Liking OpenArt more and Dzine less
I've been working hard to create my next short video, using both OpenArt and Dzine mainly, and Galaxy for simple things just to use the many credits I have there. I am liking OpenASrt more and more, and Dzine less and less. The big thing Dzine has going for it is the ability to lip sync up to four separate characters simultaneously, all in the same scene, and even talking over each other. And their lip sync itself is excellent, intelligently accompanied with appropriate body language. But especially for longer videos, I find that for some crazy reason they often make insane changes to the appearance of the character. I have a clean-shaven guy, and suddenly, over the course of just two frames, it gives him a beard! A woman with long hair combed behind her head will suddenly have her hair jump to the front of her body. A straight road going off into the distance will suddenly become sharply curved. This mess is just from lip sync! Why would lip sync put a beard on a man, or sharply bend a distant road??? Also, OpenArt's version of NanoBanannaPro is much better implemented in OpenArt than in Dzine. They let you give it up to ten reference images to guide the image generation in terms of style and atmosphere. As far as I can tell, Dzine allows only one. And of course OpenArt lets you place their Character 2.0 reference characters in an image, while with Dzine the only practical way is to upload an image already containing the characters, and hope it can keep them straight. And to top it all off, Dzine can make finding earlier generations a nightmare. I generate an image or video, change to a different AI technique, and then when I come back to review and maybe download the original generation, it's gone! It's actually still there, but it's agony hunting around to try to find it. So I have definitely made OpenArt my go-to platform, maintaining a small Dzine subscription only for multiple-character lip sync.