Generating video from a still image

In this video, we're going to continue our exploration of Gen. AI video generation, this time using an image. In the last video, we generated video from a text prompt using Gen. 3 Alpha, which is only accessible if you are on a paid plan. With Gen. 3 Alpha Turbo, you can use this on a free plan and then generate video from an image. Let's see the limitations that Gen. 3 alpha turbo has and compare that to the paid Gen. three alpha, which will produce a higher fidelity result. So I'm here on my free account. And like we did earlier, I'm going to go to text image to video. I can see the number of credits that I have available here on the top right hand corner. And keep in mind that if you actually click on here, it will tell you how much credits are needed to generate a second of video. So it's five credits for every second of Gen. 3 Alpha Turbo, more for Gen. 3 Alpha as it produces a higher quality result. You could also produce video with Gen. 2 if that is your desired result and may be good for some use cases. So in order to get started with Gen. 3 Alpha Turbo, I'm going to drop an image here. And you should know that in your exercise files, there's a 1.4 image to video folder, which has some images that I've supplied that were produced with MidJourney that you can use for this lesson. In this case, I'm going to use this shot here of a black mannequin head with red flowers and gold. I'll upload that to Runway. Since the image uploaded is 16.9, there's no need to crop it. However, if your format isn't the dimensions of 16.9, you'll have to crop your image in order for it to be exported. Underneath the video, we have a choice of whether to use this image as the first or the last frame of our video. I'm going to choose to use it as the first frame. And then we're going to enter a text prompt. And what we want to keep in mind is the movement or motion that we would like out of this image rather than trying to define everything that's in this shot. So I'm just going to paste some text that I already have saved, which is, red flower petals fall to the foreground in slow motion as the camera dollies out. The woman opens her eyes. Keeping in mind that Runway, Alpha Turbo, and Runway by default will generate 10 seconds of video unless we change it here from the menu. I'll change this to 5 seconds. I'll also pay attention to the amount of credits that I have left. And I'm now going to choose Generate. Since this is Alpha Turbo, it's going to be faster than generating with Gen. 3 Alpha, but again, not producing as good results as if we did use Gen. 3 Alpha. So here's the video. Again, I produced five seconds. I'll just press the play button. You can see there's some nice slow moving pedals there. Didn't quite pick up on opening up the subject's eyes. This again is a great reason for why keeping your video length short when you first start in order to make adjustments as needed. Now with your generated video a couple other things is that so that you could see the full prompt that you used and copy that over, not to mention the seed value, which we became familiar with when we were producing video-to-video generation, that we could use in producing other videos so that we have a similar style throughout the shots that we generate. In addition, we can download this here with a watermark on our free account, or share this as desired and even mark it as a favorite. In this case, I might want to try to reuse the prompt, which I can do again. It will copy that prompt. It still has my image loaded up. And I could draw some more detail here. I could say the woman blinks and opens her eyes and choose to generate again. And here is the result of the new video. Unfortunately, the eyes isn't open, so I might want to try a few other text prompts in order to make this happen. Now this is worth comparing with Gen. 3 Alpha and not Turbo. So what I'm going to do is take this prompt, copy it and meet you over in my paid account so that we can compare the differences. So I'm over here using Gen. 3 Alpha where I produced the same five seconds of video trying to get the statue woman to open her eyes to no result. But here is the result of the Gen. 3 Alpha. do see a bit higher quality in my opinion with the actual flower petals themselves. But overall you should note that this took twice as long to generate and also cost a little bit more with our credits. So something to keep in mind compared to that of the free Turbo account. So there's a little bit about Gen. 3 Alpha and Gen. 3 Alpha Turbo which is available for free when you're doing image to video generation. Keep in mind, Runway also has a Gen. 2 model and there are a few things you can take advantage of here, including a paintbrush to define the areas of motion, not to mention the ability to define how much movement you want in your shot. Gen. 3 in particular allows you to define the type of motion through a text prompt, which Gen. 2 wasn't as good at. But you should check out Gen. 2 and see the type of video it does produce as a comparison to Gen. 3, getting more familiar with some of the video tools that are available. Last but not least, just keep in mind, Runway is not the only video model available. I did mention some in the last movie, including Pica Art, Dream Machine, not to mention DECO here. I've also included some images as well as some video outputs that I got using Gen. 3 that you you have available in the exercise files to take advantage of.