Check out Kling 1.5 in this video ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-G2Pf7_SJQWg.html Join the AI Filmmaker Academy Today! And learn to make AI Films that make money 40% off with coupon RU-vid40 -limited time only. www.aifilmmaker.academy/ FYI: My memory card adapter ejected itself at the end of the export, causing minor visual glitches in this video. I assure you, I am real and not part of a simulation. I’ve got a new adapter now, so this won’t happen again. Thanks for watching!
I'd happily pay for a service, but there are dozens of AI tools offering almost exactly the same thing so it's impossible to know which to go for. I used Haiper while it was free, but haven't used it since.
Samson, I'm new to your channel and loving it! I'm a filmmaker, and I need to start generating some looping RU-vid and TikTok shorts. I'm learning so much here. Thank you!
You didn't happen to have monitors on while recording? There's a metallic resonance on every word that doesn't sound like room reverb, could be early stage feedback.
Thanks for the detailed overview of the neural network. Close-ups are especially great. This is important when creating (editing) a film!! As always, I like for the review!!!!!
Kling is amazing. I have had an account for a while and look forward to not having to send images and videos back and forth to my phone through te app. You are right, it is not perfect but it is really, really good. Amazing ability to consistently show natural human movement. It will get better. I hope that have a good paid solution with the new international site. On the Chinese app version there is NO way the app would have let you generate a video with all that cleavage. The international website seems to be more reasonable in that regard. LOVE your channel and the way you explain things.
In the Times of Singularity 1 year of progress in past years are like 3 month today and 1 month in 2025. The democratization of current virtual AI will only be surpassed by the democratization of Physical AI.
It’s not just that there is more movement, it’s the details in the movement. The others look stiff and anemic by comparison. Kling actually moves skin as it stretches and tightens and also allows the hair to drape over her skin as she turns her head. That’s pretty friggin realistic.
If you turn your studio monitors down or off, we won't hear the low frequency feedback. 06:27 - Stuart Little is cruising the streets in his little red sports car.
Brilliant. I have a whole catalog of music from the last 20 years finally being recorded...pretty much aural journals...this sounds like an incredible opportunity to add visual elements to the story telling...Thanks so much.
All the people behind these companies in China have their education finished in the US. Hua Su, one of the co-founders of Kuaishou, has a strong educational background in computer science. He earned his bachelor’s degree from Tsinghua University and later pursued a Ph.D. at the University of California, Los Angeles (UCLA)12
I don't really use Luma since I personally feel like their physics and movement get too wild and weird. I do use Gen3. Kling doesn't seem to be very good with text to video, which Gen3 definitely is. On the other hand, King can be shockingly good with movement when doing image to video, which Gen3 doesn't have just yet.
Yes - it seems each time they upgrade the use case changes for each AI Creation tool. We really need a current snap shot and break down of the use case for each. But things can change tomorrow. lol So that's probably why it doesn't exist. So it seems this will be a movable feast !
Kling seems to excel particularly in generating video from realistic images, you start using anything that is at all illustrated or stylized and it ends up looking like chunks of the image are being moved around rather than an actual video and yeah, the text to video functionality is way worse than Gen-3. There's a tool for each use case but I'm hoping Gen-3's image to video features are good enough when they release that I can just stick with that for now.
I tried Kling but there is no way to keep the camera in a fixed position! But it seems promising, thx for your videos, they always open new paths in my RU-vid journey as a 64 years old creator!
When one considers how much dedicated software was required to render 3d environments including SFX mimicking natural phenomena and forces, now all that is rendered obsolete by simply scripting everything which needs to appear in a scene. Look at the typical credit roll of an action Feature Film to see how many FX houses were involved.
Something odd with the upload, quite a few glitches. First I thought it was my connection but then I went back and it happened your end. Don’t know if you’re aware but you might want to sort that out before your next video!
They yesterday finally approved me on the Chinese site. On that site they have subscription running at 50% discount. The pricing is similar to other video generators but as the quality is much higher and need less rererols they are clear winners. Great that they launched global site as writing prompts and translate them in Chinese is not the most handy workflow 😅😂
✨Thanks for sharing this info ✨ I’m enjoying luma so far, but I can’t wait to see what this can provide. All of these platforms are still far from perfect and in their infancy as we all know, but they are still fun to play with. 😊📺
I used a Pan camera motion with text to video and it panned the camera across the action such that I couldn't see the beginning or end of the action. The action took place in one place and the camera simply panned into the action and continued to pan out. Silly. Another scene I had described one person being chased by another, but it kept having them run towards each other. I couldn't get them to run in the same direction, haha! I tried a couple of prompts using phrases like "running away", "chased by", "chasing", "running along"... gotta keep rolling.
Thanks for the feedback! Have you tried the similar results in Runway? Curious between the two, and if I should start using Kling versus my expensive runway lol
I guess , Luma AI is the best on the market right now especially the image to video version. It’s a great deal. I just build in Klingai and Luma AI using the same picture and i can see the cinematic feeling more in Luma . I use the same reference and same prompt.
Your mic audio has a bad resonance to it. It sounds like a table top, or a metal panel of a desk that is un-damped, and resonating when you talk a little loud. The frequency of the resonance is around 260Hz. You should thump on flat surfaces around your desk to see what "rings" at a low frequency like 260Hz... then tape some mass to it - to damp the oscillations. (this assumes that you are aware of audio rules that you don't have a speaker playing your own speech audio in the room... it could just be feedback) Besides that, the snowfall wasn't convincing because the flakes kept disappearing mid-fall, and new ones appeared out of nowhere... instead of appearing from only above the camera view. You also need to fix your own selfie camera.
when you did the comparaison at the end, did you just feed it the image without prompt ? I Find sometimes it really doesn't understand or follow the prompt and makes things more realistic without prompt usually. Thanks for the video ;)
I tried it and was not impressed - at least not yet. I also couldn't find any terms of service or indication whether the video output could be used for commercial purposes. Guess it's early days.
Hi Samson, I created a song with udio, before the update, that mixes lyrics and massive bass elements. It's very personal as it talks about my relationship with my father, my inner child who wants my father to be proud of me, and the difficulty this inner child has with my adult self. It's very unique. I'll try to make a music video using klink - it will take time, but I'll send it to you because I need to share it. I'm not interested in you talking about it in a video, I just want to show you my vision and how I can use these tools to tell a story. Thank you and congratulations on your excellent outreach work - without you, I wouldn't have known about klink's release. Thank you!
How would one use this kind of tool for congruent story telling? I'm not experienced, just dabbling with an idea. But, let's say I have built 4 characters, scenes where these characters interact. The problem I see would be to maintain consistency in the way characters look and feel or am I wrong? Based on your awesome video, I'd say it is great at generating separate clips but I fail to see how I would tell film or cartoon-like stories with it. If anyone has got an idea, please kindly advice. I'm looking to develop a process for a new 'hobbie' to create short animated films with AI about the same characters, you know...
What's the usecase of these though? Short clips only? What's the longest clip it can generate? Thanks in advance for your answer, Samson. New subscriber 🎉
I can't sign up. I can't get more than the opening page to translate to English and the sign up form doesn't work. I've tried Firefox, Chrome, and Edge.
Thanks Samson! I've run a few Kling videos and I need to use better images. Garbage in garbage out. 🤣 Also, your webcam is very glitchy. It blacks out and gets pixelated like a Cyberpunk glitch, lol.
Except that novels aren’t screenplays. Many many novels are narrative of characters’ inner thoughts. Which are impossible to translate to a full motion film, without voiceover. Which at that point, may as well listen to an audio book. This problem is why it is very difficult to adapt some books to screenplays. You have to get at the book’s themes in new ways more native to cinema. Visually poetic imagery, clever use of dialogue or actors performances to express things that were expressed in the original books narrative voice. Surely at some point you’ll be able to generate something in this way, but it’s much too complex to just feed it a book, and have it generate a decent film as it reads the text. It would need to do a deep analysis of the entire text, and then make editorial decisions in the same way a filmmaker would. That’s far beyond the ability of any current tech.
An application better-suited to this real-time generation technology would be a kind of on-the-fly video game, or make-believe ‘life simulation,’ where you could go anywhere or do anything, have it invent characters to meet and talk with etc. For example, you’d tell the AI, "Simulation based upon 20,000 Leagues Under the Sea. I’m a passenger on the Nautilus. Events can include giant squid, visiting an uncharted paradise island, schools of whales, secret grotto base. Characters should be realistic but funny, comedic dialogue. Entertaining tone." Then while you’re playing you could pause and request changes to the prompt, for example, "this is great but let’s discover Atlantis, and meet a race of mer-people."
Good video from your end, but the platform I tried it and it does not give me the same results as you get, if you can please next time you make this type of videos, do not use their premade prompts because the ai already learned those, also make longer and more complex prompts and your own custom prompts so we can experiment the real results because once we try the platform most of them are not the same results
Hi, I want to address something you say in your video about KlingAI's image generator it is actually REALLY good when you add a reference image. You can use the reference to get the style you want, it can do photorealistic very well this way.
I started generating a 5 second video yesterday and it still says it is creating it. I am starting to wonder if it is a real site or just trying to get people to sign up to get emails.
Already for days, their sign-up security check (putting the puzzle piece at the right place) doesn’t work. So, completely impossible to sign-up. No support link either 😤
Question...I have an idea that requires that I turn a regular photo or even a movie, that's shot in real time and turn that into an anime cartoon character without the big eyes or anything like that....just take a picture and turn it into a cartoon like character...do you have any ideas or suggestions?