Searching the entire internet, the only open-source application capable of handling close-up video details for lip-syncing😜. The image quality is very refined, basically on par with Keling, but the accuracy of articulation is slightly inferior. To a certain extent, it can replace Keling for lip-syncing. The main drawback is that it uses full-frame sampling, making it a bit slow... An additional advantage is that because it uses full-frame sampling, prompt words can influence the final effect, such as adding micro-expressions.
It takes 7 minutes to complete 5 seconds at 1280x720 resolution. At 1024 resolution, it is much faster. Currently, it cannot guarantee whether lengths longer than 10 seconds can be achieved, as it consumes a lot of VRAM.