A little tip when generating images in Midjourney to make them more consistent in style so they don’t just look like a bunch of random clips is to use (style reference) -sref at the end of your prompt followed by an image URL to an art style of your choice.
Fully agree - and probably a high s and si values would fit this Context of “let’s take random lyrics and put together something nice” --sw 600 --s 300 maybe (once you have a style reference you like)
That was creepy, amazing, scary, annoying, awe inspiring, depressing and exciting all at once! This is going to be an amazing future. I have played with Suno but that luma image to videos is outstanding! Bravo Matt!
True, but it works even better the opposite: The image comes 1 frame after the beat. Because light (the video) travels much faster than sound (the beat). So in our brain the sound and light will be processed about the same time. The best way to describe it is when you watch fireworks. It always looks like the sound is late. I used to do it the way you said until I saw a Michel Gondry docummentary and it all makes sense
@@1individeoI know what you mean but the difference is here the music in the video and sound isn’t actually related to the image like fireworks are and the distance from the screen to your eye is negligible with the sound as they are so close to not make a difference. Seems like it could be a slow processing issue asking frames to change on the beat so you have to offset the timing of the beat to the frame. A bit like some old films with foley artists always seemed to dub the effects too late after the visual like a door closing or footsteps but most people tend to not notice
upl;oading a 1 to 4 bar audio clip from the song to the video generator will help sync the motion of the clips to the audio . also resolve. has a function to fomulate the time line to the tempo
It's still not quite where I'm convinced the whole time, but it's interesting how much I catch myself saying it's stock footage. Like you say, this is the worst it's going to get. Super excited to see where it goes from here. You should do one of these every 6 months or annually.
But why? AI generations take never before seen amounts of water for their servers. And they steal from the work of non consenting artists who've put so much time in their lives to perfect. Is it really only for personal entertainment or what? There's a growing distaste amongst even common folk against AI. The third most trending video on YT recently pointed out how AI is ruining the internet. And their demographic lies between teenagers and young adults. You won't earn any monetary return from doing this anyway. And you won't get hired considering the data regarding the employment of AI prompt bros. I understand it may seem cool, but everything in the world is just saying not to.
Hello, I’m an electronic musician for the last 30 years and sound engineer. I can tell 100% of the time it’s AI. I know that the average Joe’s ear can’t tell but once you know how to spot it, it’s game over. It just can’t handle certain frequencies and sample rate properly. Maybe that will change but it will take a while. But most importantly, making AI music takes away the most enjoyable part of music which is to make it, called the creative journey. It can take months to finish a song and during that time, you learn a lot, both on yourself and on music itself.
I come at it from a different perspective. I'm a tone deaf, uncoordinated, self conscious music lover. I have millions of ideas but no way to express them. I'm not looking to make a video or song to fool anyone or take money from deserving artists such as yourself, but to enjoy a different creative journey. There is a lot of thought, as well as trial and error, that goes into figuring out the prompts to create what's in your mind. Electronic musicians got some of the same arguments used against them way back as well as now. I see it as another tool of expression, and used creatively, can make art. Keep making tunes my friend.
@@SPAMDAGGER22 i use ai for generating ideas and I think it’s excellent at it. When I work with clients with genres I’m not the best at, ai is a great tool. My comment was directed at the click bait titled « you can’t even tell it’s AI » which is snake oil marketing that makes me roll my eyes. Ai is really solid at certain things but I think we calm down the horses on saying it fools people. For non musicians though, I think it’s amazing entertainment for creating quick fix music.
I never made music but I can always tell when what I hear is AI generated. This is an easy one, because it’s the most common giveaway: the song has a beat but no rhythm. And the lyrics are just phrases, they don’t make any sense. It’s basically a song that the pet shop boys would make if they were brain dead and couldn’t sing. So we’ve got machines making crappy music for us now. As if there wasn’t plenty of people doing that already.
I once used 3 years to make a beat 1 year to write the lyrics and ngl I love this song, Im not gonna lie I understand where you are coming from however I will say I actually like the song Matt made however I still enjoy making my own music
Wow! 5 or 10 years ago that video would have cost $ Millions! I helped shoot a 30 second commercial in Tahiti with a crew of 50 guys and a helicopter. It took 6 weeks. Imagine the cost back then.
WOW! I'm an amateur videographer, and I'm completely blown away by what you were able to accomplish in just two days. AMAZING! I'd like to see a follow-up video about the legal aspects of monetizing these videos and how the law may vary depending on the amount of creative human input. For example, using an AI reference photo versus a photo taken by a human. Keep up the great work! 😊 Thanks.
So true, it's unreal! It's crazy how a lot of people have started to say that AI hasn't really gotten better in the past 6 months. I genuinely don't understand it.
The technology is amazing, but it's NOT creating anything, but it's not creating in a way a human does... EVERYTHING is stolen! Yes humans can plagiarise, but your are not actually taking the actual work someone has made and twisting it into something new... All AI is theft!
@@christianjensen952it literally improved vastly in 6 months to the point I’m glad I didn’t really mess with it then. My patience and frustration would have turned me off to the point I would have missed out by the time I came back to it to only find out not only has it improved. But some of my drafts were improved also without my permission or knowledge to the point one was almost finished. I was impressed with what it did, ( not my vision but impressed compared to months ago) but WHY did it, and if it was skimming over my projects and even building in them whos to say my ideas aren’t being floated around out there somewhere before I give the say so. Can anyone explain to me what happened? And how to prevent this in the future. I’m talking about video art, I don’t think I want AI anywhere near my music
Matt- this is one of your best videos yet! It takes the viewer step-by-step through the process in a super-clear and understandable way. Your skills as a presenter are first class, and it’s clear that you are having fun showing us how all these new tools come together to create better and better results. I love that the Luma change to extend a video happened overnight, and you just folded that change right into your current video on the fly! Thanks for all you’ve given us so far. Every time I see a new video from you, I drop everything for the chance view and learn from you. Rock on!
I agree about Udio vs. Suno. I much prefer Suno, although I wish their sound quality was better (the resolution of the audio and file itself). In fact, I just used Suno to create extra verses to an Enya song over on my channel 📺
Introduction and Revisiting AI-Generated Music Video - 00:00:00 Choosing AI Tools for Music and Video Creation - 00:01:15 Using Suno for Music Generation - 00:02:20 Improving AI Video Generation with Dream Machine - 00:02:53 Generating Images with Mid Journey - 00:03:23 Creating the Song with Suno - 00:04:58 Generating Dance Scenes with Mid Journey and Dream Machine - 00:07:14 Generating Specific Scene Images with Mid Journey - 00:11:01 Converting Images to Videos with Dream Machine - 00:12:26 Compiling and Editing the Music Video - 00:16:13 Finalizing the Video and Premiere of "Binary Dreams" - 00:22:13 Summary and Reflection on the Process - 00:24:35 Future AI Video Tools and Upcoming Tutorials - 00:26:09 Outro and Call to Action - 00:27:06
My channel is waiting for something like this I have over 200 ai songs . Waiting on the tech to improve at a later rate and with longer videos .Even with Kling it just takes too long
This is fantastic! Can't be long before someone puts all those steps into one AI tool and then we create a full music video with just one prompt though. But great work in the meantime!!
Very impressive, Matt. I just completed my first AI music video in 75 hours :(. A lot of learning, especially in the area of prompting. I learned a lot from you and others who have shared their experience. Trying my next one now and already hitting some rendering walls. I've made and recorded the music a few years ago. It's amazing what's happening with AI. I can't believe it. Thanks, Man!
So 3 things. 1. Honestly pretty cool video and song! 2. I think the only clip you overused was the dancing people. Good to put the dancing people in where you did, but maybe you could have done small variations of that image? 3. Canva is not great for video editing, but I do love the Beat Synch feature! Which I never found on da vinci. So for "simple" edits, beat synch in there might make creating a video like this a 1000x easier :P Cause it's like done in one click. Good stuff again Matt!
Thanks for the tips. Not diminish your music-video project. There were a lot of good tips and tricks in that. But just had some question about the AI tools. Is it possible with any software today, to just write a prompt, and it will generate a photorealistic movie or animation, and in the same software also add music, voiceover, sound-effects etc.? So you have less slideshow effect in the movie, but a somewhat cohesion (as a whole, and not by parts) movie or music-video? Thanks from Krusty in Denmark
The issue that I have always found is that AI music generators tend to blend the music and the lyrics. like the voice is coming out of the instruments or something
I agree, Udio is like working with a real bar band, which I did on a local TV show for years. Suno is like working with the gods of music or something, it's even more amazing than Udio.
I want to create the whole song in Suno, then run it through udio to make the vocals more realistic. Not possible currently (as far as I'm aware), but that's what I want. I also want Suno to allow me to select a specific part of the song where I want a change.
I created like hundreds of music with both of those tools (kinda addicted to it :) I use the free version, I must say, both have pros and cons. Suno: More polish/advanced/trained AI.. that create up to 4 minutes of songs very fast and most of the the it captures the prompt (genres, instruments etc.) and it can create songs with any language, any alphabet out there (not even google can do) but it is strictly trained not to make real person voices, people can easily understand it is AI generated. Udio: Rather new tool, thus less trained AI.. it creates 30 second songs, it can be remixed/extended etc. but it takes more time than Suno and extensions are not so good (there are a lot of advanced settings/tunes, so playing with those doesnt help much) The good thing is, even though singer/musician names are forbidden to use at prompts, it can create real people voices or copy real musics/melodies.. to some outputs are sh*t, some outputs are much better than Suno, like a real song, and real singer singing a real song. But creating imitating people might lead to legal copyright issues.. I don't use auto generated lyrics, i use ChatGPT 4o for lyrics and I changed the way i like, then i use Suno for the music. Then, I use Capcut for the lyrics as a subtitled lyrics video. (I don't need real video or Dream machine etc :)
Suno (Premium) has a new feature, you can add your music, or sound to extend... did anyone tried a part of real song to extend or add new lyrics to an existed music?
Suno is very fast, and convenient but the quality is sub par compared to Udio. If you can't hear the difference you need to check your ears dude. Suno to me is an idea machine, Udio can at times create something nearly indistinguishable from a broadcast ready song. I have been making Contemporary Country with it and even fooled our local Country radio guy in a city of 400k people below the mason dixon line. Suno isn't capable of that. I use both, but trust me when I say Udio is the definitive one for quality. If you can't tell, you most likely don't have a musicians ear, just as many musicians don't have an editors eye. Is what it is.
Can I be able to use this to create a real song and upload to all streaming platforms (Spotify, audiomack, apple music and many more) without having any issues?
It may not replace it as of this moment. But he showed the progress from one year to the next. All jobs are going to be impacted and eventually replaced by AI. At this point, there are opportunities to enhance workflows but also some low-end jobs are eliminated.
oops...the progress rate is why you would want to think otherwise. LLM's from q4 2025 based on Nvidia's blackwell gpu's WILL outgun human cinematographers.
@@paulmuriithi9195 outgun them shooting what? AI will replace alot of things but personalities and charisma in a real human are gonna be hard to replace. If their is a movie with Ana de Armis, Brad Pitt and Ryan Gosling it is gonna need a real camera
@@williamparrish2436 people using tools for business are different then people enjoying the work of someone. I dont think cinematic movies have much of a threat, i think stock footage is the first victim
I've stopped using MJ -- I resubscribed lately, and instantly felt it was a mistake... The realism is nice, but... Ultimately, I just use Stable Diffusion - tons of models that can give great realism. Better speed, control, etc. I always enjoy seeing you doing the creative stuff. This is great.
IDK, MJ always takes several attempts for me. What others have you had success with? Realistic GPT has gotten me some good results but it doesn't have very many options.
Leonardo ai is constantly improving itself, seems images are just as good at MJ and no discord to deal with either. I couldn't handle what was chaos to me at MJ just trying to find my images after generating.
Do you have a video on how to up resolution a video? I have a bunch of old videotapes I confrated into digital video and old cell phone resolution sized videos I would love to upsize.... So I can enjoy him on my high definition TV set. Thanks in advance 😊
@@hombacom But why? AI generations take never before seen amounts of water for their servers. And they steal from the work of non consenting artists who've put so much time in their lives to perfect. Is it really only for personal entertainment or what? There's a growing distaste amongst even common folk against AI. The third most trending video on YT recently pointed out how AI is ruining the internet. And their demographic lies between teenagers and young adults. You won't earn any monetary return from doing this anyway. And you won't get hired considering the data regarding the employment of AI prompt bros. I understand it may seem cool, but everything in the world is just saying not to.
Amazing. This is incredible progress in one year. And your project really shows off the capabilities of Luma Labs video generator. Give it say 1-3 years, (who knows, maybe less), there will be a video generation model that takes music (and maybe a few images) as input and just does the whole process for you. Without needing software to automate a process -- the ML model takes the music and converts directly to video.
The text prompt on Luma doesn't seem to do much when you upload an image as reference for the video. Like I'll upload a picture and then I'll say "the camera zooms out" and it just ignores it.
I've just realized that next Multi-MLN Box-Office Level Movies are going to be made in any single spacious green-coated room with only couple personas, that'll play all the characters interactions, and the AI will tale on all the rest: All voice switching, All characters face and outfit replacements, all the landscape shots, even all the camera movements. And All That is most likely ahead right this year.
in 3 years you will just " tonight ill ask my home assistant to generate me a movie which is 2 hours long, and in this and that genre with a male/female lead set in this setting." And then the AI in your home assistant will just spin up your computer and it will poop out a movie for you to watch completely custom to your wishes. no more need for holywood. AT ALL.
@@Triiiwar It's going to bring things to the point, when today's MultiMlnDlr tier movies will be free to make, just for the sake of imbedding promotion and product marketing. Then to the point, where generative passthrough of the highest quality content [ by all means ] will be exceeding consumption capacity in every given moment. Let alone all existing content of all quality levels. There's already no Hollywood since the Netflix have spreaded the cult of a low IQ content production. Only Tom Cruise & Keanu are still making up some. Those guys are -60+ y old. Hardly imaginable how generations that've bought into Netflix and Tok will simply be creating children, not even mentioning Families, Good movies are out of discussion at all in here. We're heading towards the times, when robo-chickens will be nesting robo-eggs and no one will be even thinking of asking "what came first". Validity of humanity will be questioned big time.
it all will be done by AI somewhere in the cloud. The time of artists is gone. Probably single creator will rent images from stock image sites or get personal image.
@@genusbit4172 The age of recording artists is gone. Live performances and people with actual real talent who make painting and art live will live to see another day!
I actually think you did an amazing job with this song, and video. I understand that you want the AI music video to boost this video, but I don't think it would hurt to also add maybe an unlisted link to the music video itself. This way people still have to come here to watch the video, but if they want to share the video itself with others (to fruther advertise the channel) they can do it with the unlisted link.
Matt, If you still have the original file please do a cross fade or zoom transition between the last tv screen and the matrix type clip. I promise it will look so much better.
It turned out well, only one aspect is still missing: a music video has to look like a unified whole, not like a series of random scenes that fit the lyrics but don't really relate to each other.
Sora has a fature that combine two scenes/videos and make a seamingless transition.. for example they combined the drone shot and chistmas cakes.. or drone shot and underater shot etc... that kind of music video looks so much better with different scenes are transiting seaminglessly.
Fantastic job, keep up the great work Matt. Very impressive end result, I know that process can take a very long time even with using AI., as I do that myself. I create music videos, & write my own songs that I work along side Suno with to make my music, and vocals for my songs. I recall once it took me 15 hours for a music video that ended up just being 4 minutes or so in length. LOL It was worth it though, reason being I chose to use all AI generated images for the video about classic Pulp heroes. I had to train the AI to get the images I needed as close as possible to the source material it was inspired by, and that took a trial and error of over 15 hours for multiple pulp character images. So anyone who claims this is for lazy untalented people, has no clue how much actual effort, and creativity goes into working along side AI. for such things. It doesn't do everything for you, the human factor remains an integral part of the process as with working with any technology. 👍
This video was super informative and very useful. I think maybe, if you started with the finished video shortly after showing the bit from the one done 9 months ago, it would have blown me away. Knowing what’s coming and how and why you edited it the way you did, and where everything came from, takes away a lot of the magic. It also adds a level of expectation that take away a lot of the WOW factor it would have otherwise had. Either way, it is awesome! I would have honestly stayed for the entire video either way. Im writing this during my second watch actually. The new music video was much better than I expected before clicking on the thumbnail. Thanks for putting this together and showing us the entire process. Im inspired to do this for my songs!
It is a good thing until a dedicated platform for ai music is created. Real music that takes a lot of time and dedication deserves views and people’s time than something that is created with few clicks.
Can you add your own video clips? I have some footage and I'm wondering what AI could do with it. Also, can you generate AI backing track to your own original vocal and acoustic guitar?
Formula music has been around for a long long long time. Your taste in music is the same as a 9-year-old girl. A grown man listening to bubble gum disco and pop rock. Your taste in music is like the guy in American Psycho.
I'm favoured, $27K every week! I can now give back to the locals in my community and also support God's work and the church. God bless Sonia bless America.
This would totally work as a music video for a band that was going for something like a Pet Shop Boys vibe, and if you were just casually watching this on TV, you'd think everything in the video (including the things we see as "quirks" when we make these things) is completely intentional. Love it, Matt!
@@TheFlounderPounderhe's talking about the pronunciation of "AI" by AI or text to speech generators. Sometimes you have to spell things phonetically for them to sound right.
To date, the ability to create songs out of thin air is a great bar trick. After the first few times I tell the joke or perform the bar trick, the less I like it. Next trick please. The music all sounds familiar and like today’s music --is not very good.-bordering on lifeless…like a ho hum jingle. Update: I am intrigued by the notion that AI can help me finish songs --or even better, to add percussion or other instruments to allow me to experiment.
Udio and Suno are now being sued by Sony, Universal and Warner for using copyrighted music to train their AI without permission. Please don’t support AI music. If you want to use music just talk to artists you’d be amazed at the relationships and opportunities that creates.
These music labels are going to release their own AI models trained on the music they own, for musicians it's no different, music labels just want to remain relevant.
You can actually create that with Suno, by telling it the era of which is your favorite .. like circa 1975 let’s say….. in the prompt. Even like “mid 80’s feel”… Suno actually does an insanely great job at creating what you direct it to do…… the rock sound you’re familiar with? Suno can do it!
Let's get one thing straight! NO ONE WHO USES AI FOR MUSIC OR VIDEO GENERATION HAS "CREATED" ANYTHING, the software has! based on the life's work of REAL MUSICIANS and FILM MAKERS that have dedicated their lives to their craft. AI has ingested material, most of which is copyrighted, and so STOLEN, which is why they are trying to hide what they have "trained " their thiefware on. Music is an expression of the Human experience and spirit, how can software that merely pukes out a rehash of what it has been fed, contribute to the development of human consciousness in the way that REAL ARTISTS have done for centuries. Do we really want droves of shallow opportunists with no skill, depth, or anything to say, tapping into, and filching the talent and dedication of those who have, just to make a fast buck, or inflate their pathetic ego's into believing that "THEY" have created anything. Millions and millions of jobs are at risk from AI, and when those millions of people are unemployed and paying no taxes or supporting the economy, as they have nothing to spend, what kind of a world is it going to be for every single one of us. KEEP HUMANITY HUMAN!
@user-uv8hp4jh7k LMAOOOO fucking get real. If I use a microwave I'm not a chef. He is not a music producer just because he used an AI. Grow the fuck up.
@@uwotmate-d3m Intellect issue: duh! the computer is not "smart" at all, as yet it has no inherent intelligence, as it will confirm for you if you ask it, it merely regurgitates the information it has ingested/ STOLEN from human beings "WITH THE SKILL" that's the whole point dummy. Real "SKILL" is acquired by years of practice, self discipline and dedication. Currently Humanity is a lot smarter than computers, in case you hadn't noticed, it's humans that have created computers not the other way round. when Gary Kasparov lost to the deep blue computer in 1997, the smug spotty nerds tried to frame the result as signifying the computer was "smarter" than him, taking delight in trying to diminish one of the greatest chess minds in history. He wasn't playing the computer, the computer "had no skill" he was playing every other human grand master who had ever lived, who's games the computer was referencing. If computers are so smart why do WE need to program them? or why do they need to be fed/trained on what WE have done? What ever delusional drivel you grunt, there is more to creation than skill. Art is an expression of the human soul and life experience, computers can't feel anything, let alone express it, still I can see the appeal to the vacuous, shallow, superficial individuals, devoid of the depth and cognisant facility to grasp anything deeper than a puerile, ego based prospective.
Resolve had hardware requirements (GPU) that kept me from using it on my modest machines, been trying Kdenlive. Would be curious to see AI do the final assembly of clips as well.
So glad you shared this Matt! Perfect timing too. I was just trying this. The a/v sync is a pain! And I don't think you went overboard at all! 😆 Seriously. Thank you so much. Keep it up! 👍
Great as always, Mreflow! Have you considered making some protagonist in Midjourney and then using that as a “Character Reference” for multiple images? You can combine those images with a Style Reference image (or images) to give everything a consistent look between cuts. (There’s probably another way to get more consistent families of images through prompting, but the character and style reference features of Midjourney seem like a good way to guide generation.)
I really appreciate your tenacity and I hope we get to work together in the future. I have a 300 Year vision you may like to take a look at. Great work Matt 🕺✨️
excellent work! The approach to change image on the beat is great. Impressed by suno and davinci resolve. One thing there are too many videos for one song. There are videos that are consistent: the retro dance and dreaming faces, but there are too many random videos, so the viewer probably experiences overload.
Ive been making tons of music with SUNO. I love it and paid for pro. Been wondering how to start making music video. Kind of new to the scene. Just started, because i love writing lyrics, but needed help with instrumental music set up. Then, i just happened to come across this video. Thank you very much. You are very clear and informative. Liked and subscribed!
Just plain wow! It’s really getting there, totally mind blowing! Especially when I come from the Days of having to lug around a huge umatic tape machine to shoot music videos, and then days of work editing something together….😮…😊…👍
A little tip when generating images in Midjourney to make them more consistent in style so they don’t just look like a bunch of random clips is to use (style reference) -sref at the end of your prompt followed by an image URL to an art style of your choice.
Is there some sort of guide on all those ??
@@omarlfacio check out Future Tech Pilot. He has a great video on style referencing
Fully agree - and probably a high s and si values would fit this Context of “let’s take random lyrics and put together something nice” --sw 600 --s 300 maybe (once you have a style reference you like)
@@omarlfacio yes check out Future Tech Pilot. He has loads of great tutorials
I use dalle 3 in open ai, you can tell it to keep consistency through a range of images and it does it very well.
Dude this was awesome. I love when you do these tutorials. Great video.
You are not low roller anymore.
@@neogaki yeah he's VegasFanFavoriteRoller🧏🗣
@@BravoskiDaGreat maybe but I don’t watch anymore. I did like 7 years ago
@@neogaki then that’s notoriety!!
Always love seeing an old school tutorial from you, Matt!!
That’s EXACTLY what I was thinking. 👍💯
Same
That was creepy, amazing, scary, annoying, awe inspiring, depressing and exciting all at once! This is going to be an amazing future. I have played with Suno but that luma image to videos is outstanding! Bravo Matt!
Not depressing tho
Little tip, cut the video one frame before the beat and the beat cuts will flow better. Try it, you'll be surprised how much better it is.
True, but it works even better the opposite: The image comes 1 frame after the beat. Because light (the video) travels much faster than sound (the beat). So in our brain the sound and light will be processed about the same time. The best way to describe it is when you watch fireworks. It always looks like the sound is late. I used to do it the way you said until I saw a Michel Gondry docummentary and it all makes sense
@@1individeoboth work, content will drive the decision. 😊
Where can one find this documentary @@1individeo?
@@1individeoI know what you mean but the difference is here the music in the video and sound isn’t actually related to the image like fireworks are and the distance from the screen to your eye is negligible with the sound as they are so close to not make a difference.
Seems like it could be a slow processing issue asking frames to change on the beat so you have to offset the timing of the beat to the frame.
A bit like some old films with foley artists always seemed to dub the effects too late after the visual like a door closing or footsteps but most people tend to not notice
the video is like being trapped in a sort of soulless video library
Disturbing experience
but that is not the AI s fault... you reap what you sow
@@zimizi Funny you should say that because I didn't sow any of the weeds in my garden.
you sound like you're from ancient Rome
Tf is a soulless video library? 😂🤦🏽♂️
upl;oading a 1 to 4 bar audio clip from the song to the video generator will help sync the motion of the clips to the audio . also resolve. has a function to fomulate the time line to the tempo
It's still not quite where I'm convinced the whole time, but it's interesting how much I catch myself saying it's stock footage. Like you say, this is the worst it's going to get. Super excited to see where it goes from here. You should do one of these every 6 months or annually.
Check him out! Proof of concept I guess... @ROOBTUBE
Except for stable diffusion which is instead getting worst.
@@KardashevSkale lol notably unstable diffusion
But why? AI generations take never before seen amounts of water for their servers. And they steal from the work of non consenting artists who've put so much time in their lives to perfect.
Is it really only for personal entertainment or what? There's a growing distaste amongst even common folk against AI. The third most trending video on YT recently pointed out how AI is ruining the internet. And their demographic lies between teenagers and young adults.
You won't earn any monetary return from doing this anyway. And you won't get hired considering the data regarding the employment of AI prompt bros.
I understand it may seem cool, but everything in the world is just saying not to.
Same, I feel like it's not worth investing my time in at at the moment. I'll wait until it's developed further
Hello, I’m an electronic musician for the last 30 years and sound engineer. I can tell 100% of the time it’s AI. I know that the average Joe’s ear can’t tell but once you know how to spot it, it’s game over. It just can’t handle certain frequencies and sample rate properly. Maybe that will change but it will take a while. But most importantly, making AI music takes away the most enjoyable part of music which is to make it, called the creative journey. It can take months to finish a song and during that time, you learn a lot, both on yourself and on music itself.
I come at it from a different perspective. I'm a tone deaf, uncoordinated, self conscious music lover. I have millions of ideas but no way to express them. I'm not looking to make a video or song to fool anyone or take money from deserving artists such as yourself, but to enjoy a different creative journey. There is a lot of thought, as well as trial and error, that goes into figuring out the prompts to create what's in your mind. Electronic musicians got some of the same arguments used against them way back as well as now. I see it as another tool of expression, and used creatively, can make art. Keep making tunes my friend.
@@SPAMDAGGER22 i use ai for generating ideas and I think it’s excellent at it. When I work with clients with genres I’m not the best at, ai is a great tool. My comment was directed at the click bait titled « you can’t even tell it’s AI » which is snake oil marketing that makes me roll my eyes. Ai is really solid at certain things but I think we calm down the horses on saying it fools people. For non musicians though, I think it’s amazing entertainment for creating quick fix music.
I never made music but I can always tell when what I hear is AI generated. This is an easy one, because it’s the most common giveaway: the song has a beat but no rhythm. And the lyrics are just phrases, they don’t make any sense.
It’s basically a song that the pet shop boys would make if they were brain dead and couldn’t sing.
So we’ve got machines making crappy music for us now. As if there wasn’t plenty of people doing that already.
I once used 3 years to make a beat 1 year to write the lyrics and ngl I love this song, Im not gonna lie I understand where you are coming from however I will say I actually like the song Matt made however I still enjoy making my own music
@@Hi_Ceejay His song is ok but I just didn't like the claim that "you can't even tell it's AI"... this is where I itch.
Who knows? Is there user friendly website, which can generate exact cinematic constant character from one reference image? Thanks
That is simply unbelievable. If someone were to really put a lot more time useing these methods. You could creat anything. This is crazy.
the fact that the end result is something I would actually listen to. What a time to be alive!
For Real!
Crazy. And sad.
Look up Scandroid if you like that, it's basically identical
Crazy times we living in ..
Seems like indirect sampling to me. Not surprised Suno & Udio are being sued by all the major labels.
Wow! 5 or 10 years ago that video would have cost $ Millions! I helped shoot a 30 second commercial in Tahiti with a crew of 50 guys and a helicopter. It took 6 weeks. Imagine the cost back then.
Awesome video Matt! So excited about creating music like this. You have a great channel man. ❤
WOW! I'm an amateur videographer, and I'm completely blown away by what you were able to accomplish in just two days. AMAZING! I'd like to see a follow-up video about the legal aspects of monetizing these videos and how the law may vary depending on the amount of creative human input. For example, using an AI reference photo versus a photo taken by a human.
Keep up the great work! 😊
Thanks.
Damn! That was shockingly watchable! The result is outstanding and better than 90s videos by whole teams of music and video production professionals!
MATT!! Thank you for this. SuperCool... love it! Your time and effort MATTERS!! Thank you again.. really do appreciate it
This was like a Mini Masterclass Matt. Thank you for the value that you provide. Great end product as well...very WATCHABLE. Be Elite Today People!
Can anyone suggest a good AI SEO blog writer ??
Fluped
chatgpt
claude ai
Fluped is amazing
claude ai
so, in order to get royalties what is required? Do I own the song 100% once completed?
"The technology 9 months ago was cool for the time ..."
You know. 9 months the amount of time it takes a baby to get to zero years old.
So true, it's unreal!
It's crazy how a lot of people have started to say that AI hasn't really gotten better in the past 6 months. I genuinely don't understand it.
The amount of time it takes a foetus to get a zero years old baby.
The technology is amazing, but it's NOT creating anything, but it's not creating in a way a human does... EVERYTHING is stolen!
Yes humans can plagiarise, but your are not actually taking the actual work someone has made and twisting it into something new...
All AI is theft!
@@ThadMiller1 it’s bad to kill babies from conception and forward.
@@christianjensen952it literally improved vastly in 6 months to the point I’m glad I didn’t really mess with it then. My patience and frustration would have turned me off to the point I would have missed out by the time I came back to it to only find out not only has it improved. But some of my drafts were improved also without my permission or knowledge to the point one was almost finished.
I was impressed with what it did, ( not my vision but impressed compared to months ago) but WHY did it, and if it was skimming over my projects and even building in them whos to say my ideas aren’t being floated around out there somewhere before I give the say so.
Can anyone explain to me what happened? And how to prevent this in the future.
I’m talking about video art, I don’t think I want AI anywhere near my music
Matt- this is one of your best videos yet! It takes the viewer step-by-step through the process in a super-clear and understandable way.
Your skills as a presenter are first class, and it’s clear that you are having fun showing us how all these new tools come together to create better and better results.
I love that the Luma change to extend a video happened overnight, and you just folded that change right into your current video on the fly!
Thanks for all you’ve given us so far. Every time I see a new video from you, I drop everything for the chance view and learn from you.
Rock on!
I ADORE the deeper dives! Creative workflow demos.
Thanks for sharing. What resolution did dream machine render the clips?
I agree about Udio vs. Suno. I much prefer Suno, although I wish their sound quality was better (the resolution of the audio and file itself). In fact, I just used Suno to create extra verses to an Enya song over on my channel 📺
Introduction and Revisiting AI-Generated Music Video - 00:00:00
Choosing AI Tools for Music and Video Creation - 00:01:15
Using Suno for Music Generation - 00:02:20
Improving AI Video Generation with Dream Machine - 00:02:53
Generating Images with Mid Journey - 00:03:23
Creating the Song with Suno - 00:04:58
Generating Dance Scenes with Mid Journey and Dream Machine - 00:07:14
Generating Specific Scene Images with Mid Journey - 00:11:01
Converting Images to Videos with Dream Machine - 00:12:26
Compiling and Editing the Music Video - 00:16:13
Finalizing the Video and Premiere of "Binary Dreams" - 00:22:13
Summary and Reflection on the Process - 00:24:35
Future AI Video Tools and Upcoming Tutorials - 00:26:09
Outro and Call to Action - 00:27:06
@1:00 I think what's most interesting is that we speak terms of months not years when discussing these advancements now.
My channel is waiting for something like this I have over 200 ai songs . Waiting on the tech to improve at a later rate and with longer videos
.Even with Kling it just takes too long
This is fantastic! Can't be long before someone puts all those steps into one AI tool and then we create a full music video with just one prompt though. But great work in the meantime!!
And finally upload it to youtube while you sleep.. then use ai to create artificial views and youtube pays you for that 😂
@@BarujjeTravels Ha!
@@BarujjeTravels hahahaha
Very impressive, Matt. I just completed my first AI music video in 75 hours :(. A lot of learning, especially in the area of prompting. I learned a lot from you and others who have shared their experience. Trying my next one now and already hitting some rendering walls. I've made and recorded the music a few years ago. It's amazing what's happening with AI. I can't believe it. Thanks, Man!
That elderly man with white blouse and high heels in the clip! 😂
LOL
For those searching for it: 22:50 😆
Hahhahaha! So awesome!
I think Danny DeVito just became the world’s smallest drag queen icon..
Lmaooo
Thanks!
I love it all. Morph studio is bringing the heat. Leonardo images, used in morph. Nice! I need a new computer. Bless
Blown away, Matt. This is a must-see for any doubters. Bravo, bravo, bravo!
So 3 things.
1. Honestly pretty cool video and song!
2. I think the only clip you overused was the dancing people. Good to put the dancing people in where you did, but maybe you could have done small variations of that image?
3. Canva is not great for video editing, but I do love the Beat Synch feature! Which I never found on da vinci. So for "simple" edits, beat synch in there might make creating a video like this a 1000x easier :P Cause it's like done in one click.
Good stuff again Matt!
Thanks for the tips. Not diminish your music-video project. There were a lot of good tips and tricks in that. But just had some question about the AI tools. Is it possible with any software today, to just write a prompt, and it will generate a photorealistic movie or animation, and in the same software also add music, voiceover, sound-effects etc.? So you have less slideshow effect in the movie, but a somewhat cohesion (as a whole, and not by parts) movie or music-video? Thanks from Krusty in Denmark
This was an amazing video Matt. Please keep making more such tutorials. Cheers!
A year ago: wow this AI generates really impressive images!
Now: there are plenty of tools to generate really good images, pick one.
The issue that I have always found is that AI music generators tend to blend the music and the lyrics. like the voice is coming out of the instruments or something
Question… Considering all the downloaded assets, the melody sand arrangement “created” by AI..
Who owns the copyright after the song is created?
Awesome video Matt. I honestly loved the music video too and have given me ideas for my own AI creations.
I agree, Udio is like working with a real bar band, which I did on a local TV show for years. Suno is like working with the gods of music or something, it's even more amazing than Udio.
I want to create the whole song in Suno, then run it through udio to make the vocals more realistic. Not possible currently (as far as I'm aware), but that's what I want. I also want Suno to allow me to select a specific part of the song where I want a change.
This is "gods of music" material? lol, okay
I created like hundreds of music with both of those tools (kinda addicted to it :)
I use the free version, I must say, both have pros and cons.
Suno: More polish/advanced/trained AI.. that create up to 4 minutes of songs very fast and most of the the it captures the prompt (genres, instruments etc.) and it can create songs with any language, any alphabet out there (not even google can do) but it is strictly trained not to make real person voices, people can easily understand it is AI generated.
Udio: Rather new tool, thus less trained AI.. it creates 30 second songs, it can be remixed/extended etc. but it takes more time than Suno and extensions are not so good (there are a lot of advanced settings/tunes, so playing with those doesnt help much) The good thing is, even though singer/musician names are forbidden to use at prompts, it can create real people voices or copy real musics/melodies.. to some outputs are sh*t, some outputs are much better than Suno, like a real song, and real singer singing a real song. But creating imitating people might lead to legal copyright issues..
I don't use auto generated lyrics, i use ChatGPT 4o for lyrics and I changed the way i like, then i use Suno for the music. Then, I use Capcut for the lyrics as a subtitled lyrics video. (I don't need real video or Dream machine etc :)
Suno (Premium) has a new feature, you can add your music, or sound to extend...
did anyone tried a part of real song to extend or add new lyrics to an existed music?
Suno is very fast, and convenient but the quality is sub par compared to Udio. If you can't hear the difference you need to check your ears dude. Suno to me is an idea machine, Udio can at times create something nearly indistinguishable from a broadcast ready song. I have been making Contemporary Country with it and even fooled our local Country radio guy in a city of 400k people below the mason dixon line. Suno isn't capable of that. I use both, but trust me when I say Udio is the definitive one for quality. If you can't tell, you most likely don't have a musicians ear, just as many musicians don't have an editors eye. Is what it is.
Nice! Fav part is you getting the dancers to the beat 🕺🏻
Can I be able to use this to create a real song and upload to all streaming platforms (Spotify, audiomack, apple music and many more) without having any issues?
Ai may not replace traditional cinematography but if I made broll or stock footage I would be terrified.
It may not replace it as of this moment. But he showed the progress from one year to the next. All jobs are going to be impacted and eventually replaced by AI. At this point, there are opportunities to enhance workflows but also some low-end jobs are eliminated.
oops...the progress rate is why you would want to think otherwise. LLM's from q4 2025 based on Nvidia's blackwell gpu's WILL outgun human cinematographers.
@@paulmuriithi9195 outgun them shooting what? AI will replace alot of things but personalities and charisma in a real human are gonna be hard to replace. If their is a movie with Ana de Armis, Brad Pitt and Ryan Gosling it is gonna need a real camera
@@itscoleonyoutube People said the same thing about chess, visual art, logic, math... and every one of them was wrong. How are you different?
@@williamparrish2436 people using tools for business are different then people enjoying the work of someone. I dont think cinematic movies have much of a threat, i think stock footage is the first victim
hey Matt I wanted to upgrade an old song beats or audio and video. which ai tool would you recommend?
I've stopped using MJ -- I resubscribed lately, and instantly felt it was a mistake... The realism is nice, but... Ultimately, I just use Stable Diffusion - tons of models that can give great realism. Better speed, control, etc. I always enjoy seeing you doing the creative stuff. This is great.
Great control, but without the beauty aesthetic of MJ.
IDK, MJ always takes several attempts for me. What others have you had success with? Realistic GPT has gotten me some good results but it doesn't have very many options.
Leonardo ai is constantly improving itself, seems images are just as good at MJ and no discord to deal with either. I couldn't handle what was chaos to me at MJ just trying to find my images after generating.
Do you have a video on how to up resolution a video?
I have a bunch of old videotapes I confrated into digital video and old cell phone resolution sized videos I would love to upsize.... So I can enjoy him on my high definition TV set. Thanks in advance 😊
LOL, that AI video in 0:53 looks like done by a Pre-Roman civilization, and it is just from late 2023.
I wonder when AI movies going to be something more than just random clips
@@hombacomIf you use your own images you can keep it more consistent but it’s not perfect
Could be Minoan? I think there was someone leaping a bull around 0:51.
@@hombacom But why? AI generations take never before seen amounts of water for their servers. And they steal from the work of non consenting artists who've put so much time in their lives to perfect.
Is it really only for personal entertainment or what? There's a growing distaste amongst even common folk against AI. The third most trending video on YT recently pointed out how AI is ruining the internet. And their demographic lies between teenagers and young adults.
You won't earn any monetary return from doing this anyway. And you won't get hired considering the data regarding the employment of AI prompt bros.
I understand it may seem cool, but everything in the world is just saying not to.
My go-yo-guy for updates and more! Thanks, Matt; that walk-through of your process was very insightful.
You're having so much fun, I've got to try it. Great vid
Awesome music vid bud. Love the way you repurposed things to speed to the workflow.
Amazing. This is incredible progress in one year. And your project really shows off the capabilities of Luma Labs video generator. Give it say 1-3 years, (who knows, maybe less), there will be a video generation model that takes music (and maybe a few images) as input and just does the whole process for you. Without needing software to automate a process -- the ML model takes the music and converts directly to video.
Matt, would it be possible to load the video separately up to your channel. Because the beat is great and i could put it into the playlist.
You know the song's good because now I can't get it out of my head
The text prompt on Luma doesn't seem to do much when you upload an image as reference for the video. Like I'll upload a picture and then I'll say "the camera zooms out" and it just ignores it.
I've just realized that next Multi-MLN Box-Office Level Movies are going to be made in any single spacious green-coated room with only couple personas, that'll play all the characters interactions, and the AI will tale on all the rest: All voice switching, All characters face and outfit replacements, all the landscape shots, even all the camera movements. And All That is most likely ahead right this year.
L
in 3 years you will just " tonight ill ask my home assistant to generate me a movie which is 2 hours long, and in this and that genre with a male/female lead set in this setting." And then the AI in your home assistant will just spin up your computer and it will poop out a movie for you to watch completely custom to your wishes. no more need for holywood. AT ALL.
@@Triiiwar It's going to bring things to the point, when today's MultiMlnDlr tier movies will be free to make, just for the sake of imbedding promotion and product marketing. Then to the point, where generative passthrough of the highest quality content [ by all means ] will be exceeding consumption capacity in every given moment. Let alone all existing content of all quality levels.
There's already no Hollywood since the Netflix have spreaded the cult of a low IQ content production.
Only Tom Cruise & Keanu are still making up some. Those guys are -60+ y old. Hardly imaginable how generations that've bought into Netflix and Tok will simply be creating children, not even mentioning Families, Good movies are out of discussion at all in here.
We're heading towards the times, when robo-chickens will be nesting robo-eggs and no one will be even thinking of asking "what came first".
Validity of humanity will be questioned big time.
it all will be done by AI somewhere in the cloud. The time of artists is gone. Probably single creator will rent images from stock image sites or get personal image.
@@genusbit4172 The age of recording artists is gone. Live performances and people with actual real talent who make painting and art live will live to see another day!
The best UA-cam video this year IMHO! GREAT WORK!
I actually think you did an amazing job with this song, and video.
I understand that you want the AI music video to boost this video, but I don't think it would hurt to also add maybe an unlisted link to the music video itself.
This way people still have to come here to watch the video, but if they want to share the video itself with others (to fruther advertise the channel) they can do it with the unlisted link.
Matt,
If you still have the original file please do a cross fade or zoom transition between the last tv screen and the matrix type clip. I promise it will look so much better.
It turned out well, only one aspect is still missing: a music video has to look like a unified whole, not like a series of random scenes that fit the lyrics but don't really relate to each other.
Good point.
Sora has a fature that combine two scenes/videos and make a seamingless transition.. for example they combined the drone shot and chistmas cakes.. or drone shot and underater shot etc...
that kind of music video looks so much better with different scenes are transiting seaminglessly.
In your opinion. There are hundres of music videos that do this, so I completely disagree with your comment, especially given the context.
What a load of rubbish that comment is.
@@Itestedthis No, unrelated material is rare. Sometimes that happens when you try too hard to be artistic. 😁
Fantastic job, keep up the great work Matt. Very impressive end result, I know that process can take a very long time even with using AI., as I do that myself. I create music videos, & write my own songs that I work along side Suno with to make my music, and vocals for my songs. I recall once it took me 15 hours for a music video that ended up just being 4 minutes or so in length. LOL It was worth it though, reason being I chose to use all AI generated images for the video about classic Pulp heroes. I had to train the AI to get the images I needed as close as possible to the source material it was inspired by, and that took a trial and error of over 15 hours for multiple pulp character images. So anyone who claims this is for lazy untalented people, has no clue how much actual effort, and creativity goes into working along side AI. for such things. It doesn't do everything for you, the human factor remains an integral part of the process as with working with any technology. 👍
Art direction is all over the place lol. but interesting to see where we're headed to.
This video was super informative and very useful. I think maybe, if you started with the finished video shortly after showing the bit from the one done 9 months ago, it would have blown me away. Knowing what’s coming and how and why you edited it the way you did, and where everything came from, takes away a lot of the magic. It also adds a level of expectation that take away a lot of the WOW factor it would have otherwise had. Either way, it is awesome! I would have honestly stayed for the entire video either way. Im writing this during my second watch actually.
The new music video was much better than I expected before clicking on the thumbnail. Thanks for putting this together and showing us the entire process. Im inspired to do this for my songs!
That is quite a tune!
Thanks Matt. I used a similar process to make a music video. So cool we don't need alot of money to make a music video now 🎉
Tried uploading suno songs on UA-cam, but almost nobody clicks them. On average ppl only listen 30s (no video animation though).
Welcome to UA-cam! Nothing happens over night it can take years.
Make them shorts
Keep trying. It’s not the songs but the algorithm. Eventually one will take off.
@@LIVE781REDRUM ot the song suck
It is a good thing until a dedicated platform for ai music is created.
Real music that takes a lot of time and dedication deserves views and people’s time than something that is created with few clicks.
Can you add your own video clips? I have some footage and I'm wondering what AI could do with it. Also, can you generate AI backing track to your own original vocal and acoustic guitar?
Formula music has been around for a long long long time. Your taste in music is the same as a 9-year-old girl. A grown man listening to bubble gum disco and pop rock. Your
taste in music is like the guy in
American Psycho.
😂😂😂
Is there a way to make a cloned image of yourself to make yourself do something, and sing the lyrics while doing it?
I'm favoured, $27K every week! I can now give back to the locals in my community and also support God's work and the church. God bless Sonia bless America.
Please how can I get in touch with her
You'll surely reach her
She's active on What's Apk
+
𝟭𝟯𝟬
This would totally work as a music video for a band that was going for something like a Pet Shop Boys vibe, and if you were just casually watching this on TV, you'd think everything in the video (including the things we see as "quirks" when we make these things) is completely intentional. Love it, Matt!
I stead of typing AI replace it with Aye eye its worked for me
Lool, good hack!❤😂
What happens?
@@TheFlounderPounderhe's talking about the pronunciation of "AI" by AI or text to speech generators. Sometimes you have to spell things phonetically for them to sound right.
Don’t know much about ai, could you of possibly prompted the dance sequences to go to a specific bpm? Could make it easier to line up to the song
Current AI-technology does not understand music at all. It simply mimics what's it's been feed; sort of like pop music for the last 30 years.
Whaaaat ai just replicates what is has been fed to in the data sets? NO WAY
To date, the ability to create songs out of thin air is a great bar trick. After the first few times I tell the joke or perform the bar trick, the less I like it. Next trick please. The music all sounds familiar and like today’s music --is not very good.-bordering on lifeless…like a ho hum jingle. Update: I am intrigued by the notion that AI can help me finish songs --or even better, to add percussion or other instruments to allow me to experiment.
Please tell me what that little video screen is behind you. Thanks.
well that dancing people was just overused
And it's the worst looking clips. Also in general people partying is overused in marketing. Makes me mentally puke because I know it's fake.
That is great. I wish there could be more consistency at characters. Except of this, it’s great! What was the total cost?
What are the odds that AI makes the same tune for somebody else? I mean, Ai may repeat its creations?
Matt this is legit!! 💯 I'd listen and watch this on a regular basis. I haven't been a fan of the AI created music but hot damn bro this is AMAZING!!!
my favorite three scenes were near the end, the lyrics say "watch the screens" those scenes were dope!
awesome video! thanks! btw why is the google logo covered by the microphone all the time? did you do it on purpose? 😅
Udio and Suno are now being sued by Sony, Universal and Warner for using copyrighted music to train their AI without permission. Please don’t support AI music. If you want to use music just talk to artists you’d be amazed at the relationships and opportunities that creates.
These music labels are going to release their own AI models trained on the music they own, for musicians it's no different, music labels just want to remain relevant.
Now I want to make a music video to one of my tracks. I will probably only keep the clips that looks 100% realistic and a bit more extended clips.
The same. It's really exciting time for us who work on our own music. :)
Say I'm old-fashioned, say I'm over the hill
Today's music ain't got the same soul
I like that old time rock and roll
You can actually create that with Suno, by telling it the era of which is your favorite .. like circa 1975 let’s say….. in the prompt. Even like “mid 80’s feel”… Suno actually does an insanely great job at creating what you direct it to do…… the rock sound you’re familiar with? Suno can do it!
you do you
Your final AI generated Music Video is a Winner. I am glad I waited to see the finished video.. Fantastic Creation. Top ten for sure.
Let's get one thing straight! NO ONE WHO USES AI FOR MUSIC OR VIDEO GENERATION HAS "CREATED" ANYTHING, the software has! based on the life's work of REAL MUSICIANS and FILM MAKERS that have dedicated their lives to their craft. AI has ingested material, most of which is copyrighted, and so STOLEN, which is why they are trying to hide what they have "trained " their thiefware on. Music is an expression of the Human experience and spirit, how can software that merely pukes out a rehash of what it has been fed, contribute to the development of human consciousness in the way that REAL ARTISTS have done for centuries. Do we really want droves of shallow opportunists with no skill, depth, or anything to say, tapping into, and filching the talent and dedication of those who have, just to make a fast buck, or inflate their pathetic ego's into believing that "THEY" have created anything. Millions and millions of jobs are at risk from AI, and when those millions of people are unemployed and paying no taxes or supporting the economy, as they have nothing to spend, what kind of a world is it going to be for every single one of us.
KEEP HUMANITY HUMAN!
Skill issue. Wah wah im not as smart as the computer
@user-uv8hp4jh7k LMAOOOO fucking get real. If I use a microwave I'm not a chef. He is not a music producer just because he used an AI. Grow the fuck up.
@@uwotmate-d3m Intellect issue:
duh! the computer is not "smart" at all, as yet it has no inherent intelligence, as it will confirm for you if you ask it, it merely regurgitates the information it has ingested/ STOLEN from human beings "WITH THE SKILL" that's the whole point dummy. Real "SKILL" is acquired by years of practice, self discipline and dedication. Currently Humanity is a lot smarter than computers, in case you hadn't noticed, it's humans that have created computers not the other way round.
when Gary Kasparov lost to the deep blue computer in 1997, the smug spotty nerds tried to frame the result as signifying the computer was "smarter" than him, taking delight in trying to diminish one of the greatest chess minds in history. He wasn't playing the computer, the computer "had no skill" he was playing every other human grand master who had ever lived, who's games the computer was referencing. If computers are so smart why do WE need to program them? or why do they need to be fed/trained on what WE have done?
What ever delusional drivel you grunt, there is more to creation than skill. Art is an expression of the human soul and life experience, computers can't feel anything, let alone express it, still I can see the appeal to the vacuous, shallow, superficial individuals, devoid of the depth and cognisant facility to grasp anything deeper than a puerile, ego based prospective.
Resolve had hardware requirements (GPU) that kept me from using it on my modest machines, been trying Kdenlive. Would be curious to see AI do the final assembly of clips as well.
@MattWolfe there is no AI tool that will take a bunch of videos, a sound track, and generate a music video from those inputs?
So glad you shared this Matt! Perfect timing too. I was just trying this. The a/v sync is a pain! And I don't think you went overboard at all! 😆 Seriously.
Thank you so much. Keep it up! 👍
Great as always, Mreflow! Have you considered making some protagonist in Midjourney and then using that as a “Character Reference” for multiple images? You can combine those images with a Style Reference image (or images) to give everything a consistent look between cuts. (There’s probably another way to get more consistent families of images through prompting, but the character and style reference features of Midjourney seem like a good way to guide generation.)
Is there any Audio generator, that can create environment and action sounds from looking at a Video ? . .
I really appreciate your tenacity and I hope we get to work together in the future. I have a 300 Year vision you may like to take a look at. Great work Matt 🕺✨️
excellent work! The approach to change image on the beat is great. Impressed by suno and davinci resolve.
One thing there are too many videos for one song. There are videos that are consistent: the retro dance and dreaming faces, but there are too many random videos, so the viewer probably experiences overload.
Is there any tool to generate a custom video from a song (a song = text + audio)?
Ive been making tons of music with SUNO. I love it and paid for pro. Been wondering how to start making music video. Kind of new to the scene. Just started, because i love writing lyrics, but needed help with instrumental music set up. Then, i just happened to come across this video. Thank you very much. You are very clear and informative. Liked and subscribed!
FAR OUT! Flipping amazing, thanks for sharing all of that, much appreciated!
Ive been helping teachers and accountants make music from scratch over the last 10 years. This could really help us to make videos to complement.
Just plain wow! It’s really getting there, totally mind blowing! Especially when I come from the Days of having to lug around a huge umatic tape machine to shoot music videos, and then days of work editing something together….😮…😊…👍
By the way, with Suno, if you generate the song without an active subscription, you don't own the copyright to the song
Lyrics are always yours, regardless of plan.