









I’m hitting a frequent bottleneck with Kling 3.0. Even in silent shots, characters often trigger unprompted lip-sync—specifically that weird, unintelligible babble that sounds like Chinese.
It’s not constant, but it happens enough to kill the workflow and waste a ton of credits on ruined generations. I’ve already tried the usual "mouth closed" commands and negative prompts for talking/lip-sync, but the model keeps overriding them.
If you’ve dealt with this, how did you get around it? Looking for any successful experiences or workarounds.
Attached clip of the error for reference.
I’ve been using Kling Ultra for a while, but recently I feel like the video quality has been getting worse.
Not sure if it’s just me getting pickier, or if something actually changed on Kling’s side.
Also, I’ve noticed that when I change a character’s outfit, the facial consistency sometimes drops, which is a bit frustrating.
I haven’t tried Seedance yet, so I’m curious — for those who’ve used both, how do they compare?
Any noticeable differences in quality or workflow?
Would love to hear real experiences before I switch.
​
I’ve been using Kling Ultra for a while, but recently I feel like the video quality has been getting worse.
Not sure if it’s just me getting pickier, or if something actually changed on Kling’s side.
Also, I’ve noticed that when I change a character’s outfit, the facial consistency sometimes drops, which is a bit frustrating.
I haven’t tried Seedance yet, so I’m curious — for those who’ve used both, how do they compare?
Any noticeable differences in quality or workflow?
Would love to hear real experiences before I switch.