Hi! Thanks for your comment. The progress indeed is amazing, although - as you mentioned - the results are still not ideal. We believe that judging the AI filmmaking software should be done by comparing the current state with what was available earlier - and the progress is visible. Runway has introduced Gen-3 image-to-video 4(!) days after Kling AI went public. It makes us think Runway was in a bit of a hurry here :) And this leads us to expect some improvement in Runway's image-to-video function very soon.
What would be great is showing how many credits it took to get the final shots...😅 The amount of times gen3 just completely ignores prompts and says you know what, im.doing my own thing.
Welcome to the channel! Thanks for your question, although the answer is not simple. First of all, there's a huge difference between text-to-video and image-to-video methods of generating video. When it comes to achieving cinematic, realistic scenes (and this is what we are focused on) that reflect cinematographer's intentions, the image-to-video is still the way to go. All the scenes you can see in our video were generated in one or two takes. Two things are important here. The first one is how good your initial (first) frame is. Are there any elements in this image that you don't want to be animated (but AI can choose to animate them)? Is the camera angle and lighting you wanted and do you know how you want it to develop in the scene? The second one is what you ask AI to do in your scene. If you expect a lot of animation in the scene and, on top of that, you put a complicated camera movement in the prompt, chances are that AI (be it Runway Gen-3 or its competition) will lose it at some point. My advice would be to carefully create the first frame image, repeat the ambiance, color tones, mood, etc in the animation prompt and make sure you don't ask for too much to be going on in the scene. For example, if you want to have a lot of animation in the scene, ask for a still camera shot. If not much is to be animated (actors, props), you can prompt a more spectacular camera movement. Also, make sure that the camera movement you ask for will not spoil the scene as it would in real life. For example, if your actor is backlit in the initial frame image and you ask for camera panning, you may get blinded by the sun, just as it would happen on a traditional film set. Apologies for a long reply - unfortunately, I have just touched the tip of the iceberg here. We'll cover many of these topics in our future videos here and in the full masterclass course (coming soon) so stay with us. Good luck with your AI films!
Hello and welcome to the channel! Thanks for your opinion. We pay quite a lot of attention to the type of background music we choose and we actually like this score pretty much :)
Hi, thanks for chiming in! That's really a tough question and a close call. In some cases Kling is spot on, in others Runway is so much better. At AI Filmmaking Academy we focus mainly on realistic, cinematic footage. For such job there is no easy, jack of all trades answer. When we create a movie, we usually use more than one software. I'd say both offer a high level of animation proficiency but we would not be happy to be forced to use only one of them for all tasks :)
Welcome to the channel! Sorry to hear that, but let me suggest that you revise your prompts - their structure and all the info that should be included to achieve the desired result. Usually, the problem lies in the prompting. Good luck!
Welcome to the channel! Yes, you can, as long as the person you want to lip sync is realistically looking. The good thing is that Runway does not "freeze" the person and only moves lips - the rest of the animation you asked for is still going on when you add the text and lips movement.
Hi! Welcome to the channel! Thank you for your comment, we appreciate it. It's more of a "news" type of video, not a tutorial, so we decided we could pump the music volume a bit :) But you may be right that we pumped it a notch too high - we'll do better next time. Have a great day!
I have not replied as we waited for others to chime in. Here at AIFilmmaking Academy we focus exclusively on realistic cinematic movies, so cartoon/animation is an area we would not like to offer guidance. Nevertheless, I have watched the trailer, and I enjoyed the characters (although I can see some issues with their consitency), the music, and the camera movements. Keep on creating, let's see how the story develops! Good luck!
i just have started playing with the new option and sill there are a lot of poor quality results - but is better than before - ) progress is noticable
Did you have to pay?
Hi! Thanks for your comment. The progress indeed is amazing, although - as you mentioned - the results are still not ideal. We believe that judging the AI filmmaking software should be done by comparing the current state with what was available earlier - and the progress is visible. Runway has introduced Gen-3 image-to-video 4(!) days after Kling AI went public. It makes us think Runway was in a bit of a hurry here :) And this leads us to expect some improvement in Runway's image-to-video function very soon.
Gen-3 Alfa is a paid option. The least expensive option to test the image-to-video feature costs $12/month.
What would be great is showing how many credits it took to get the final shots...😅 The amount of times gen3 just completely ignores prompts and says you know what, im.doing my own thing.
Welcome to the channel! Thanks for your question, although the answer is not simple. First of all, there's a huge difference between text-to-video and image-to-video methods of generating video. When it comes to achieving cinematic, realistic scenes (and this is what we are focused on) that reflect cinematographer's intentions, the image-to-video is still the way to go. All the scenes you can see in our video were generated in one or two takes. Two things are important here. The first one is how good your initial (first) frame is. Are there any elements in this image that you don't want to be animated (but AI can choose to animate them)? Is the camera angle and lighting you wanted and do you know how you want it to develop in the scene? The second one is what you ask AI to do in your scene. If you expect a lot of animation in the scene and, on top of that, you put a complicated camera movement in the prompt, chances are that AI (be it Runway Gen-3 or its competition) will lose it at some point. My advice would be to carefully create the first frame image, repeat the ambiance, color tones, mood, etc in the animation prompt and make sure you don't ask for too much to be going on in the scene. For example, if you want to have a lot of animation in the scene, ask for a still camera shot. If not much is to be animated (actors, props), you can prompt a more spectacular camera movement. Also, make sure that the camera movement you ask for will not spoil the scene as it would in real life. For example, if your actor is backlit in the initial frame image and you ask for camera panning, you may get blinded by the sun, just as it would happen on a traditional film set. Apologies for a long reply - unfortunately, I have just touched the tip of the iceberg here. We'll cover many of these topics in our future videos here and in the full masterclass course (coming soon) so stay with us. Good luck with your AI films!
Good video!
Welcome and thank you for your comment! Stay with us, we'll be publishing next one today :)
please change your background music choice
Hello and welcome to the channel! Thanks for your opinion. We pay quite a lot of attention to the type of background music we choose and we actually like this score pretty much :)
Thank you very much for your efforts !!!
Welcome! Thank you! Stay tuned, more videos are coming very soon :)
what the best, runway 3 or kling?
Hi, thanks for chiming in! That's really a tough question and a close call. In some cases Kling is spot on, in others Runway is so much better. At AI Filmmaking Academy we focus mainly on realistic, cinematic footage. For such job there is no easy, jack of all trades answer. When we create a movie, we usually use more than one software. I'd say both offer a high level of animation proficiency but we would not be happy to be forced to use only one of them for all tasks :)
"very good"
Thank you and welcome to the channel!
I spent all my credits in one night, all bad generations… very disappointed
Welcome to the channel! Sorry to hear that, but let me suggest that you revise your prompts - their structure and all the info that should be included to achieve the desired result. Usually, the problem lies in the prompting. Good luck!
CAN you do lip sync in gen3?
Welcome to the channel! Yes, you can, as long as the person you want to lip sync is realistically looking. The good thing is that Runway does not "freeze" the person and only moves lips - the rest of the animation you asked for is still going on when you add the text and lips movement.
@@AIFilmmakingAcademy thanks
You're welcome!
That background music is too much! Yikes...
Hi! Welcome to the channel! Thank you for your comment, we appreciate it. It's more of a "news" type of video, not a tutorial, so we decided we could pump the music volume a bit :) But you may be right that we pumped it a notch too high - we'll do better next time. Have a great day!
@@AIFilmmakingAcademy It's still a great video!
Maya Vanam (Mysterious Forest) | Teaser 2 | Animated | Comic | Web Series
th-cam.com/video/IJtww6Iqo8c/w-d-xo.htmlfeature=shared
Guide my work 😊
I have not replied as we waited for others to chime in. Here at AIFilmmaking Academy we focus exclusively on realistic cinematic movies, so cartoon/animation is an area we would not like to offer guidance. Nevertheless, I have watched the trailer, and I enjoyed the characters (although I can see some issues with their consitency), the music, and the camera movements. Keep on creating, let's see how the story develops! Good luck!