Thank you. I just happened upon your channel. I am an old lady trying to learn about AI and making content. I had never heard of Luma before but you have made it so easy to understand and use. I am now a subscriber!
Pans up Static camera shot Camera pans left Camera pans right Tilting the camera up Tilting the camera down Zoom in Zoom out Camera pullback Rotate the camera clockwise Rotate the camera counterclockwise Combining multiple motions (Zooming and tilting) Camera tracks a subject walking forwards Panoramic camera shots Turning to face the camera Looking up Looking down
Yes. I went from free to subbed and it's way faster. Also I used about a dozen free images and when I subbed it takes them into account. Mighty frustrating.
Thank you, this is great. I've been having the same issue with panning motion going the wrong direction, glad to see it's not just me. Love your "camera pullback" tip, going to give it a try.
Thank you very much for your great videos. I want to create 9:16 videos suitable for Instagram reels. I write the 9:16 keyword in the text box. But it creates 16:9 videos. What is the best keyword for 9:16 video generating? Thank you 💖💖❤️🌸🙏
Luma only generates landscape videos at the moment. You could crop the videos after downloading for now. They'll probably add in aspect ratio control at some point.
do you think Sora Ai will release sooner now that theirs competition now out? and how long do you think before this technology goes from (1 year ago will smith spaghetti, to this 1 year to etc?)
I seriously doubt it. Right now OpenAi is the big dog and everyone is compared to it so they have currently have time to refine and improve their video generator. Also with Luma out they can see what users like and what they dislike and correct it for their release. So, my estimate is they will stick to their schedule and may even delay it now
It's hard to say, Open Ai has so many different projects from ChatGPT, DaLLE and now Sora. When it does release I'm sure it will be really high quality.
Another great tip is use the most extreme words possible. Like if you want characters running or fighting say "frantic, extreme, chaotic, murderous, exhausting". AI cant produce extreme results but it gets your characters bopping about.
Ok, but you didnt get zoom, you got dolly. Sometimes I get dolly in shot even though I prompt that I want a static camera and there is movement anyway. I was unable to actually get a static camera in some shots. It looks like the AI sometimes doesn't even read my prompts or just doesnt care :)
Hopefully they will add a camera control feature soon. In the meantime, there is a new keyframes feature which makes controlling motion much easier. I'm working on a guide for that
@ I think the video generation and the camera movement inherently has gotten better but still the camera movement goes off in very few or some videos generations
Thank you so much, this explanation on correct prompt was a lifesaver, I started having the prompt and result I want from my image. I love LUMA. AI I can't wait for them to improve it and make it longer I need at least 1 minute prompt on this. remove the "enhance prompt" help a lot, it would send my image into crazy mode and add things I did not request on the image like extra character that did not fit.So Far I am in love with Luma AI.
I've been using luma. It's very hit and miss with what I had imagined. Sometimes my prompts work perfectly, other times I will get desired affect with no prompts at all. In specific I'm wondering can you share some pointers for morphing one character into another?
I found it works best if the characters look fairly similar, otherwise it's a shot in the dark. Like if you use Midjourney to create 2 photos of the same character. I made a video guide for that here: th-cam.com/video/OmmQcu17OLs/w-d-xo.html
You know what I wish for? I want to be able to grab a snippet of a video doing the movement I want ( I was trying to get a man to go down on one knee and shake hand/paw with a dog) then show the snippet to the ai 5:30 video generator and prompt “Do this!” And it will make my characters do the gesture. You can prompt all day long as of now and never get your characters to do something this simple (like stand up from sitting at a desk) so the audience has to mentally fill in the gaps. “Oh he’s walking now so he must have stood up” That’s just wrong and I can’t wait for the day I can laugh at such crude things we used to do.😅
They just released a keyframes feature that gives a lot more control over character animations It can't really complex motions, but it actually works pretty well for simpler things. I'm working on a video for that now.
Thank you! More tips please :) What keywords do I need to use to get it to make an infinity loop . Like for animation. So I can just the creations for backgrounds
I haven't seen it be able to create looping videos on it's own. You could try combining it with the flowframes software to get looping videos. I have a tutorial on that here: th-cam.com/video/pwWriaYCdg4/w-d-xo.html
Its definitely lacking control but i feel they will get that under control soon. Id like to keep a static camera but have landscape movement. Do you know how to prompt somethin like that?
@@taoprompts I do apologize. My mistake. I use to having a very long prompts and did not realize that it was a short one. Thank you again. It was very helpful. By the way the rummers are that Runway Gen 3 is out tomorrow.
How many money do u use for trying the prompts and give us this great information❤❤🎉 im very happy and thank u so much. Now i can spend my time and money...
Hlo ur video is so helpful,i havea doubt i used a childhood of me standing on step,i want to make a video stepping down but when im making the video the result is the face is changing is not really static the face is not the same as in the image stepping down wht to do for that pls reply
If the face takes up a small region of the image, it is more likely to distort. You could try rerunning it a few times or go with a smaller movement than stepping down
Great work as always. Do you think Luma is stable enough to be used commercially? I used all the free credits and only about 1/4 of the generation looks good, the rest contains artifacts and e.g. morphing. This puts me off the paid plan a bit.
That depends on what you mean by commercially. It's still going to be hard for Ai to generate videos that look 100% natural, especially if you have complex motions & movements in them. Hand & arm movements are especially difficult. For image-to-video Luma is one of the best I've seen especially for human subjects though.
@@taoprompts I tried and I love it. Last two movies on my channel was generated in Luma and I think it's a huge step in AI videos. Can't wait for other tools.
Is it possible for me to get 1 image and morph it into another without changing the original images? Every time I try it with 2 images it completely changes them.
Are you trying to use the keyframes feature? If your two images have similar structure in then, it should do a decent job of preserving the general shapes in the images.
@taoprompts what do you think? It's easy to work with Pika but when creating anime (image to video) image almost doest move. Hair movement, breathing that's it. Only plus is lip sync. From paid ones, which program is the best (image to video)? Considering latest updates. Runway is good for realistic human movement I know. But I'm interested in anime or 3D cartoon.
For anime I haven't seen anything that is really high quality yet. I think Luma is better than Pika for sure, it has more movement. For 3D cartoon style, Luma looks really good imo.
I agree with with you. This AI is the worst! Not just camera movements. Extreme morphing, distortion, extra extremities on human and creatures, weird motions and other problems in the output video. Better make a video on what makes this AI the worst. I wasted 2/3 of paid credits on useless video. More power.
Thank you. I just happened upon your channel. I am an old lady trying to learn about AI and making content. I had never heard of Luma before but you have made it so easy to understand and use. I am now a subscriber!
It's great that you're learning about Ai now, this is the best time to get started 👍
Wow! You just saved us a ton of frustration and you got to this first because you don’t waste time testing new tools. Thanks!
Thanks Suzanne. I was having a lot of trouble with this myself so I did a lot of testing to see what could be done to fix it.
Pans up
Static camera shot
Camera pans left
Camera pans right
Tilting the camera up
Tilting the camera down
Zoom in
Zoom out
Camera pullback
Rotate the camera clockwise
Rotate the camera counterclockwise
Combining multiple motions (Zooming and tilting)
Camera tracks a subject walking forwards
Panoramic camera shots
Turning to face the camera
Looking up
Looking down
Thank you for the tips.
Did you test these more than once?
This is by far the best channel for AI image and video tips. I've binged watched every one of Tao's videos and look forward to more. Great work!
Thanks! I saw your alien video, great work 👍
@@taoprompts Wow, thank you so much for watching it!
would you recommend getting the standard subscription for dream machine? the free version takes forever to load (3-4 hours in queue times)
I have the standard plan. It's definitely way faster than 3-4 hours in queue. I can do a bunch of videos reasonably fast.
Yes. I went from free to subbed and it's way faster. Also I used about a dozen free images and when I subbed it takes them into account. Mighty frustrating.
I can't even imagine what this stuff will look like in the next 2 years. Crazy.
It'll change film and content creation for sure.
Damn ur Luma and Midjourney "camera" help is the best online.
Thanks man 👍. I know a lot of people had issues with this.
Good video. Straight to the point, no hype and helpful.
Thanks Tao ! This was a great Tutorial. I'm glad you also addressed the "Enhance" feature.
Thank you. I was shocked at how much of a difference removing the "enhance prompt" option made, it improved the user experience a lot.
Thank you, this is great. I've been having the same issue with panning motion going the wrong direction, glad to see it's not just me. Love your "camera pullback" tip, going to give it a try.
Appreciate that man. I'm excited to see any videos you make about Luma 👍
@@taoprompts What about "rotate right" instead of clockwise?
@@Beunibster That may work, although I don't think the Ai really understands left from right.
Very nice job. You just saved me $30 worth of the generation and at least 2 hours of my time. Thank you for the tips.
Useful video as always, you are one of the few with a useful guide.
Thank you! Luma has fairly basic functions and already produces high quality videos. I'm eager to for them to release their full set of features.
Amazing video and very useful,thank you very much, greetings from Santiago, Chile 🇨🇱
Thanks man 🙏. Great to hear from Chile!
Thanks. Very helpful.
Thanks Matt 👍
Heureca! I am so grateful, this is an amazing tutorial and clarification !!
For sure man, camera control makes a big difference 👍
I'm saving all these videos. Thank you Tao!!
You saved my life with this video, liked and subscribed!
Great to hear this helped you out! I know camera motions was a problem for a lot of people. And thanks for subbing.
Thank you for sharing the details! it helps a lot :)
There's still a lot more experimenting to be done with Luma! Glad you liked this vid.
Amazing content! Thanks friend!
Excellent video Tao, very helpful as always!
Thanks Val, Luma still has a lot of cool features planned. This is just the beginning.
Thank you very much for your great videos. I want to create 9:16 videos suitable for Instagram reels. I write the 9:16 keyword in the text box. But it creates 16:9 videos. What is the best keyword for 9:16 video generating? Thank you 💖💖❤️🌸🙏
Luma only generates landscape videos at the moment. You could crop the videos after downloading for now. They'll probably add in aspect ratio control at some point.
You deserve more views!
Thank you!
Always good and downloaded one of you PDF thanks
Thanks Stephen, I hope that PDF helps with your prompts 👍
Really helpful thanks, subbed I'm on the learning curve
Thanks for the sub, there's still a lot of testing to be done with Luma 👍
do you think Sora Ai will release sooner now that theirs competition now out? and how long do you think before this technology goes from (1 year ago will smith spaghetti, to this 1 year to etc?)
I seriously doubt it. Right now OpenAi is the big dog and everyone is compared to it so they have currently have time to refine and improve their video generator.
Also with Luma out they can see what users like and what they dislike and correct it for their release. So, my estimate is they will stick to their schedule and may even delay it now
It's hard to say, Open Ai has so many different projects from ChatGPT, DaLLE and now Sora. When it does release I'm sure it will be really high quality.
How do you stop all camera motion? Thank you for the great video.
Try turning off the "enhance prompt" checkbox
Another great tip is use the most extreme words possible.
Like if you want characters running or fighting say "frantic, extreme, chaotic, murderous, exhausting". AI cant produce extreme results but it gets your characters bopping about.
Thx for the tips, I'll have to try out some more human animations.
Ok, but you didnt get zoom, you got dolly. Sometimes I get dolly in shot even though I prompt that I want a static camera and there is movement anyway. I was unable to actually get a static camera in some shots. It looks like the AI sometimes doesn't even read my prompts or just doesnt care :)
Hopefully they will add a camera control feature soon. In the meantime, there is a new keyframes feature which makes controlling motion much easier. I'm working on a guide for that
doesnt work for me no matter how i prompt the movement is unpredictable
Amazing work... as usual
Thanks man! Greatly appreciated.
Hello all, how can i change resolution to 9:16 Text to Image on LUMA AI?
I don't think that's possible atm
Thanks for the tips! What is the maximum resolution Luma Ai can create the video?
For wide screen, I get 1360x752
@@taoprompts thank you
Hey there, thank you for making the video but I do not see the enhanced option
they've changed the luma interface recently, I haven't tried the new version yet
@ I think the video generation and the camera movement inherently has gotten better but still the camera movement goes off in very few or some videos generations
Thank you so much, this explanation on correct prompt was a lifesaver, I started having the prompt and result I want from my image. I love LUMA. AI I can't wait for them to improve it and make it longer I need at least 1 minute prompt on this. remove the "enhance prompt" help a lot, it would send my image into crazy mode and add things I did not request on the image like extra character that did not fit.So Far I am in love with Luma AI.
It's a great tool, I think they have a lot of upcoming features planned like video inpainting 👍
I've been using luma. It's very hit and miss with what I had imagined. Sometimes my prompts work perfectly, other times I will get desired affect with no prompts at all. In specific I'm wondering can you share some pointers for morphing one character into another?
I found it works best if the characters look fairly similar, otherwise it's a shot in the dark. Like if you use Midjourney to create 2 photos of the same character. I made a video guide for that here: th-cam.com/video/OmmQcu17OLs/w-d-xo.html
does any body know how to keep the video from not moving side to side. what words do you put in the prompt .
Thank you so much! Very helpful!
Glad this tutorial was helpful!
Te ganaste mi suscripción y mi respeto, Thanks for your videos
Se lo agradezco mucho. Me alegra saber que este tutorial fue útil.
Awesome video. Is there a tool where I can choose which part of the image to animate? Most of these tools animate whatever it pleases.
Runway Gen2 had a motion brush tool which worked like that. They haven't added it to Gen3 yet though.
@@taoprompts thanks I’ll take a look!
This was very helpful, thank you!
You know what I wish for? I want to be able to grab a snippet of a video doing the movement I want ( I was trying to get a man to go down on one knee and shake hand/paw with a dog) then show the snippet to the ai 5:30 video generator and prompt “Do this!” And it will make my characters do the gesture. You can prompt all day long as of now and never get your characters to do something this simple (like stand up from sitting at a desk) so the audience has to mentally fill in the gaps. “Oh he’s walking now so he must have stood up” That’s just wrong and I can’t wait for the day I can laugh at such crude things we used to do.😅
They just released a keyframes feature that gives a lot more control over character animations It can't really complex motions, but it actually works pretty well for simpler things. I'm working on a video for that now.
Thank you! More tips please :) What keywords do I need to use to get it to make an infinity loop . Like for animation. So I can just the creations for backgrounds
I haven't seen it be able to create looping videos on it's own. You could try combining it with the flowframes software to get looping videos.
I have a tutorial on that here: th-cam.com/video/pwWriaYCdg4/w-d-xo.html
@@taoprompts Thank you. Also I subscribed to your channel. I'm sure dream machine will improve in time. Since it's still new and all
@@Esbikenh Thanks for the sub! They just released a really cool new keyframes feature. I'm working on a video for it now.
Thx a lot, great stuff!
HI. Camera movement is important. In some tools like runway gen 3 it is there.
I'm definitely waiting for Gen 3. I hope it lives up to the advertising.
Its definitely lacking control but i feel they will get that under control soon. Id like to keep a static camera but have landscape movement. Do you know how to prompt somethin like that?
If you are using image to video, I think just turning off enhance prompt and asking for "static camera {landscape description}" should work
@@taoprompts thank you
Is there any way you could possibly make a video where you show the whole prompt and then results in Luma AI?
I did show the entire prompts here. I used image-to-video to attach reference images and animated them with very simple prompts.
@@taoprompts I do apologize. My mistake. I use to having a very long prompts and did not realize that it was a short one. Thank you again. It was very helpful. By the way the rummers are that Runway Gen 3 is out tomorrow.
How many money do u use for trying the prompts and give us this great information❤❤🎉 im very happy and thank u so much. Now i can spend my time and money...
Hey, I have the standard plan for $30/month which was enough for the last 2 videos I made. But I'm out of credits now 😂
Thank you!
Hlo ur video is so helpful,i havea doubt i used a childhood of me standing on step,i want to make a video stepping down but when im making the video the result is the face is changing is not really static the face is not the same as in the image stepping down wht to do for that pls reply
If the face takes up a small region of the image, it is more likely to distort. You could try rerunning it a few times or go with a smaller movement than stepping down
I'm more interested with the infamous fever dream generation
The meme videos use keyframes for that. Just upload a start and end frame and it will animate what's in between
How can we get a first person walking movement? i really cant seem to get it...
Maybe try the prompt for the specific type of situation like "walking tour", "FPV camera", "first person POV" or something like that.
GoPro
please can you list all camera motions in description, I think it will help your SEO as well
I added your list to the description 👍
Great video
In an image, if there are multiple people, what kind of command would make the people in the image perform different poses or actions?
I haven't tested much with multiple people yet. In general this will be hard for the AI to animate correctly though.
I usually use square pictures 1024x1024..but the luma sometimes makes kinda 3:4 shots out of it..any ideas?
I don't think it's supposed to do that, it could just be a temporary bug on their side.
How can I download the PDF?
I don't have a pdf for Luma Ai yet, I left the keywords in the video description.
Thank you genius🎉
excelent. thanks
🙏💯I appreciate that man
Great work as always. Do you think Luma is stable enough to be used commercially? I used all the free credits and only about 1/4 of the generation looks good, the rest contains artifacts and e.g. morphing. This puts me off the paid plan a bit.
That depends on what you mean by commercially. It's still going to be hard for Ai to generate videos that look 100% natural, especially if you have complex motions & movements in them. Hand & arm movements are especially difficult. For image-to-video Luma is one of the best I've seen especially for human subjects though.
@@taoprompts I tried and I love it. Last two movies on my channel was generated in Luma and I think it's a huge step in AI videos. Can't wait for other tools.
Thanks!
I am really struggling with good motions for my channel videos. I’m hoping this is going to help
This should give you a bit more control over the camera and Luma typically does a good job of animating the subjects inside the videos.
hello! my luma generator works very slowly. do you know better option to get render easily? it took more than 40 min
It is quite slow, I have purchased the standard monthly plan which is a bit faster than free users.
I knew the enhanced prompt box was bad news haha
Yeah, It seems to add a ton of unnecessary motion randomly
Is it possible for me to get 1 image and morph it into another without changing the original images? Every time I try it with 2 images it completely changes them.
Are you trying to use the keyframes feature? If your two images have similar structure in then, it should do a decent job of preserving the general shapes in the images.
tried on 2 browsers to get prompt guide , gumroad not processing
That could be a temporary gumroad issue. Google chrome seems to work fine for me.
it Maximum concurrent usage limit exceeded. how to fix it?
If you are on a free plan, there will be some restrictions on how many videos you can create.
Is there a limit to how many imagines you can create if you pay for the service?
thanks
So sora is better for controlling the camera motions.
I haven't tried Sora, maybe it will be
still no static camera that actually works on Luma or any other AI video gen I've tried!.:(
how many features in Luma Ai . ...............Video generator or 3d camera motion
It is still pretty basic right now. I think they plan to add in video inpainting soon.
@@taoprompts great
who we can slow motion it? i put slow motion in the begning and the kling just ignore it :(
You can always slow down the speed in a video editor.
Hi. Can you make a long take with two actors talking in the same room and then dub with lipsync?
Yes, that can be done in Runway
useful
Static does nothing for me EVER constantly panning and I'm really starting to get annoying. Even tried prompts I got from support
subtle motion, static camera, cinematic still, motionless
What if you don't want any camera motion at all?
You can try prompting for static camera and turn off the enhance prompt option
I have the paid version but doesn't remove watermark... kinda dissapointed by that
The download without watermark feature seems to be working for me atm.
@@taoprompts Yes I feel Like an idiot now, I discovered it later after my comment
👏👍
❤
🙏
Do you think Luma is better than Pika?
Yeah it’s better 😂 less morphing 😂
It's like comparing blue ray to vhs.
4k to 480p. Eating a strawberry to a coffee bean. 😂
@taoprompts what do you think?
It's easy to work with Pika but when creating anime (image to video) image almost doest move. Hair movement, breathing that's it. Only plus is lip sync.
From paid ones, which program is the best (image to video)? Considering latest updates.
Runway is good for realistic human movement I know. But I'm interested in anime or 3D cartoon.
@@piggywahwah I have paid to pika but I was completely satisfied. So definitely want to try other one.
For anime I haven't seen anything that is really high quality yet. I think Luma is better than Pika for sure, it has more movement. For 3D cartoon style, Luma looks really good imo.
SOUND EFFECT CREATE ,AI HELP NEXT VIDEO POST ??
I usually use elevenlabs for Ai voices. And then Runway for lip sync. Perhaps Luma will add in a lip sync feature also.
If only there was a way to keep the camera stationary consistently :*(
Try prompting for "static camera, cinematic still, subtle motion", words like that have helped me
That is not a ZOOM in, the camera is moving.
Is your video Avatar AI made ? It looks like a German video Avatar ! Very Human !
That's really me lol
I agree with with you. This AI is the worst! Not just camera movements. Extreme morphing, distortion, extra extremities on human and creatures, weird motions and other problems in the output video. Better make a video on what makes this AI the worst. I wasted 2/3 of paid credits on useless video. More power.
Kling or Minimax are better options right now
Thanks for the great info! I'm having a lot of fun animating my stories with Luma and Runway ML th-cam.com/users/shortsQQtaNRMn-FA?feature=share
Great work 👍
kinda too limiated, yet
omg this luma ai is horrible
does any body know how to keep the video from not moving side to side. what words do you put in the prompt .
Thank you, very helpful