You don't say what type of camcorder your source video is coming from -- but there is no reason both versions (or any versions) of this program should not produce identical results when using the same settings and the same footage.
The basic codec and workflow system this program uses has not been changed.
As Steve points out, there should be no reason why Export/Share should differ. The one difference between the full-paid version and the trial is the watermarking in the trial's export.
Please tell us about your source footage, your Project Preset and your exact Export settings. Also, if possible, tell us how these differ from the Project in PrE 3. The difference will be in that data somewhere. It might be easiest to do a table with two columns: PrE 3 & PrE 8, with four rows: Assets, Project, Export/Share settings and results differences.
Before I undertake that project -- estimated to consume 3-4 hours, replete with output samples -- I would appreciate someone verifying whether the watermarking process of the trial version does or does not affect fidelity.
Source footage is an m2t file direct from a Kodak digital camcorder. Output has been in WMV, MPG, and Flash -- all show unacceptable artifacts, pixelation, and fuzziness from PREL 8.
Export quality also depends on the project setting which you are using. Could you specify more clearly about your project settings, export preset which you are using and the capture mode of your camcorder?