"Is your mobile phone missing an AI that can automatically generate videos?
Short Video for free are really not difficult these days. Recently, many friends have asked me how to use AI to make decent videos. In fact, there are many methods. The key depends on how much time you are willing to spend exploring.
Let's start with the simplest one, use ready-made AI tools. For example, platforms like Coze specifically help you turn text into videos. You just need to write a script, and it will automatically match pictures and dub. A friend of mine tried it a while ago and finished three Short Video in five minutes. He told me that as long as the script is clearly written, AI can hardly make mistakes.
However, most of these tools require registration of an account. If you don't want to be bound, there are open source options. Software like Runway ML is a bit complex to operate, but it is completely free. My colleague spent two days learning, and now he can adjust the parameters himself and generate the video clips he wants. He said that the most troublesome thing is to find material. He has to search for a bunch of pictures and video materials online.
Another compromise is to use AI tools to generate some content. For example, let AI write scripts first, and then use free software to synthesize videos. I tried doing this recently, and it worked pretty well. Send the script to ChatGPT first to change it to make it more colloquial. Then find Bard to generate several sets of keywords and go to Pexels to download matching materials. Use CapCut to cut the material and match it with some AI-generated music. The whole process took only half an hour for a video.
There are many points to pay attention to. First of all, the pictures generated by AI sometimes get stuck. I have seen someone make videos using Stable Diffusion, but the pictures keep jumping. He said the solution was to try a few more times or manually fill the frame himself. Secondly, dubbing is prone to problems. The text read by AI sometimes breaks wrong sentences, which has to be adjusted manually later.
There is also a trick to mix different tools. For example, using D-ID to generate a real avatar, and then using Runway's animation effect. I know a blogger who did this. He said it would make the video more vivid. However, be careful about copyright issues. Some AI-generated materials cannot be commercially available.
If you really don't want to go through it, consider finding ready-made templates. Design software like Canva now has video templates. You just need to fill in the content and it will automatically typeset. I tried a news broadcast template, put text in it, and generated a video in a few seconds. Of course, this template has low degree of freedom and is suitable for quick release.
To be honest, free tools always have limitations. I suggest deciding what type of video you want to make first. For example, if it is a verbal broadcast class or a scene class, it will be easier to find the corresponding tool. I have a client who specializes in food videos. He directly uses AI to generate close-ups of the ingredients and then takes pictures of the cooking process himself. The effect is great.
AI tools are now updated quickly, and there may be a simpler solution after a while. The key is to try more and find the right combination for you. I have met many people who were confused at first, but after trying a few tools, they gradually found the trick. "
