With synthetic the worries are much bigger. As an example of what can happen consider tryptophan made in Japan during the late 80's. Genetically modified bacteria may have produced highly poisonous tryptophan. The poison, (the molecules in error) were so similar to tryptophan and such a small part of the mix that very sensitive spectral analysis would not detect the poison.
Since the Japanese manufacturer (Showa Denko) of the bad TRP isn't talking, I presume we don't really know what sort of process analyses were regularly run. I say this because analytical techniques in common use today should be able to uncover problems like this. Liquid Chromatographic quantitation of impurity peaks could tell you that something has changed in the process if a new peak suddenly appears. In the tryptophan case, the impurity, 1,1'-ethylidenebis[tryptophan], was over twice the molecular weight of tryptophan, and would have eluted far from the tryptophan peak. What we don't know is even if some nameless Chinese manufacturer runs regular LCs, and they see a new peak, would they halt production in order to sort it out? Or would they say "it's way less than 1%, so what?"
Yet, in comparing the cell culture resveratrol to the knotweed extract, how do you know that the ostensibly "natural" knotweed extract doesn't have something bad in it? There's really nothing "natural" about running flash chromatography with a bunch of weird solvents on an herb, and concentrating the hell out of a fraction of it, then ingesting gram+ doses for the rest of your life. It's just a different bunch of technology. In the end, I would want either product, herbal or cell culture, (or chemically synthesized) to have all impurities that are present in physiologically relevant quantities to be characterized.
For what it's worth, the Showa Denko tryptophan disaster was an extremely unusual event. Sometimes airplanes crash, but we keep flying on them. At the start of the thread, Anthony mentioned an 8X price difference in the two products. It that's the case, then I'll most likely go for the cheaper one. However, I'd heard elsewhere that the fermentation resv was way cheaper than that. If it were actually cheaper than the herbal extract, and demonstrably more pure, I'd probably go that route.