The first Transformers movie did their voice actor recordings very differently to most big budget cgi films.
The majority of big budget CGI films create the characters, sets etc then have the voice actors come in, and record their dialogue.
From there the film is animated so as to correctly match up to the voice actors dialogue, which is also heavily used for timing of actions and movements as well.
This process allows for ad-libbing, script changes and character traits to be introduced by the voice actors.
Transformers got the voice actors in after all was said and done minus final editing (this was most likely due to the constant changing nature of the script, and the very long render times) and the voice actors had to match their dialogue to the pre rendered animation.
This limits the voice actors to only using dialogue that fits with the animation, leaving little to no room for changes to dialogue or improvisation and allows for no input to the characterisation of the character from the voice actor.
Cheaper budget animated films often skip precise lip synching as it is extremely time consuming and difficult and requires many more animators hours to get correct, instead they rely on basic lip synching, which is based off 15-25 mouth movements shared amongst word sounds (for example, the letters/sounds "b", "p" and "t" share a mouth movement).
So if the voice actors are on set already perhaps they will be recording dialogue the more conventional way, before rendering begins?
Looking For: Wreckers Saga TPB Collection (with Requiem)