A comparison by intended output would make this list easier to use. A five-second visual concept and a narrated product explanation place different demands on a tool. For the latter, I would test script revision, readable on-screen text and the ability to correct one scene without starting over. Leadde is relevant to that explanation workflow, but I would not treat it as another underlying model in the same ranking. Would you consider separating model-level comparisons from complete authoring workflows?