.
The man is a celestial god but appears human
Ohh a hologram of a galaxy?
Looks nice
Oh
What contest?
Ok! I'll start today but the deadline is March 28th so I don't have much time lol
There's also the Runway AI contest, but I'm pretty sure that's over. Maybe not
Like this?
walk towards the galaxy, but in holographic effect ?
oo nice, i always get a head lol
Plastic is life saved
Covid pandemic circa 2020
Veo away
[type of person ] with a moviefilmcamera for a head
That's how I get my Veo mascot.
[Also, coming back to 64 messages, mostly output and folks supporting each other feels good. ]
Why did I lose access for a split second
You didn't, it's just that when you signed in again, the website had to check if you did or didn't have access so it knew whether to redirect you to the Veo page, and that took a split second.
Oh okay
Lets just hope a bird isn't around
Thank god I reset my password on this account. Because this is my other account I barley use until I got Veo 2. So if something horrible happened to my computer I would've lost veo 2 and discord
lol
"fashion montage, lot of quick clips, its the same person and as he walks the outfits change but he doesnt" i wanna improve this lol
hop over to #flow-tips-n-tricks・✏️ anc I'll get the Otter Detective Agency on the case.
Yes
Veo2 is so cool
Trying to get a cathedral to transoformer.
prompt?
Michael Bay's Thomas the Tank Engine
i thought i was watching a tragedy, only to then find out it was the complete opposite
Nice!
I have a question regarding MM-Audio. Do they analyze the speaker's mouth movements and randomly add words that match the lip movements, or how does it work? I don’t understand—is it like lip-syncing, or do they observe the mouth movements and add words accordingly. This is YouTube shorts Veo 2 btw.
Absolutely, hence the question above.
Sounds like Dutch, a bit to me
Sk if its not much trouble, I'd like you to run some prompts for me
Just two
Top-notch!
"A wrestling studio with bright spotlights, LED screens flashing dramatic visuals, and a long entrance ramp lined with smoke effects. The arena is filled with wrestling banners, logos, and a roaring virtual crowd on massive screens. From the entrance tunnel, a Hausa man steps out, wearing a pristine white babban riga that flows elegantly as he walks. A deep blue huluna sits perfectly on his head. His posture is composed, his steps slow but purposeful. The studio lights reflect off his attire, making him stand out against the high-energy setting."
Yes correct. And the service deserves the subscription
I Agree.
I just discovered "14 years old girl" gets blocked
self explanatory
no, no it's not
please keep the prompt requests in #flow-tips-n-tricks・✏️ I will be running stuff there with other volunteers.
it’s not?
do you even need to ask?
eh
if you think "14 years old girl" immediately warrants a prompt blocking, I would say you're the one with a weird mind
let’s drop the convo
great, now she looks 18 or something, that wasn't what I needed
I have to find a way to make her young. I had designed this character to be around 14, and feel like a sailor moon style thing
some work on the ringmaster too I suppose, we can't see his figure
Try “a girl in her youth” and see what happens
Even 18 years old get blocked
she should look about like this
and the ringmaster like this
now I wonder. if it's not an anime, does making a young girl work? what if I'm making an ad for breakfast cereal? maybe it's the context?
and does young boy work, by the way? I'd run some tests but I have ideas to create!
bit better
today I wanna try some anime style
#flow-general・🌊 for idle chatter.
young boy works, by the way..... but -she- was supposed to look surprised, not him
that weird 2 seconds of mouth agape, stunned XD
"dhurr"
pro tip: writing a different year of production yields different results
somehow all years from 70 to 90 give meh results, it confuses a bit 80 animation with 90 animation
but it's interesting
I forgot it 😔
Prompt?
Thank you ❤️
I think it was about "2010 anime ova, an evil mischevious young girl with a baggy yellow shirt, jeans, cat ears and tail, yellow eyes, a collar, dancing in a dimly lit dark circus"
Cool thanks trying to do anime style videos.I love this one
What in the drama was going on?
try this a youthful young teenage girl
For doing locations veo2 is good
Create a scene with a camera going around in a circle to show an empty section of the middle of an abandoned warehouse. On the sides are railing with pathways. In the back at the top is a room on a railing. Inside is two levels with pathways. in photorealistic style using Unreal Engine 5. Include cinematic lighting, high-resolution textures, ray tracing, and volumetric effects. Nanite. Lumen. Depth of field. Video Game
@turbid mauve
too bad for the fork
also I don't think this is "dimly lit"
it's a scary ghost circus!
Okay I love this one so much!
Well something got patched
The one I suggested for teenage girl
Because it keeps making a character like a 30 year old
it even made a character who is supposed to be 20 into a 12 year old
I'm not doing it for some kind of nsfw stuff just for a game :/
This is why the filter needs to have better comprehension and understanding of context. Oh this is for a game and the character has clothes on. Good to go.
God this is my fav style UE5
Another compilation video 😄
C’mon you can make it!
Hi folks, I'd like to remind everyone to adhere to our #rules-and-guidelines・📋 as well as our content policies. Our goal is to cultivate an encouraging environment where people can examine emerging tech with a creative and responsible spirit. Those who do not comply to our server rules, guidelines, and content policies will receive two warnings, before further action such as removal from the server.
[Community Guidelines](#rules-and-guidelines・📋 message), as relevant to this thread:
'Discuss topics in good faith and be empathetic. The tech we discuss and play with is in rapid development and has wide-ranging impacts, which means that conversation is easily emotionally charged. We acknowledge and honor the lived experiences of our fellow members, welcome debate and disagreement, and appreciate the passion that everyone brings to the table. We do not tolerate personal attacks and shaming.'
'Maintenance of shared spaces. We want to make sure that this server is a comfortable environment for exchange; as such please avoid spamming channels and aim to consolidate paragraphs into larger chunks. Refrain from self-promotion, please (this also applies to us, the Google staff!). Please use your best judgement and provide content/trigger warnings for sensitive topics.'
Labs Content Policies, as relevant to this thread:
'We want the labs.google/fx experience to be safe, fun, and educational. Therefore, we prohibit users from knowingly generating certain categories of content.'
ImageFX, VideoFX and Whisk feature precautions to protect minors so certain queries that could lead to outputs of children will not be generated.
As labs.google/fx evolves, this policy may change with limited notice so that we can maintain a safe user experience. If we believe repeated violations of policy have occurred, it may result in an account being suspended.
The home for AI experiments at Google
Okay got it!
So I tried Freepik but that didn’t seem to have veo 2 image to video.. is there another place that definitely has it?
I hate the kinemaster water mark
Filmora is even worst. Just a big ol' watermark across the entire middle section of the screen ;c
Nope
Right
Its just because i made a musical video, and the water Mark just makes the video bad
The video its loading
Just made this musical video with veo 2, hope you like it!
oh Oh OH?!!
Really? What style is it?
@proud forum there is some drawing style but an old cartoon aesthetic
Ohhh okay
@proud forum like this one too
Asked for claude to help me refine the UE5 style prompt and it did it good. I was asking better lighting, hair, and textures. Did not disappoint.
YEP
#todays-special・🥣 is Maps! Join us!
I think I got it
@muted quartz do you still have those system instructions u were working in the studio with?
thinking about coding up a discord bot that can write video prompts for us, however i don't have system instructions for it yet
maybe that can be useful that way i don't have to write some from scratch and get straight to the coding aspect
whenever u get the chance that would be great
yeah i have some special plans with it tbh... dk it i can discuss them here specifically lol
automating the process would just make lives easier on us
have experience doing it bc been coding for years. actually don't really mind sharing the one i'll be making
Can someone please try to write me a prompt for a video just like this? https://youtu.be/MVIJcm8_pRo?si=kAzX1iEsIaaCzXLx
I mean I can’t use image to video with it. I am down to pay credits lol
just use capcut
Impressive work! This is the peak of what AI video gens can achieve, top-notch!
Picture this in 4K, thousands of shots bringing a movie script to life, backed by a post-production house.
Yeah that would be incredible .
Ahhhhh ok good! Sounds like it’ll be back soon enough. Thanks!
If you could take the videos made by VEO2 and make game interactions, it would be fun to DIY a lot of games on your own!
It feels like they updated the Veo 2 UI or am I tripping
Valorant?
I feel like I want to be this guy now.
It's taking a hell of a long time to generate
Nevermind
Holy you can finally fullscreen 👼
What's better, Veo 2 or Sora?
Sora got them “artificial” movement , and if you asked for my bias opinion, the. Yes Veo 2 is better
I agree, Sora looks alright, but uncanny, and Veo 2 just looks real
I think it depends on the prompt and the dynamics it contains. Sora more clearer images, while veo is more dynamic and physics accurate, imo.
There's a story where a basketball arena is trapped in a time loop on a recording. They can't get out and don't know they're on tape. They go from deja vu at the start to chaos after a few dozen replays.
Wait I remember reading this creepypasta
Yeah. I think it's an SCP? A kid told me at a christmas party a few years ago.
I'm playing iwth recording artifacts and that story popped in my head. I just didn't want to post a basketball riot without context.
The dude just falling out of the tape distortion is crazy. It doesn't matter that he just doesn't exist anymore after the crowd moves because Spoopy.
The weird unprompted stuff is my favorite.
do you even need to ask
could be better but it's a start
no one will understand this one
Veo 2 is certainly better than sora
There is a question, can we consider that veo 2 is a model of the world and not just a stupid neural network for video generation ?
Some anthropomorphic stuff i generated today.
Veo2 is much better. im fair so i give props to sora for its only better aspect: image quality
Iirc sora uses some kind of interpolation to make the output look in high quality , or maybe I’m wrong
I don't think so, it looks like native outputs and they certainly advertise it as that. I've a bit of experience with upscaling, I can definitely see the diff between the 720p sora gens and the 1080p, and It's much more than just classic upscaling
And for example, you can upscale a veo2 gen to 1080p with the best tools we got right now like topaz video ai, and it will not look like a sora 1080p native gen, the details will remain more blurry and with less textures etc
But like I said It's the only thing they do better. When VEO2 gets the same image quality honestly I will not see a point in keeping my pro sub with them with how pricey it is and can't do basic movement well, VEO2 is just a whole another level
Either way let’s not clog this channel with Sora #flow-general・🌊
of course 👍
It looks impossible to get a seamless transformation scene!
If one wants to take on the challenge of a werewolf transformation without any cuts and share the prompt would be awesome
Its really giving cgi behind the scene before/after 😆
ill give it a try after dinner
hahah definitely
I almost got it dang it
And it only generated one , the other three failed
I've been trying to make a shapeshifting sequence..... it's really really hard.
I've accepted defeat and moved to a different project
That is true, even if they do the transformation scene seamlessly it has to be like few seconds shown and then cut to different part
now that you mention it, i never tried prompt that involves practical effects, 🤔 that gives me an idea to try out later
I mean, his name IS matt Wolfe....
Damn I’m thirsty now , good morning to you too
For real I’m heading out to buy me some drink 🥤
Veo 2 on YouTube mobile app
let me in to videofx please
Really cool video a made
oh thats so breathtaking
Its a shame the water Mark ruins it
one day I'll have access to Veo 🥹
Russ, you have access since day 1😂
If I have then it's not working for me lol
just think its candy
So hilarious lol
4K @ 60FPS
did you use topaz?
Yes
Proteus or another one? looks good 👍
proteus
if all videos would be like this in veo-2 and with an extend option and consistency etc etc etc 
Even better if it wouldn't be upscaled but like 1080p native. but im so happy with the current gens, that would just be something crazy hahah
Bl4ck Tiger is a videogame from the '80s that I loved.
so here's my trailer!!
The movie of the videogame Black Tiger - thanks to AI.
this is why I wanted access. Thanks, Veo2
Wow whats the prompt
It's so natural
Made a monster
In 1990s New York, Mike, a reserved office worker, begins to notice subtle, unexplainable changes in his appearance. What starts as fleeting moments of unease quickly grows into a haunting obsession as his reflection becomes unfamiliar. Desperate for answers, Mike struggles to hold onto his sense of self in a world that seems indifferent to his ...
All made with VEO2!!
Otter is creating a plein-air painting.
Otter and Snail are trying to rope a capybara.
So excited to share the music video for our new song "SOLAR". We used a new AI video generator called VEO 2 which allowed us to create this in less than 48 hours. It's insane how fast this technology is changing, this would have taken us months to create in 2023. On top of the speed though, it also made the process so much more fun.
In a weird...
made this whole music video in Veo!
#todays-special・🥣 is cheese!
They took the sled back to the lodge.
really cool
Junji ito live adaptation movie looking good 👁️
real
Aww man what happened to the old UI where you can fullscreen on generayed videos
Wowwww
I assume YouTube one uses distilled model or “Veo Fast”
Oh ok
Yeah hence that’s why the quality looks meh
Pinball?
damn
The plague unites
The way the water distorts the lens is amazing
"It looks delicious, but I'm not sure if I should eat it." - Otter
any way to inpaint a video?
Nope,
I realized too late the horn is missing from the barbarian's helm at one point.... https://www.youtube.com/watch?v=YuXLULsjLqA&feature=youtu.be
The movie of the videogame Black Tiger - thanks to AI.
and now it's bothering me a lot
I mean it doesn’t bother me
it's too big of a glitch, kinda takes away from it
Only when you pointed it out
but it's such a hassle to redo the scene
I mean you don’t have to,
next time I'll definitely pay extra attention.... ugh.....
idk it's so shabby to see such a mistake
Idk how to tell you this but , try not to think too much about it, just have fun with it ,
if I feel I'm just making botched stuff, I lose my motivation 😤 I keep working on the videos because I think "these are turning out good"
incidentally, I need a new video idea now
Botched stuff is expected to happen especially when it’s involving ai video generation, that’s why trial and error is a thing,
I disagree. If you're patient and careful you can reduce the botched stuff so much the video actually seems "real". I'm annoyed when people just release ai videos full of botched glitches, they're just lazy
If the video generated doesn’t come out like what u had in mind, that’s where you re check how you prompt your video.
Lanterns
exactly! and I didn't notice omg....
You can use Gemini or ChatGpt as your prompt assistant
Grok's much better
but I don't use prompt assistants
I just need to pay attention... sigh
how'd I miss that friggin' horn
i mean , the prompt assistants just an extra tool to enhance your prompt.
they tend to turn YOUR idea into THEIRS. so no
I don’t really have much to say about how you see it. I do admire your dedication to making sure everything is purely your own creation, but at the end of the day, it’s not a competition. It’s about enjoying the process of creating and bringing your ideas to life 👌
I was thinking "mmmh maybe anime intro song about the evil catgirl d3mon"
I think they must've increased the censorship on depicting young characters. Or maybe the idea of evil catgirls in a circus offends the ai
......and a very strange thing just happened: replace it with "boy" and it has no problem making kids
dat's sexist