I wouldn't say open-source AI tools aren't as good. You have to work harder to get your results but you can also do more with them. As a VFX artist, I've been most impressed with what ComfyUI can do.
This is so true. I'm not a trained VFX artist but I've found WAY better quality with motion capture using the open source tools (Live Portrait, MimicMotion) vs the commercial tools (Mootion, Viggle)
Open Source AI tools would be cheaper compared to paid commercial tools. From experience, for every 5 AI videos, you get 1 good video. If you have 25 clips for a video, you would need to make 100 to 150 clips to get decent videos for the project. So costs will mount up quickly! If you get a decent machine to render of the videos, it will pay for itself in the long term. Also, the ComfyUI community will develop new tools to gain more control.
Wheres the best place to learn this. I'm kinda put off by all this cmd prompt/python/bat file. Most of the time the walkthrough assume u know what to do in cmd prompt. They talk with jargon I'm unfamiliar with.
The new Chinese yext to video platform "Mini Max" is amazing made some of the best text to video generations I've seen yet, very very realistic and good movement with very little to no warping at all, and it's completely free and unlimited "as of now" only issue is no image to video yet, but once they get thay going, I will definitely rival sora
Runway Gen 2-3 are still my go to generators. Although I think Gen 2 is better. G3 can go wonky sometimes. Luma is getting better too. I like it for the aspect ratio the most. And Haiper is a true underdog. For me, I have been more insisting on "big movements" in my Ai vids and Haiper has always brought "Action Packed" to my MJ images. Which is now a series. I haven't tried Kling yet. I love this channel!!
Kling beats them all, especially its End frame add option, looks like not a lot of people actually know about it but it's a beast :), Haiper is the next best, especially its 3 frame keyframing, it creates some amazing results, as long as you have three somewhat consistent shots.
@@Jarbrain Theres 2 formats : 1. You can set the end frame so the animation ends on that frame. 2. You can have a start frame and an end frame. So bear in mind theres 3 options. A. Traditional : you start with image and prompt B. Set image to end frame: the prompt will end on that last scene. C. Start and end frame. The video will go from your start image to your end image Hope that makes sense. Good example is a kiss. If you have a couple kissing then you stuck with kiss then pull away. But if you set the kiss to end frame (in luma and kling). You can have the couple coming together and finish on the kiss. In luma theres a tiny arrow above the upload images, you have to click on that to activate "finish on this image"
@@Jarbrain just like luma or haiper, you add two very similar shots and it does the interpolation and makes the end movement very smooth and realistic looking.
Kling is definitely the best yeah. I've tried both of them on their premium plans, and Kling destroys Runway in a lot of ways. Gen3 is such unpredictable and sometimes ends up doing just a push in. None of that with Kling.
@@Trankilstef exactly! And there end-frame tool is extremely powerful. I usually just generate a lot of videos with the original frame, and when I find a frame that I like within those videos, I use it as the end frame. Once Kling adds camera control and more control over the animation, it’ll be enough to create films
I would put Kling, Runway Gen-3 and Luma Labs at the same level, at least for image-to-video. I've created hundreds of video clips on each one and they each have their strengths and weaknesses for different types of shots
My experience with Kling has been really good. And I'd like to point out that Standard Mode is nothing like Gen2. It can give you weird results, but often it's very much on par with Luma or Gen3. Actually, I rarely use the Professional Mode, as often Standard result are pretty good, but also because Pro can be sooo painfully slow. Also, in Kling you can add an end frame as well, like in Luma. The only thing is that you can't give it JUST the end frame, it always needs the first frame as well. Luma is also pretty good, I noticed it tends to be a bit stubborn and static at times, while Kling may go a bit crazy here and there, but this way you have more options to experiment. I'm very impressed with Gen3, but I need to play with it a bit more. Also, it should be mentioned there also two options: the normal Gen3 Alpha and Gen3 Alpha Turbo, and it's a bit like with Kling, where you have a faster/cheaper version and a more "Pro" one, but the cheaper one is often as good as the Pro. And one thing I can say about Gen3 Turbo is it's hands down the fastwst video generator I know. Oh, and to remedy Pixverse a little, they also have a built in 4k upscale option. :) I should also add that ComfyUI is the most pro tool today when it comes to image generation, it's light years ahead of any image generator out there (cause it can do everything those generators can and much much more, with waaaaaaay more control). Haven't used ot for video though. Yet. ;)
I have recently been trying Minimax which is really impressive and it's free for now. Unlimited video generation and relatively fast. Runway is no doubt fast and of good quality but i feel minimax might be a game changer
Kling is probably the best, but their terms state that in order to use their videos in advertisements or for commercial purposes, it must be noted that it was created by Kling or a watermark must be included, which could be problematic when working with a larger company. What do you think?
@@curiousrefuge hey, of course I did see the whole video like others in the past. I couldn’t find you exploring the video to video generators how they hv evolved so far. Hence asked.
The conversion from 120, 60 or 30 to 24 fps on Davinci Resolve it's perfect and cost nothing, so that criteria of the 24 fps says nothing about the AI Video Generator it's useless
Which would be the best AI Video Generator for creating drawing type animations such as anime or comic book characters? All of the examples I've seen on youtube are usually 3D Pixar type or photo realistic type.
Great question! 24fps is the frame rate used to shoot the majority of movies. This is typically industry standard and will help in creating more cinematic and realistic shots.
Character consistency IS the real game changer or at least it should be. Very few options at the moment. I complete my short thru LTX platform. Same issues.
I wanted to try Kling and I couldn't pass de verification stage. You know, when you must drag a puzzle piece on the enpty space? I simply can't drag it! Maybe it is because I use a Mac M1, I don't know. Anyway. I won't loose my time with it. Competition is there.
damn, timing sucks for you that Minimax was just released and is better than all of them. The only thing is it doesnt have tools yet... but its unlimited and free
MiniMax is mainly used for Whatsapp videos etc, meaning meant for mobile phones users with short video creation rather than long youtube video or bilibili. Maybe their target market is for Africa, since lots of Africa leaders come to Beijing these few days for better cooperation. Why need to use pcs to do video creation when your own phone processing power is enough.
My biggest issue is these sites deny any form of payment I’ve put in for a subscription. Not sure what the issue is - wish they’d take PayPal or Apple Pay
@@curiousrefuge Any debit card I've entered gets denied. So, I've tried using the cashapp option, but the site will not allow me to click on the 'enter address' portion of that screen. I've tried on desktop and phone. I reached out to their support team weeks ago with no reply, sadly.
@@curiousrefuge When I try to sign up for a subscription, the site denies any debit card I input - unsure why, when there's no issue with the bank or funds. So, I tried using the cashapp option, but the site will not allow me to click on the 'enter address' section of that page either - I've tried on a computer and a phone. I reached out to the support team weeks ago but haven't heard back, sadly.
Too bad you couldn't include the free MiniMax that was just released. It seems to be superior to all of these in motion and realism. It's only text to video right now with 25 fps and 5 second clips. But it is very prompt adherent with big actions.
@@curiousrefuge their is another channel that i subscribed ( ruclips.net/video/NpvQReYeDHw/видео.html ) and he stated KLING is top to your top video ai:
What counts to me is the outcome. So Kling is the clear winner in my eyes. Usability is a joke in all those tools, if you can't manage them you maybe should, uhm, wait until your 6th birthday. And guys please stop mentioning Sora. Aaand stop using videos from the websites... could they're heavily cherry picked. Funny I have to tell you that 😉
Thanks! At the time of recording minimax didn't (and perhaps still doesn't) have image2video. We did not consider LTX as it's a completley different workflow. However we will consider making an LTX video in the future!
I put your prompt in runway and it did not adhere to the prompt. no astronaut was displayed and the spaceships were very fake lol. I would give them a D.
I don’t get it. You’re trying to test physics with these models, but then you’re giving them prompts and pictures of stuff from outer space where the rules of physics are totally different? You’re really something, dude!
This is a great and very thorough review, thank you! Sadly however, there are restrictions with most of the online accounts I've explored out there with regard to content. I have only discovered one or two so far that don't 'impose' those limitations on their subscribers and broadly treat adults like children. There will always be those who want to push the boundaries toward what is deemed inappropriate, but that doesn't mean that we all should suffer because of that inevitability. The solution might be to give each individual the so called benefit of the doubt and the opportunity to take responsibility for themselves. I suppose I shouldn't hold my breath for that to happen in this notoriously hypocritical world...
Nice video and thanks for the effort but I would like to share my findings over most of these apps over a longer period of time testing. First of all Your imagers are showing prompts that already have been tested but simple from scratch the results are terrible. My goal is more video from image so that is what I want to focus on. Well 90% suck, have no better word. E.g. Haiper, image to video, causation woman, in focus, good quality, looking into the camera, head shoulder shot, with proper clothing. No Keyframe Conditioning Simple prompt, smile, wink one eye, kiss, no hands And there we go , totally face mutilation, ripping off her her face no wink, no smile, and hands in front of face . So I made the prompt longer like: make subject smile into camera, then give a kiss with her mouth, wink her eyes into the camera and don't show her hands. Guess what : She is sticking out her tongue and giving me huge hands But this is not only Haiper, most of the do this so utterly wrong. The least bad !!! is RUNWAY, tried it in Gen 2 mode cause I could not get in to the Gen 3 queue, it likes to zoom , so no zoom had to be added, although not 100% avoided, eyes are getting scaled bigger than the rest of the head during the video, a simple kiss with the mouth is to hard for all the AI applications, how hard can that be, also a wink with an eye or two, is to hard to understand, at least no hands and red lipstick vanish during the video . In the end NOT usable. So basically they all are beneath par, and I would never pay a cent to go into the paid version for this kind of lack of quality if you want to go for image to video. The prompt to image versions of the ones mentioned are less bad and can be tuned into a desired result more easy but it takes a lot of effort to get hands, arms legs and feet in order. Anyways, thanks for sharing
Starting to get sick of this AI content. Not because of the tools but because to creators. I watched 10 different youtube videos on 10 different subjects but somehow the creators must of been thinking collectively as they all had the same script outline. Its pathetic. Use it to add to your talent and spice up your style not replace it completely because your super lazy and don't want to do anything at all.
Please stop using "the best of the best". These people cherry pick videos you might not even be able to make. That's not TESTING or ranking, that's drinking Kool-Aid. Also, no Web site can be "close to Sora" as "Sora" doesn't actually exist except in a Wall Street prospectus deck.
Good points ,but you know, all that shit fuckin looks the same with the same emotional weirdness at the end of the day … just film real stuff with a good story
I used Gen-3, Luma AI, and Kling AI, and in my opinion, Kling AI is the best. Gen-3 worked well for me only in the unlimited subscription mode.
thanks for watching!
I wouldn't say open-source AI tools aren't as good. You have to work harder to get your results but you can also do more with them. As a VFX artist, I've been most impressed with what ComfyUI can do.
This is so true. I'm not a trained VFX artist but I've found WAY better quality with motion capture using the open source tools (Live Portrait, MimicMotion) vs the commercial tools (Mootion, Viggle)
Open Source AI tools would be cheaper compared to paid commercial tools. From experience, for every 5 AI videos, you get 1 good video. If you have 25 clips for a video, you would need to make 100 to 150 clips to get decent videos for the project. So costs will mount up quickly!
If you get a decent machine to render of the videos, it will pay for itself in the long term. Also, the ComfyUI community will develop new tools to gain more control.
Wheres the best place to learn this. I'm kinda put off by all this cmd prompt/python/bat file.
Most of the time the walkthrough assume u know what to do in cmd prompt. They talk with jargon I'm unfamiliar with.
Open source is great but comes at a biiiiiggg price - the user experience !
@@zabiisutois it a pain to download and run on your computer?
Excellent breakdown, currently Runway is my favorite due to all the extra tools that you mentioned
We appreciate you watching!
The new Chinese yext to video platform "Mini Max" is amazing made some of the best text to video generations I've seen yet, very very realistic and good movement with very little to no warping at all, and it's completely free and unlimited "as of now" only issue is no image to video yet, but once they get thay going, I will definitely rival sora
It takes so much time to load in browser 😢
We love the early results of Minimax, we are very much hoping for Image2Video soon!
@@curiousrefuge links to the ai sites would be useful thanks
@@curiousrefuge it seems to have Image to VIdeo now (HaiouAI, can't spell it ever)
Runway Gen 2-3 are still my go to generators. Although I think Gen 2 is better. G3 can go wonky sometimes. Luma is getting better too. I like it for the aspect ratio the most. And Haiper is a true underdog. For me, I have been more insisting on "big movements" in my Ai vids and Haiper has always brought "Action Packed" to my MJ images. Which is now a series.
I haven't tried Kling yet. I love this channel!!
Interesting that you are still using G2. Very cool!
Kling beats them all, especially its End frame add option, looks like not a lot of people actually know about it but it's a beast :), Haiper is the next best, especially its 3 frame keyframing, it creates some amazing results, as long as you have three somewhat consistent shots.
I am curious, how would you use the end frame in your workflow?
Did you know you can use those end frames to make loops just like in Luma?
@@Jarbrain
Theres 2 formats :
1. You can set the end frame so the animation ends on that frame.
2. You can have a start frame and an end frame.
So bear in mind theres 3 options.
A. Traditional : you start with image and prompt
B. Set image to end frame: the prompt will end on that last scene.
C. Start and end frame. The video will go from your start image to your end image
Hope that makes sense.
Good example is a kiss.
If you have a couple kissing then you stuck with kiss then pull away.
But if you set the kiss to end frame (in luma and kling).
You can have the couple coming together and finish on the kiss.
In luma theres a tiny arrow above the upload images, you have to click on that to activate "finish on this image"
The end frame on Kling is definitely a worthy feature!
@@Jarbrain just like luma or haiper, you add two very similar shots and it does the interpolation and makes the end movement very smooth and realistic looking.
Kling AI is definitely better than Runway, and it’s cheaper. Also has the end-frame tool that gets ignored. I used it constantly
I tried the free plan and it got stuck at 99% for 3 hrs....moved to Runway!
Kling is definitely the best yeah. I've tried both of them on their premium plans, and Kling destroys Runway in a lot of ways. Gen3 is such unpredictable and sometimes ends up doing just a push in. None of that with Kling.
@@sunitjoshi3573 yeah, my sister is using Kling’s free plan and she’s been stuck on 99% for days. They put free users on a queue
@@Trankilstef exactly! And there end-frame tool is extremely powerful. I usually just generate a lot of videos with the original frame, and when I find a frame that I like within those videos, I use it as the end frame. Once Kling adds camera control and more control over the animation, it’ll be enough to create films
yes for sure kling is the best
Sora should note even be mentioned at all. Sora is just a bunch of hype and marketing from OpenAI
Totally fair point!
I would put Kling, Runway Gen-3 and Luma Labs at the same level, at least for image-to-video. I've created hundreds of video clips on each one and they each have their strengths and weaknesses for different types of shots
Interesting, could you share more details. I found Luma to be pretty bad except for some start & end frame video gen
Many would agree with that!
@@nadk666 It really depends upon the starting image. I would just experiment without any prompts to start off (using image-to-video)
What do you use as your main focus now?
My experience with Kling has been really good. And I'd like to point out that Standard Mode is nothing like Gen2. It can give you weird results, but often it's very much on par with Luma or Gen3. Actually, I rarely use the Professional Mode, as often Standard result are pretty good, but also because Pro can be sooo painfully slow.
Also, in Kling you can add an end frame as well, like in Luma. The only thing is that you can't give it JUST the end frame, it always needs the first frame as well.
Luma is also pretty good, I noticed it tends to be a bit stubborn and static at times, while Kling may go a bit crazy here and there, but this way you have more options to experiment.
I'm very impressed with Gen3, but I need to play with it a bit more. Also, it should be mentioned there also two options: the normal Gen3 Alpha and Gen3 Alpha Turbo, and it's a bit like with Kling, where you have a faster/cheaper version and a more "Pro" one, but the cheaper one is often as good as the Pro. And one thing I can say about Gen3 Turbo is it's hands down the fastwst video generator I know.
Oh, and to remedy Pixverse a little, they also have a built in 4k upscale option. :)
I should also add that ComfyUI is the most pro tool today when it comes to image generation, it's light years ahead of any image generator out there (cause it can do everything those generators can and much much more, with waaaaaaay more control). Haven't used ot for video though. Yet. ;)
You're right, I was kinda angry when he said 'Standard Mode' is comparable to Gen2 Runway. Even Standard Mode is by far superior to Runway Gen2.
We appreciate the feedback!
Way better video than the last one (Luma one)! 👌
Really great in-depth analysis on all key parameters 👏
Thanks for watching!
10:35 her right eye! I'm practically in tears!!
Thx for watching!
I have recently been trying Minimax which is really impressive and it's free for now. Unlimited video generation and relatively fast. Runway is no doubt fast and of good quality but i feel minimax might be a game changer
It's a competitive space for sure!
Kling is probably the best, but their terms state that in order to use their videos in advertisements or for commercial purposes, it must be noted that it was created by Kling or a watermark must be included, which could be problematic when working with a larger company. What do you think?
Certainly makes it difficult for commercial use!
Thank you, this was quite helpful.
Glad you enjoyed this!
Minimax has excellent prompt understanding and the others are useful in other ways. Disappointed with Luma though.
We appreciate you engaging with us!
Could you please put timestamps in the video? It would be much appreciated. Subbed and will hopefully join an in-person event soon
Absolutely! Just added them. :)
EXCELLENT video, I've been researching each of these tools as well. You've summed these up wonderfully.
Glad it was helpful!
Which is the best video to video ai generator?
Watch the video you bot
Check out the video for yourself :)
@@curiousrefuge hey, of course I did see the whole video like others in the past.
I couldn’t find you exploring the video to video generators how they hv evolved so far. Hence asked.
I think image to video runway gen3 alpha and kling ai the best ai tool available on internet platforms
Makes sense!
The conversion from 120, 60 or 30 to 24 fps on Davinci Resolve it's perfect and cost nothing, so that criteria of the 24 fps says nothing about the AI Video Generator it's useless
We disagree that FPS conversion and changes are a useless point.
Kling-ons 😂😂😂 definitely going to use that one going forward!
We think you'll love our Dino-Cat video (made with RunwayML)
Feel free to share!
@@curiousrefuge ruclips.net/video/1iP_KgvgjGs/видео.htmlsi=Heu7wAJNg1xTDUcj
Which would be the best AI Video Generator for creating drawing type animations such as anime or comic book characters? All of the examples I've seen on youtube are usually 3D Pixar type or photo realistic type.
Likely a mix of MJ and Runway. Pika does well too with anime
@@curiousrefuge I'll check them out. Thanks so much!!
It would be nice to see your prompts
Thanks, next time we will add it! :)
Thanks for the video! Why does 24 frames per second is better then 30 second?
Isn't the fact that there are more frames in the video make it better?
Great question! 24fps is the frame rate used to shoot the majority of movies. This is typically industry standard and will help in creating more cinematic and realistic shots.
Great video!
Thanks !
😊
Our pleasure!
Have you tried KREA AI and Leonardo AI? How do either of these stack vs the ones you covered in this video?
Yes we've tried them all! :)
Thanks a lot! Could you tell me please how you make that background in your videos?
It's a projector :)
Character consistency IS the real game changer or at least it should be. Very few options at the moment. I complete my short thru LTX platform. Same issues.
For this reason, many people rely on character consistency via MJ and then use Image2Video :)
I wanted to try Kling and I couldn't pass de verification stage. You know, when you must drag a puzzle piece on the enpty space? I simply can't drag it! Maybe it is because I use a Mac M1, I don't know. Anyway. I won't loose my time with it. Competition is there.
The M1 likely won't stop you, but rather a verification page may be due to your region.
Can you please tell which one is best which include all and create realistic one
Watch and you'll see which styles that you prefer and like most!
Shame Runway is quite expensive, as it seems the best all-in-one solution.
Hopefully it all comes down a bit :)
damn, timing sucks for you that Minimax was just released and is better than all of them. The only thing is it doesnt have tools yet... but its unlimited and free
No image which is pretty bad :(
It doesn't have Image2Video yet! That's one of our qualifiers ;).
gotchya
MiniMax is mainly used for Whatsapp videos etc, meaning meant for mobile phones users with short video creation rather than long youtube video or bilibili. Maybe their target market is for Africa, since lots of Africa leaders come to Beijing these few days for better cooperation. Why need to use pcs to do video creation when your own phone processing power is enough.
Interesting...we didn't know that!
My biggest issue is these sites deny any form of payment I’ve put in for a subscription. Not sure what the issue is - wish they’d take PayPal or Apple Pay
What do you mean deny?
@@curiousrefuge Any debit card I've entered gets denied. So, I've tried using the cashapp option, but the site will not allow me to click on the 'enter address' portion of that screen. I've tried on desktop and phone. I reached out to their support team weeks ago with no reply, sadly.
@@curiousrefuge When I try to sign up for a subscription, the site denies any debit card I input - unsure why, when there's no issue with the bank or funds. So, I tried using the cashapp option, but the site will not allow me to click on the 'enter address' section of that page either - I've tried on a computer and a phone. I reached out to the support team weeks ago but haven't heard back, sadly.
Can you upscale Kling video and change the framerate to 24 in Topaz Video?
You certainly can change the framerate in Topz!
minimax? P.S. your content is ❤🔥
Minimax doesn't have Image2Video...yet!
Too bad you couldn't include the free MiniMax that was just released. It seems to be superior to all of these in motion and realism. It's only text to video right now with 25 fps and 5 second clips. But it is very prompt adherent with big actions.
Keep in mind ---- Image2Video is a key qualifier for us in which Minimax isn't there yet ;)
the video that we were waiting for ;)
Glad you enjoyed it!
so of KLING vs Runway: seem that i have to find a video on youtube that you did on KLING?
What do you mean?
@@curiousrefuge their is another channel that i subscribed ( ruclips.net/video/NpvQReYeDHw/видео.html ) and he stated KLING is top to your top video ai:
kling is the best one
A fellow Klin'on :)
What counts to me is the outcome. So Kling is the clear winner in my eyes. Usability is a joke in all those tools, if you can't manage them you maybe should, uhm, wait until your 6th birthday.
And guys please stop mentioning Sora. Aaand stop using videos from the websites... could they're heavily cherry picked. Funny I have to tell you that 😉
We appreciate the feedback!
Its totaly true not a feedback
I try all off them with same picture and same text, Luma also good.
I think Runway is the best so far. And, you forgot LTX, Vidu, and Minimax.
Thanks! At the time of recording minimax didn't (and perhaps still doesn't) have image2video. We did not consider LTX as it's a completley different workflow. However we will consider making an LTX video in the future!
you forgot vidu..and what about image to video??
We'll have to test out Vidu next time !
super video 🔥
Thanks.
which one is the best for anime or cartoons?
We like Pika and Runway for that type. Go ahead and test them!
@@curiousrefuge a try vidu I think is the best for image to video. Do you have a video about vidu?
I put your prompt in runway and it did not adhere to the prompt. no astronaut was displayed and the spaceships were very fake lol. I would give them a D.
Let us see it in our discord!
AWESOME Vid! Thank You!
Glad you enjoyed this !
Whats The best Free ones?
Unfortunately free/freemium/paid is a revolving door so it's hard to say at this point in time.
Which tier is Minimax? :)
Doesn't have Image2Video...yet!
Any of these do children’s animations that last about 10 minutes? Also, do any of these Ai voice generate?
Yes some of these do - like runway!
@@curiousrefuge I will have to check it out. Thanks
When generating images....prompt....Photo....or photographic ...realistic....cinematic
Thanks for watching!
Kling is the best.
Tons of people are Kling'ons :)
need unlimited option for kling, i really like the professional mode but its expensive
I'm Haiper Hedra and Genmo
Those are good tools!
Actually based on your test Kling did a more better job than all of them.
Kling is great!
I don’t get it. You’re trying to test physics with these models, but then you’re giving them prompts and pictures of stuff from outer space where the rules of physics are totally different? You’re really something, dude!
The physics testing is moreso seeing if the generations meet expectations of viewer/prompter.
Does anyone in the comments make videos for training or education? If so, which generators worked best for you?
25:00 so sad moment 😂😂😂
thanks for watching!
This is a great and very thorough review, thank you! Sadly however, there are restrictions with most of the online accounts I've explored out there with regard to content. I have only discovered one or two so far that don't 'impose' those limitations on their subscribers and broadly treat adults like children. There will always be those who want to push the boundaries toward what is deemed inappropriate, but that doesn't mean that we all should suffer because of that inevitability. The solution might be to give each individual the so called benefit of the doubt and the opportunity to take responsibility for themselves.
I suppose I shouldn't hold my breath for that to happen in this notoriously hypocritical world...
We appreciate your perspective.
@@curiousrefuge Thanks
Really, you didn't mention the Generator who every one is talking about ! MiniMax
Unfortunately it's only text2video right now!
Can someone tell me is it free all the softwarees ?
Most of the software are freemium or paid.
Criteria #10... Do the hands suck?
;)
haha true...the classic ol' 6th finger!
All website is for commercial use
Check all the T&C's!
my eyes prefer Klings movement
Kling is definitley popular!
@@curiousrefuge 99% done, and just one week to wait for a 2 sec video 😆
It’s only Luma, kling, and runway. Everyone else is awful
Heard about minimax..
Pikalab so stagnant now, rip to those who bought yearly plan
We'll see what happens with Minimax soon!
@@curiousrefuge if it does image to video it's number 1
UP TO 10 SECONDS IS A JOKE!!! HOW AM I SUPPOSE 2 MAKE A MOVIE WITH 10 SECONDS DUHHHH THESE ALL STINKY DINK!!!
You have to edit clips together and scenes!
You forgot leonardo!
Leonardo is certainly a good one! We are hoping they come up with an update soon!
Surreality could be a factor ?
We'll add that for next time :)
I like your Pajamas. Are they AI?
Haha, not pajamas :)
Minimax ai is missing
Minimax doesn't have image2video yet ;)
pyjamas
Haha looks like it huh?
Rocking the Hugh Hefner look 😂 I love it
For me Heygen is better than Runway ML
We'll have to keep testing :)
He said "objective opinion" lol, classic
haha, woops!
Why try to be perfect with these models? Thats a waste of time, bc the tech is not there yet. Just embrace the flaws and create :)
Embrace the chaos! :)
Kling ai is best
Kling is certainly a good one!
Nice video and thanks for the effort but I would like to share my findings over most of these apps over a longer period of time testing.
First of all Your imagers are showing prompts that already have been tested but simple from scratch the results are terrible.
My goal is more video from image so that is what I want to focus on.
Well 90% suck, have no better word.
E.g. Haiper, image to video, causation woman, in focus, good quality, looking into the camera, head shoulder shot, with proper clothing. No Keyframe Conditioning
Simple prompt, smile, wink one eye, kiss, no hands
And there we go , totally face mutilation, ripping off her her face no wink, no smile, and hands in front of face .
So I made the prompt longer like: make subject smile into camera, then give a kiss with her mouth, wink her eyes into the camera and don't show her hands.
Guess what : She is sticking out her tongue and giving me huge hands
But this is not only Haiper, most of the do this so utterly wrong.
The least bad !!! is RUNWAY, tried it in Gen 2 mode cause I could not get in to the Gen 3 queue, it likes to zoom , so no zoom had to be added, although not 100% avoided, eyes are getting scaled bigger than the rest of the head during the video, a simple kiss with the mouth is to hard for all the AI applications, how hard can that be, also a wink with an eye or two, is to hard to understand, at least no hands and red lipstick vanish during the video . In the end NOT usable.
So basically they all are beneath par, and I would never pay a cent to go into the paid version for this kind of lack of quality if you want to go for image to video.
The prompt to image versions of the ones mentioned are less bad and can be tuned into a desired result more easy but it takes a lot of effort to get hands, arms legs and feet in order.
Anyways, thanks for sharing
We appreciate you sharing all your findings!!
Minimax is the best
It's great but only text2video currently!
Dude, making your video in your pyjamas? Seriously?
Don't like the drip?!
Are you wearing your pajamas?
No :)
runway A? lol
Like Runway the best?
Seni güzellik
thanks for watching!
Why not just change the frame rate in premier... dont be stupid.
It's certainly an option but has it's pros as cons.
Please stop saying 'it begs the question". BTW You should find out what it really means. You might be surprised.
We appreciate the feedback!
Starting to get sick of this AI content. Not because of the tools but because to creators. I watched 10 different youtube videos on 10 different subjects but somehow the creators must of been thinking collectively as they all had the same script outline. Its pathetic. Use it to add to your talent and spice up your style not replace it completely because your super lazy and don't want to do anything at all.
We are sorry you feel this way :(. We appreciate you watching !
What is a generation? Can y you unsubscribe at any point?
Genrate is when you let the software create something for you.
Please stop using "the best of the best". These people cherry pick videos you might not even be able to make. That's not TESTING or ranking, that's drinking Kool-Aid. Also, no Web site can be "close to Sora" as "Sora" doesn't actually exist except in a Wall Street prospectus deck.
Fair point!
Good points ,but you know, all that shit fuckin looks the same with the same emotional weirdness at the end of the day … just film real stuff with a good story
Storytelling is the heart of all good content!
Luma is complete shi.
Really? We like it!
none of these are worth my money or time
Understood!