Started using it last night with my version of the man in the blue business suit prompt “an African-American woman driving a convertible flying car in a rural area”. There has been NO text to video AI model that has gotten remotely close… until Minimax. Wow what a difference! Thank you so much for promptly creating a prompt creator (I couldn’t resist). As a professional filmmaker I’m really looking forward to the future this thing, especially when it comes to character performances.
man, that is a GREAT test-- and I'm more than sure a lot of generators have botched pretty badly! Consistent Characters are getting close! We're ALMOST there!
I love it! That's one of my favorite parts about doing these videos: the "community spotlight" part-- I LOVE seeing how everyone can get these really unique outputs that I would never think of!
The tools to do "Filmmaking" are more accessible, but don't forget that the "Filmmaking" process, still been the same, and everyone need to learn that process, and that takes time, "It's not just blowing hot glass and make bottles and fine wine cups", the most important thing still being the creativity and vision of a human that knows how to use the right tools that will makes possible real nice new films/video content.
@@woltergeist9175 Yes, I am doing short stories with AI and proper editing tools, and are going amazing, I have about 25 years experience in Photography and Video/Filmmaking, I did about 9 short films already (with real equipment and people). I hope to share the AI Short Films soon, but it takes time, just like a regular Filmmaking workflow.
Thanks Tim, one of your best videos yet, and you built a gpt to help! Wow,! I already have it bookmarked. And yes, please make a video about camera movement and how it's motivated onscreen. I'd love to hear your thoughts on the topic. Cheers!
Great to hear! I'll pop that one on the "to do" list! I might roll the topic into a larger piece-- but yeah, clearly I think it'll be a good video to make!
Tim, thanks for the tutorial. I played around with MiniMaz a bit last week and then went back to Luma. Your prompt generator brought me back to MiniMax and if they come out with Image to Video, then I will must likely make MiniMax my main squeeze!
Your prompt builder is absolutely amazing, thank you!! Turns a vague thought into consistently great scenes. Used it for most of my latest video "Santa's here".
I have been using MiniMax for my next AI short film and to me it seems to be the most advanced AI video generator. I’m really excited to see how their image-to-video function works when released. I would also love MiniMax to introduce the character reference function that Vidu has. Anyways, this is a major step towards the possibility to actually create a movie - alone - in you basement (or perhaps a more comfortable setting of your choice). Great informative video, as always!
"The scene begins with the camera executing a smooth, wide pan across a massive spaceship hovering majestically above Earth. Shot with an IMAX camera, the spacecraft's design merges futuristic technology with Roman architectural grandeur-its colossal hull is embellished with marble columns, imposing arches, and intricate mosaics that glisten under the artificial lights. The camera glides effortlessly, capturing the hyper-detailed textures and craftsmanship of the ship’s exterior, blending extreme realism with cinematic HD photorealism. As the camera moves closer, it transitions to a slow zoom, revealing the ship's observation deck. Here, a line of Roman centurions stands at rigid attention, their polished armor reflecting the cold blue glow of the ship's lights. The atmosphere is tense and foreboding. The camera then centers on the centurion leader, who steps forward with commanding authority. In a dramatic motion, the leader raises a massive sword, its blade pulsing with a sinister red glow, vivid and threatening. The lens performs a tight, dynamic zoom on the leader’s sword as he points it menacingly toward Earth, specifically aiming at the Netherlands below. The shot is a masterpiece of award-winning cinematography, capturing the surreal fusion of ancient power and advanced technology in stunning ultra-high definition." can u improve that i want a scence with a space ship with angry space roman legionaires preparing to attack earth
@@ffffff7593 Your shot prompt is very detailed, which is good. However - the prompt asks for a lot of things to be shown in a 5-second video. I would cut your storytelling into separate shots - I would tell what you want to tell in about 3-5 different clips that you can then edit together. I would use detailed, but shorter prompts. It seems that MiniMax can get confused if your prompt is very long. I would also suggest to put the most relevant things at the start of the prompt and less relevant things after. Also - if you want consistent style in your clips and film, I would suggest prompting in the lighting conditions and color grading so that they would remain the same. I hope this helps!
@@Earthball_Productions I’ve created my short films with Kling and it’s awesome! I’m really excited about the update and can’t wait to see what quality I can produce with it for my next film :)
Another excellent vid. I've "paused" spending much time with AI vid because it's too wild, but Minimax is getting closer. Your GPT prompter is terrific!
Thank you for the Minimax Cinemotion Prompter; it's helping me create scenes that surpass anything I've produced with KlingAI and RunwayML. I would be thrilled if you developed a prompter for them too. Mastering the art of writing prompts for various programs has been challenging. I adore everything you're doing-please continue the excellent work. Your prompter is poised to become the next big thing. The minimax seems even better than Sora now that we have a great prompter.
@@TheoreticallyMedia I'm using it as much as I can to finish a medieval peasants' revolt short film I started when Haiper came out and was unlimited free.
When it comes to character consistency, I believe that, as long as the body frame of the person is close enough each output, the workaround would be just use a face swap app to keep the same person throughout the videos. Could you do something like that and make a video or tutorial on it? 🙏🏽
🙌 *Tim is tha man!* 1) Thanks so much for the gpt text gen. I'm using it and LOVE it. 2) Really appreciate all your advice/tips; very helpful. _Please keep sharing as your content is stellar!_ 🌟
I ABSOLUTELY LOVE THIS , as someone who wanna start a channel im here stock piling videos like this in media wallet so when i reading to begin my UA-cam journey ill be able to be prepared and revisit it in media wallet
I've been making Ai music and just started getting into learning about making music videos for my songs. This brings so many new ideas, gonna have me a lot of fun!
Minimax is awesome tools I have be using it for few days now, but not until any tool can have up to 1min video. All we do with all the available tools is to practice with them.
Thanks for that prompt generator, it works. I am hoping for minimax character consistency and an api so I can automate prompting and filing, or getting. But I will try your idea for consistency and try to use the ultra long prompt.
Lovely work breaking the document down for everyone, great explanation. It works a treat really impressed with how it interprets prompt commands. I have many a bunch of things with it too much fun
I think with a bit more work-- like, longer generations (6 seconds is a bit cramped), and Image to Video, it'll be pretty close to a shot at the crown. That said: right now, it is the BEST FREE video generator around. It's kind of insane that this is free.
Thank you so much for sharing the prompt GPT! That is super useful! I put a prompt into Minimax in English and I did not get what I was looking for at all 😂 Gonna be using the GPT going forward!
Just started playing with minimax/hailuoAI and it's astounding. I would love to see the ability to not only add character and set reference, but also to do iterative generations from prior videos, allowing you to refine and tweak your output. I'm sure it's just a question of time.
Indeed. One that is already doing that is Vidu. So, the technology is out there, it’s just a question of when it will be implemented. So much stuff exploding into so many directions!
I tried LUMA this week, and just now trying Minimax with some of my movie prompts adapted to their format but none of them so far can really replicate the beautiful face of Byrwho (the star of my 90 minute moonvalley ai text2video movie I am working on for Jan 2025). It's quite important you fall in love, and hence care for, the fate of Byrwho emotionally, that requires gorgeousness as well as realistic actions/reactions in the scene. But it's free and I've only done 4 prompts vs 500k prompts on moonvalley (now closed). LUMA does amazing car action sequences from my movie prompts however but the free plan is highly restrictive on LUMA and generations per dollar are real stingy, for the tinkerer/dabbler not world record smashing ambitious types like me.
Super Masterclass. It was hella MORBING. Excellent point about the Mandarin prompt. Gracias! Now off to make scenes from: "Bloodswords Are Go!" Directed by Sim Timmons.
Great overview, as per usual Tim, Really appreciate the prompt generator ..though I personally noticed that if I translate English to Italian, there is more detail and character/flair to the output…cheers.
Hey man!! Yeah- it's pretty impressive. I mean, like all AI video, it'll also give you some duds, but the fact that this level of video exists somewhere in the training data (my thought is: Big set of data and a good world model) means that Image to Video is really going to rock with this thing.
i love it.. its the main video ai generator that i use.. and i cant wait to see it in a couple of months time.. and looking forward to video from picture and as you said just 6 guys working on it.. wow
I legit was going to go over to Udio/Suno and have it generate a song with lyrics from that prompt! I thought it would have been a good cutaway gag, but-- also probably would have made half the audience fall asleep! haha
For sure, nothing sedating about you nor your content man. One of the few people that can still make me laugh, oh sh** have I gotten jaded? Welp. Thanks Tim for keeping us up to speed and forging the road. No one like you. Enter Guitar solo here.
I have been using it for about a week. It produces really good results, but not always true to the prompt. For each prompt I generate several clips and one or two will be what I asked for. I hope they make it do image to video and start and end images. I use a seed number to get fairly consistent scenes and characters.
Hi, Tim. I am doing the volume 2 of my graphic novel, "Blue Angel", using Flux and Mage. The image quality is excellent but it lacks consistency and variety. I usually generate an average of 30 images to get closer of what I need. Most of the generated images show an inexpressive character looking at the camera... I see Minimax as something similar. Amazing visuals, but if you can't have character consistency from one shot to another or if you can't generate the pose and attitude you want, it's just useless, just a toy. And I am not talking about lip sync (no pun intended). I would love to make my series using this tool, it would be much better than 3D, but I would trade some of the quality for consistency and pose variety.
I use a seed number to get fairly consistent scenes and characters with this. It's worth a try. Sometimes it adds the seed number as text on the video so if I like the clip I will use a watermark remover to get rid of it.
This is a really nice tool. I would love a paid 2k /4k render version of Minimax , also one thing I noticed is repeating same characters cars and other things is not that easy . But the results are good even at 720, you think we get a better more improved paid version of minimax or something like it soon with more options with regards to repeatability
Slight technical note (putting here because the algorithm likes engagement)… Mandarin is a spoken dialect of Chinese. The system you saw is written Chinese, and the character system is called Hanzi. So, in terms of taking about writing, you would simply say “Chinese”. (Or Hanzi to be precise.) I guess an equivalent is if you said the Latin alphabet was “English.”
Ahhh- Thank you so much! That makes sense. For some reason I got it in my head that Mandarin was the written version-- I know there is a lot of nuance in asian languages when it comes to modern communication (hence "Chinese Simplified" as an option)-- I'll dig into it more, as clearly I'm going to be using many more applications that originate from China in this AI craziness!
Oh, that's another one people confuse too: Chinese Simplified versus Chinese Traditional is something like using sans serif versus serif fonts. (It only applies to written and not spoken Chinese.) The simplified characters have fewer brush strokes and details.
It's pretty good! Admittedly, a little slow currently-- but not as bad as some I've seen at launch. I mean, I'm sure they're getting hammered pretty hard-- considering...free. But, you also can't beat that price!
Level Up Your AI Videos: Minimax Prompt Masterclass! Discover advanced techniques to optimize and enhance your AI-generated content with expert strategies and tips for maximizing impact. Unleash your creativity and take your videos to the next level with cutting-edge AI tools!
best way I'd describe it is terribly amazing, it still doesn't look what I'd consider to be good or usable yet but at the same time its amazing that it could create these from just a prompt, still one year off being an actual useful tool i think.......we patiently wait for sora to enter the chat
"Terribly Amazing" should be the official tagline of this channel! Haha, I love it. It's totally accurate. We're still early in all of this-- it gets better and better everyday, but for sure, we're the early pioneers of this technology. That's what I love about it.
Minimax’s greatest feat is weening me off my Kling inspiration points addiction…if their image to video is as good as their text to video then I’d say minimax is Ai video 2.5 and Luma, Kling and Runway are in the dirt.
Now that anyone can make cinematic shots...yes, would 100% love to see you give lessons on the overusage of camera movemen (one of the biggest and most common issues the ai filmmaking community has right now). How to use the language of film needs to be passed onto the new ai artists...like yesterday 🤣
0:45 MiniMax is run by SIXXXXXX people!? That's NUTTS. Either they're geniuses or this let's us know that AI video is just going to go hyperbolic once people start ramping up use and education and information of the technology becomes a bit more mainstream for more regular people to create custom models. I think it's both.
Ive been using it for 4 days now and its wild that its free and unlimited. It must be costing them over a million dollars a week to offer to the public, this has to be heavily subsidized by the CCP.
Hollywood vet here too-- 16 years in the business. To be totally fair, H'wood has been bleeding out from damage done to itself for awhile now. AI Video is just another cut. The Studios were too late to recognize the value of streaming early in, the whole film development process was "moneyballed" to death-- leading to bloated blockbusters no one cares about, inflated upsells like 3d, and a real lack of taking chances to farm interesting and new voices. I mean, wrap that all up, and coupled with the fact that "Hollywood" was always a myth (I mean, really-- isn't "hollywood" really Atlanta and Vancouver?), and the LA housing market is totally out of control? Yeah...It doesn't look good. That said: I think the future looks amazing for anyone who wants to create a film, and use AI to augment it to a high production value. That's the real story here!
Thank you! Man-- they have the WORST branding going on. I'll try to update it next time around. This is like when I kept called Udio "Ooooh-De-Oh" only to find out it is "You-Dee-Oh"
That's awesome, thanks for mentioning it, love your videos. I do have a question, how can I remove my past videos from the "mine" section in minimax after I render them? I don't want to keep a long list of videos I don't need. Thanks again, keep up the good work.
Yeah i know, i was one of the first that i start to try this model the same time one of you, youtubers talk about it. I created a short story about werewolf -rabbits and MAN HO MAN!!! i created even the intro O M G!!!! Is really amazing, so i created with Suno and Udio the song in Metallica Enter sandman style, with my virtual singer voice names Lexa , and then i just to put those scemes into cupcut without editing and the result is amazing. Really. I created more things, like alice in wonderland in claymation, miniature people walking on a giant room, some things don't understand, like a man behind a hirl cutting her hair, sometimes the man is forward, or is in another room , especially if you put a mirror in the scene, in another hand, understand Giant people but is really hard to create microscopic people, even using a macro photography in the prompt, to create my videos with microscopic people i expend a lot of time. But we are teally getting closer to consistent characters and more quality and more longer videos. But is really amazing, AW! i created the most wet fantasy of Marvel MCU fantasy fans. TOM CRUISE AS SUPERIOR IRON MAN!!! standing in a Post apocalyptic NY city. The result is amazing, bright white armor and golden armor with deeper or concerns expression, Tom watching looks like saying: Oh crap, we destroyed the whole city. 😂😂is really amazing. If you wanna see my superior iron man just ket me know. Ill send uou to your email and you can use it in the next videos. Just let me know Tim. And Make Hollywood great again with Minimax!!!😂🎉
@@TheoreticallyMedia I think it will be free for some time, like Ideogram used to be, but then Ideogram added subscriptions and started giving only 10 credits a day to free generations, but that's also good)
hi，tim，I would like to know, when comparing the video quality generated by Minmax and RUNWAY, without considering the time factor, which one produces better video quality? Among the many AI video generation tools available right now, which one offers the best results?
Hi thanks for your videos. How do you set up the character to video, not image to video, function in runway gen 3 turbo, as in vidu? To maintain the consistency of the character actors with the face? Vidu is very reliable
Yeah-- that totally looks like Jared Padalecki's long lost brother! At some point I have to go back and finish watching Supernatural. I think I went up to about Season 6? or what...like 23? haha
Not complaining. Minimax and Kling... They are the ones I use the most. And I also like to use chinese instead of english. Not because of prompt lenght, but I have a feeling it would be more accurate in some cases. Although, I'm never really sure if the translation is correct... :D But, my Claude translator project haven't failed, yet.
am I the only one who feels that giving my phone number to 'someone' in China is actually way too expensive... or is everyone here just buying burner phones to bypass any potential privacy or security concerns? If you can really use it without giving that away then I might feel brave and dip my toe in... Great video as always Tim.
So in their discord they mentioned they’re dropping the phone number thing. I think it’s just more common in China? Currently I don’t think there is any requirement to log in now though!
Could MiniMax be used to create a time lapse video, to create the illusion of a much longer video, then have DaVinci Resolve Studio fill in the missing frames?
Yeah, currently no way to remove the watermark. I suspect that'll come with the paid plan. Although, the one trick you can mess with is just doing the old "Faux Letterbox" black bars.
Can you share the Chinese Minimax documentation you show at 1:16? I can't find it on the internet and i want to try and train a Groq version of this ChatGPT (for people who are on free tier and hit the limit)
Thank You TIm, your video is awesome ! I was thinking about to made up a GPT myself and then I found you ! Merci beaucoup ! I tried Runaway and Kling before Minimax and had a lot of problems with consistent characters, but it may be because I ask for a non human one ? I wanna generate a cute little cloud in a 3D Pixar style and mixed up with some self made 3D animation... I let you know how that goes now that I can use your empowered Minimax Prompt GPT! Bonjour de France !
Impressive, though character inconsistency is definitely an issue which could be a significant limiting factor. I'm just trying to think of clever ways to mask this - perhaps degrading the video giving it a retro look, B&W and fast cuts in the hope viewers don't notice with each scene change. Dunno.
What did you use for more accurate translations for the Mandarin stuff ? ChatGPT? Claude? I'm studying Mandarin and wanna know what's the best LLM technology to use for it lol
So far I've been having the most success with ChatGPT. At least as far as I can tell-- Claude would likely be very good. Gemini for sure was SUPER wrong. I was really baffled and confused when I was first reading the translations...like, "what are they talking about?" But then trying it in ChatGPT, it was like: Oh...now everything makes sense! Google, man...they just can't get it together.
Thanks so much for the GPT link and the UPDATE... I have been killing myself to get the SMS message to go through for years... and then I just skipped to the 2nd green "Create button" and I was off to the races... Did you by any change create a GPT for Kling? 😀
Started using it last night with my version of the man in the blue business suit prompt “an African-American woman driving a convertible flying car in a rural area”. There has been NO text to video AI model that has gotten remotely close… until Minimax. Wow what a difference! Thank you so much for promptly creating a prompt creator (I couldn’t resist). As a professional filmmaker I’m really looking forward to the future this thing, especially when it comes to character performances.
man, that is a GREAT test-- and I'm more than sure a lot of generators have botched pretty badly!
Consistent Characters are getting close! We're ALMOST there!
@@TheoreticallyMedia Indeed!
This is why we love you, Tim! A Minimax Cinemotion Prompter! You're a legend, thank you!!!
Filmmaking just became accessible to everyone, excited to see all the original content that takes off 🔥
I love it! That's one of my favorite parts about doing these videos: the "community spotlight" part-- I LOVE seeing how everyone can get these really unique outputs that I would never think of!
It already was with phones. People made crap.
The tools to do "Filmmaking" are more accessible, but don't forget that the "Filmmaking" process, still been the same, and everyone need to learn that process, and that takes time, "It's not just blowing hot glass and make bottles and fine wine cups", the most important thing still being the creativity and vision of a human that knows how to use the right tools that will makes possible real nice new films/video content.
The first AI film that is uploaded to youtube and totally worth watching is right around the corner!
@@woltergeist9175 Yes, I am doing short stories with AI and proper editing tools, and are going amazing, I have about 25 years experience in Photography and Video/Filmmaking, I did about 9 short films already (with real equipment and people). I hope to share the AI Short Films soon, but it takes time, just like a regular Filmmaking workflow.
As a filmmaker I used prompts with a detailed description of camera movements, lights, props and direction of actors and I obtained amazing results.
Thanks Tim, one of your best videos yet, and you built a gpt to help! Wow,! I already have it bookmarked. And yes, please make a video about camera movement and how it's motivated onscreen. I'd love to hear your thoughts on the topic. Cheers!
Great to hear! I'll pop that one on the "to do" list! I might roll the topic into a larger piece-- but yeah, clearly I think it'll be a good video to make!
Tim, thanks for the tutorial. I played around with MiniMaz a bit last week and then went back to Luma. Your prompt generator brought me back to MiniMax and if they come out with Image to Video, then I will must likely make MiniMax my main squeeze!
Your prompt builder is absolutely amazing, thank you!! Turns a vague thought into consistently great scenes. Used it for most of my latest video "Santa's here".
Mini Max is blowing up fast!, im bout to generate as much as i can while its still surprisingly free
That's the key right there: Get to the buffet early-- when the dinner rush starts, the prices go up!!
...also, I eat dinner at like 4:30...hah.
I have been using MiniMax for my next AI short film and to me it seems to be the most advanced AI video generator. I’m really excited to see how their image-to-video function works when released. I would also love MiniMax to introduce the character reference function that Vidu has. Anyways, this is a major step towards the possibility to actually create a movie - alone - in you basement (or perhaps a more comfortable setting of your choice). Great informative video, as always!
It will good than vidu and kling,in image to video
"The scene begins with the camera executing a smooth, wide pan across a massive spaceship hovering majestically above Earth. Shot with an IMAX camera, the spacecraft's design merges futuristic technology with Roman architectural grandeur-its colossal hull is embellished with marble columns, imposing arches, and intricate mosaics that glisten under the artificial lights. The camera glides effortlessly, capturing the hyper-detailed textures and craftsmanship of the ship’s exterior, blending extreme realism with cinematic HD photorealism.
As the camera moves closer, it transitions to a slow zoom, revealing the ship's observation deck. Here, a line of Roman centurions stands at rigid attention, their polished armor reflecting the cold blue glow of the ship's lights. The atmosphere is tense and foreboding. The camera then centers on the centurion leader, who steps forward with commanding authority. In a dramatic motion, the leader raises a massive sword, its blade pulsing with a sinister red glow, vivid and threatening. The lens performs a tight, dynamic zoom on the leader’s sword as he points it menacingly toward Earth, specifically aiming at the Netherlands below. The shot is a masterpiece of award-winning cinematography, capturing the surreal fusion of ancient power and advanced technology in stunning ultra-high definition." can u improve that i want a scence with a space ship with angry space roman legionaires preparing to attack earth
@@ffffff7593 Your shot prompt is very detailed, which is good. However - the prompt asks for a lot of things to be shown in a 5-second video. I would cut your storytelling into separate shots - I would tell what you want to tell in about 3-5 different clips that you can then edit together. I would use detailed, but shorter prompts. It seems that MiniMax can get confused if your prompt is very long. I would also suggest to put the most relevant things at the start of the prompt and less relevant things after. Also - if you want consistent style in your clips and film, I would suggest prompting in the lighting conditions and color grading so that they would remain the same. I hope this helps!
Kling 1.5 is a rival
@@Earthball_Productions I’ve created my short films with Kling and it’s awesome! I’m really excited about the update and can’t wait to see what quality I can produce with it for my next film :)
Another excellent vid. I've "paused" spending much time with AI vid because it's too wild, but Minimax is getting closer. Your GPT prompter is terrific!
Minimax is the superhero we needed
Yo!! This is gold Tim! Thanks for sharing! I gave it a shot last night and upload my first video on my birthday 😅
Happy Birthday!!
Thank you for the Minimax Cinemotion Prompter; it's helping me create scenes that surpass anything I've produced with KlingAI and RunwayML. I would be thrilled if you developed a prompter for them too. Mastering the art of writing prompts for various programs has been challenging. I adore everything you're doing-please continue the excellent work. Your prompter is poised to become the next big thing. The minimax seems even better than Sora now that we have a great prompter.
We've come such a long way just a year! Thanks for the gpt! Brilliant work!
Thanks for the Minimax gpt prompter!! loving Minimax, best free option to ever come out so far.
Excellent to hear! and yeah-- hit that buffet while it's still free!!
@@TheoreticallyMedia I'm using it as much as I can to finish a medieval peasants' revolt short film I started when Haiper came out and was unlimited free.
Thank you for the GPT Prompter. Just tried it and got an excellent result.
Your GPT prompt builder is excellent, thank you! Very helpful indeed.
When it comes to character consistency, I believe that, as long as the body frame of the person is close enough each output, the workaround would be just use a face swap app to keep the same person throughout the videos. Could you do something like that and make a video or tutorial on it? 🙏🏽
🙌 *Tim is tha man!* 1) Thanks so much for the gpt text gen. I'm using it and LOVE it. 2) Really appreciate all your advice/tips; very helpful. _Please keep sharing as your content is stellar!_ 🌟
btw Tim, for future ref, I noticed the tool is actually called *Conch AI* and MiniMax is their model. _Cheers!_ 🥂
been using this since day one and it's results are so inconsistent but also sooooo good when it gets it right
Someone in the comments called it “terrible and amazing”
Haha, I can’t think of a better description!
@@TheoreticallyMedia it truly is. its results vary .. very much
Minimax is by far the best of all the video generators out there, and it's free.
still stunned that we're getting this for free! Mind you, for now-- I know that shoe will drop at some point, but for now: let's enjoy the ride!!
Aaand as always, Tim is here to the rescue 🤟 Planning to try Minimax tomorrow, the guide will come in handy. Thanks a lot Tim!
I ABSOLUTELY LOVE THIS , as someone who wanna start a channel im here stock piling videos like this in media wallet so when i reading to begin my UA-cam journey ill be able to be prepared and revisit it in media wallet
I've been making Ai music and just started getting into learning about making music videos for my songs. This brings so many new ideas, gonna have me a lot of fun!
Minimax is awesome tools I have be using it for few days now, but not until any tool can have up to 1min video. All we do with all the available tools is to practice with them.
Thanks for that prompt generator, it works. I am hoping for minimax character consistency and an api so I can automate prompting and filing, or getting. But I will try your idea for consistency and try to use the ultra long prompt.
Lovely work breaking the document down for everyone, great explanation. It works a treat really impressed with how it interprets prompt commands. I have many a bunch of things with it too much fun
It is the best video generator, even thought it is text to video only ,! Extraordinary results
I think with a bit more work-- like, longer generations (6 seconds is a bit cramped), and Image to Video, it'll be pretty close to a shot at the crown. That said: right now, it is the BEST FREE video generator around. It's kind of insane that this is free.
I totally agree,thanks for the GPT@@TheoreticallyMedia
thank you much for your great work. watch your stuff a lot since months or years and your information style and help is always a blast!
Awesome breakdown. And thanks for sharing the prompt builder!
This tutorial is amazing! Thank u so much for supporting Hailuo Ai!
Appreciate that!! Keep up the stellar work on your end!
Thanks so much Tim! The prompt generator is gold!
Thank you so much for sharing the prompt GPT! That is super useful! I put a prompt into Minimax in English and I did not get what I was looking for at all 😂 Gonna be using the GPT going forward!
Just started playing with minimax/hailuoAI and it's astounding. I would love to see the ability to not only add character and set reference, but also to do iterative generations from prior videos, allowing you to refine and tweak your output. I'm sure it's just a question of time.
Indeed. One that is already doing that is Vidu. So, the technology is out there, it’s just a question of when it will be implemented.
So much stuff exploding into so many directions!
I tried LUMA this week, and just now trying Minimax with some of my movie prompts adapted to their format but none of them so far can really replicate the beautiful face of Byrwho (the star of my 90 minute moonvalley ai text2video movie I am working on for Jan 2025). It's quite important you fall in love, and hence care for, the fate of Byrwho emotionally, that requires gorgeousness as well as realistic actions/reactions in the scene. But it's free and I've only done 4 prompts vs 500k prompts on moonvalley (now closed). LUMA does amazing car action sequences from my movie prompts however but the free plan is highly restrictive on LUMA and generations per dollar are real stingy, for the tinkerer/dabbler not world record smashing ambitious types like me.
Super Masterclass. It was hella MORBING. Excellent point about the Mandarin prompt. Gracias! Now off to make scenes from: "Bloodswords Are Go!" Directed by Sim Timmons.
Great overview, as per usual Tim, Really appreciate the prompt generator ..though I personally noticed that if I translate English to Italian, there is more detail and character/flair to the output…cheers.
Respect for Tarkovsky shout-out😁👍
minimax is almost as good as the video you made about it explaining how it works. great job man
Wow I had no idea you had such control over the outputs. Especially detailing purpose of the cameras movement.
Hey man!! Yeah- it's pretty impressive. I mean, like all AI video, it'll also give you some duds, but the fact that this level of video exists somewhere in the training data (my thought is: Big set of data and a good world model) means that Image to Video is really going to rock with this thing.
Definitely! MiniMax is so far the best program available (to the public, that is) at following detailed prompts for creating video.
Great prompting. Love your channel!
i love it.. its the main video ai generator that i use.. and i cant wait to see it in a couple of months time.. and looking forward to video from picture and as you said just 6 guys working on it.. wow
Really great info, thanks. I’m enjoying using MiniMax for animation. Made a few nice ones. It works really well at realising my ideas.
Radio head music LOL nice one :) Thanks Tim!
I legit was going to go over to Udio/Suno and have it generate a song with lyrics from that prompt! I thought it would have been a good cutaway gag, but-- also probably would have made half the audience fall asleep! haha
For sure, nothing sedating about you nor your content man. One of the few people that can still make me laugh, oh sh** have I gotten jaded? Welp. Thanks Tim for keeping us up to speed and forging the road. No one like you. Enter Guitar solo here.
I have been using it for about a week. It produces really good results, but not always true to the prompt. For each prompt I generate several clips and one or two will be what I asked for. I hope they make it do image to video and start and end images. I use a seed number to get fairly consistent scenes and characters.
Can you explain what the seed number is? Thanks
I really hope a full movie of "The Hobbit Sheriff" hits our screens soon. The love interest will of course be Daniëlle van de Donk
Holy crap, I need to make this movie now! The Magnificent Seven Rings?
Hi, Tim. I am doing the volume 2 of my graphic novel, "Blue Angel", using Flux and Mage. The image quality is excellent but it lacks consistency and variety. I usually generate an average of 30 images to get closer of what I need. Most of the generated images show an inexpressive character looking at the camera...
I see Minimax as something similar. Amazing visuals, but if you can't have character consistency from one shot to another or if you can't generate the pose and attitude you want, it's just useless, just a toy. And I am not talking about lip sync (no pun intended).
I would love to make my series using this tool, it would be much better than 3D, but I would trade some of the quality for consistency and pose variety.
I use a seed number to get fairly consistent scenes and characters with this. It's worth a try. Sometimes it adds the seed number as text on the video so if I like the clip I will use a watermark remover to get rid of it.
@@TheKarenRoach Thank you. They will probably create the seed control as a feature soon.
This is a really nice tool. I would love a paid 2k /4k render version of Minimax , also one thing I noticed is repeating same characters cars and other things is not that easy . But the results are good even at 720, you think we get a better more improved paid version of minimax or something like it soon with more options with regards to repeatability
Great guide! thanks so much. I will definitely put some of these advanced prompts into making my next music video
🤗🤗🤗🤗🤗🤗
already made a whole music video with this! crazy stuff
Bless you for making that gbt to help us
Oh, it's such a time saver. I mean, that's what AI is for, right?!
Thank you for the GPT , only 6 people and this quality??? OMG
As always, great info!
Awesome job, thanks
Slight technical note (putting here because the algorithm likes engagement)… Mandarin is a spoken dialect of Chinese. The system you saw is written Chinese, and the character system is called Hanzi. So, in terms of taking about writing, you would simply say “Chinese”. (Or Hanzi to be precise.) I guess an equivalent is if you said the Latin alphabet was “English.”
Ahhh- Thank you so much! That makes sense. For some reason I got it in my head that Mandarin was the written version-- I know there is a lot of nuance in asian languages when it comes to modern communication (hence "Chinese Simplified" as an option)-- I'll dig into it more, as clearly I'm going to be using many more applications that originate from China in this AI craziness!
Oh, that's another one people confuse too: Chinese Simplified versus Chinese Traditional is something like using sans serif versus serif fonts. (It only applies to written and not spoken Chinese.) The simplified characters have fewer brush strokes and details.
Minimax looks really cool.
"Hit the X Button".
; )
It's pretty good! Admittedly, a little slow currently-- but not as bad as some I've seen at launch. I mean, I'm sure they're getting hammered pretty hard-- considering...free.
But, you also can't beat that price!
Level Up Your AI Videos: Minimax Prompt Masterclass! Discover advanced techniques to optimize and enhance your AI-generated content with expert strategies and tips for maximizing impact. Unleash your creativity and take your videos to the next level with cutting-edge AI tools!
great video mate! thanks for the gpt :)
Thank you for all the clarifications, hero Theoreticallymedia
best way I'd describe it is terribly amazing, it still doesn't look what I'd consider to be good or usable yet but at the same time its amazing that it could create these from just a prompt, still one year off being an actual useful tool i think.......we patiently wait for sora to enter the chat
although i did think the Claymation one was pretty good
"Terribly Amazing" should be the official tagline of this channel! Haha, I love it. It's totally accurate.
We're still early in all of this-- it gets better and better everyday, but for sure, we're the early pioneers of this technology. That's what I love about it.
Minimax’s greatest feat is weening me off my Kling inspiration points addiction…if their image to video is as good as their text to video then I’d say minimax is Ai video 2.5 and Luma, Kling and Runway are in the dirt.
Now that anyone can make cinematic shots...yes, would 100% love to see you give lessons on the overusage of camera movemen (one of the biggest and most common issues the ai filmmaking community has right now). How to use the language of film needs to be passed onto the new ai artists...like yesterday 🤣
I was blown away, free and fast. Im scared to imagine what it will make if i pay for it 😏
I love your humor.
0:45 MiniMax is run by SIXXXXXX people!? That's NUTTS. Either they're geniuses or this let's us know that AI video is just going to go hyperbolic once people start ramping up use and education and information of the technology becomes a bit more mainstream for more regular people to create custom models. I think it's both.
Ive been using it for 4 days now and its wild that its free and unlimited. It must be costing them over a million dollars a week to offer to the public, this has to be heavily subsidized by the CCP.
Top dollar, Tim. Thank you very much.
Thanks for the ChatGPT Filter!
As a veteran film maker, Hollywood is done. Add AI paintbrush and re-rendering precision, it's game over.
Hollywood vet here too-- 16 years in the business. To be totally fair, H'wood has been bleeding out from damage done to itself for awhile now. AI Video is just another cut.
The Studios were too late to recognize the value of streaming early in, the whole film development process was "moneyballed" to death-- leading to bloated blockbusters no one cares about, inflated upsells like 3d, and a real lack of taking chances to farm interesting and new voices.
I mean, wrap that all up, and coupled with the fact that "Hollywood" was always a myth (I mean, really-- isn't "hollywood" really Atlanta and Vancouver?), and the LA housing market is totally out of control?
Yeah...It doesn't look good.
That said: I think the future looks amazing for anyone who wants to create a film, and use AI to augment it to a high production value.
That's the real story here!
@@TheoreticallyMedia This is very insightful! I'm glad I found your channel! 🙏🏻🙂
video on camera move motivation would be interesting
Very interesting. Thank you.
Always brilliant
It's actually Conch ("Hailuo") Video. Minimax is the company.
Thank you! Man-- they have the WORST branding going on. I'll try to update it next time around. This is like when I kept called Udio "Ooooh-De-Oh" only to find out it is "You-Dee-Oh"
thx for the gpt hermano. looking forward to playing with it. will report back
That's awesome, thanks for mentioning it, love your videos. I do have a question, how can I remove my past videos from the "mine" section in minimax after I render them? I don't want to keep a long list of videos I don't need.
Thanks again, keep up the good work.
+1 for film school prompts
Fantastic!! onto the list it goes!
As someone who loves generative art, but has zero filmmaking background, I'd love to see this too.
Just remember the almost always true statement of "If it's free, you're the product."
Love your content! Curious about fashion…what’s your favorite brand for your awesome black t-shirts? 4:26
Yeah i know, i was one of the first that i start to try this model the same time one of you, youtubers talk about it.
I created a short story about werewolf -rabbits and MAN HO MAN!!! i created even the intro O M G!!!! Is really amazing, so i created with Suno and Udio the song in Metallica Enter sandman style, with my virtual singer voice names Lexa , and then i just to put those scemes into cupcut without editing and the result is amazing. Really. I created more things, like alice in wonderland in claymation, miniature people walking on a giant room, some things don't understand, like a man behind a hirl cutting her hair, sometimes the man is forward, or is in another room , especially if you put a mirror in the scene, in another hand, understand Giant people but is really hard to create microscopic people, even using a macro photography in the prompt, to create my videos with microscopic people i expend a lot of time. But we are teally getting closer to consistent characters and more quality and more longer videos. But is really amazing, AW! i created the most wet fantasy of Marvel MCU fantasy fans. TOM CRUISE AS SUPERIOR IRON MAN!!! standing in a Post apocalyptic NY city. The result is amazing, bright white armor and golden armor with deeper or concerns expression, Tom watching looks like saying: Oh crap, we destroyed the whole city. 😂😂is really amazing. If you wanna see my superior iron man just ket me know. Ill send uou to your email and you can use it in the next videos. Just let me know Tim. And Make Hollywood great again with Minimax!!!😂🎉
MiniMax is a good AI tool, but it's a pity that you can't generate video from photos, but I think they will add it.
Sometime in the next week, from what they're saying. It'll be here before you know it! (and hopefully it remains free!)
@@TheoreticallyMedia I think it will be free for some time, like Ideogram used to be, but then Ideogram added subscriptions and started giving only 10 credits a day to free generations, but that's also good)
he said it multiple times that they are soon
Does your GPT function in other paltforms like runway/kling?
Amazing.
hi，tim，I would like to know, when comparing the video quality generated by Minmax and RUNWAY, without considering the time factor, which one produces better video quality? Among the many AI video generation tools available right now, which one offers the best results?
LOTR + cowboys = Men of the West 🤠⚔️
Merci pour le prompt
Awsome content Tim!
Hi thanks for your videos. How do you set up the character to video, not image to video, function in runway gen 3 turbo, as in vidu? To maintain the consistency of the character actors with the face? Vidu is very reliable
i love this thank you
11:11 This is definitely Sam from Supernatural
Yeah-- that totally looks like Jared Padalecki's long lost brother!
At some point I have to go back and finish watching Supernatural. I think I went up to about Season 6? or what...like 23? haha
Not complaining. Minimax and Kling... They are the ones I use the most. And I also like to use chinese instead of english. Not because of prompt lenght, but I have a feeling it would be more accurate in some cases. Although, I'm never really sure if the translation is correct... :D But, my Claude translator project haven't failed, yet.
am I the only one who feels that giving my phone number to 'someone' in China is actually way too expensive... or is everyone here just buying burner phones to bypass any potential privacy or security concerns? If you can really use it without giving that away then I might feel brave and dip my toe in...
Great video as always Tim.
So in their discord they mentioned they’re dropping the phone number thing. I think it’s just more common in China?
Currently I don’t think there is any requirement to log in now though!
@@TheoreticallyMedia by the way, the GPT prompt generator was pretty stellar!
I didn't give a phone number
Could MiniMax be used to create a time lapse video, to create the illusion of a much longer video, then have DaVinci Resolve Studio fill in the missing frames?
Is it possible to use this in youtube videos without the watermark? Or do we need to wait until they have some sort of subscription?
Yeah, currently no way to remove the watermark. I suspect that'll come with the paid plan. Although, the one trick you can mess with is just doing the old "Faux Letterbox" black bars.
link?
Can you share the Chinese Minimax documentation you show at 1:16? I can't find it on the internet and i want to try and train a Groq version of this ChatGPT (for people who are on free tier and hit the limit)
Thank You TIm, your video is awesome ! I was thinking about to made up a GPT myself and then I found you ! Merci beaucoup ! I tried Runaway and Kling before Minimax and had a lot of problems with consistent characters, but it may be because I ask for a non human one ? I wanna generate a cute little cloud in a 3D Pixar style and mixed up with some self made 3D animation... I let you know how that goes now that I can use your empowered Minimax Prompt GPT! Bonjour de France !
Impressive, though character inconsistency is definitely an issue which could be a significant limiting factor. I'm just trying to think of clever ways to mask this - perhaps degrading the video giving it a retro look, B&W and fast cuts in the hope viewers don't notice with each scene change. Dunno.
What did you use for more accurate translations for the Mandarin stuff ? ChatGPT? Claude? I'm studying Mandarin and wanna know what's the best LLM technology to use for it lol
So far I've been having the most success with ChatGPT. At least as far as I can tell-- Claude would likely be very good. Gemini for sure was SUPER wrong. I was really baffled and confused when I was first reading the translations...like, "what are they talking about?"
But then trying it in ChatGPT, it was like: Oh...now everything makes sense!
Google, man...they just can't get it together.
Is anybody else getting 30 ads in this 13 min video?!!! Wow, the highest ever on a yt vid I've watched
Thanks so much for the GPT link and the UPDATE... I have been killing myself to get the SMS message to go through for years... and then I just skipped to the 2nd green "Create button" and I was off to the races... Did you by any change create a GPT for Kling? 😀
Do you think the image to video version is going to be free or are they going to become expensive as other platforms?😅