A user tested the Seedance 2.5 video generation model using a short fashion and editorial sequence featuring a single character in various outfits, compositions, and visual styles.
- The test aimed to observe how well the model preserves coherent character identity while other elements change simultaneously.
- The sequence included close-up portraits, fashion-focused shots, graphic compositions, outfit changes, and red-and-black editorial scenes.
- Overall character identity remained fairly coherent across the sequence, though facial features shifted slightly between shots.
- Small inconsistencies became more noticeable when clothing, framing, movement, and composition all changed together.
The testing suggests that evaluating how much visual variation a video model can handle before character identity breaks down is more useful for creative workflows than simply assessing frame realism.