I'm working on an audio book and I want to use signal processing to indicate when the narrator is saying something or thinking something. These tracks are anywhere from 4 to 45 minutes long. I want to be able to use effects on sentences and/or individual words to enhance the dialogue.
 
What's the best way to do this? Right now I can't get my head beyond riding the effect level throughout. There has to be a better way. Thanks.

A good question. Perhaps clone the track, and apply the effect to that track, and use volume automation on the two tracks?

if the effect is say reverb, put your reverb on a bus, create a send to the reverb bus from the track, create an envelope for that send in the track, automate the envelope with nodes.

timidi
if the effect is say reverb, put your reverb on a bus, create a send to the reverb bus from the track, create an envelope for that send in the track, automate the envelope with nodes.

I do that all the time with delay sends, to for example only apply delay to the last word of a phrase etc. 

If it's a single effect, and it has an automatable wet/dry control, then the easiest and cleanest way is to automate that parameter via a track envelope on the dialog track.
 
If it's a combination of effects, create a bus for them and automate either a send envelope on the track or a volume envelope on the bus. This is a little trickier than the first method because you might have to compensate for the volume increase when the fx bus is active. 

And if they consist of different effects at different places, with no obvious tail to consider, split the track into clips and apply clip Fx

Thanks for the response. I'm going to test out each of these ideas and see what's best. I have to hit the books to get myself up to speed on most of it because I seldom use any type of automation. Jonesey's idea about clip effects sounds interesting and easy, but I'll try everything you guys suggest. I'm off to do some book learnin' now.

I've done this mostly as James mentions above ... I ended up multing to separate tracks so I could adjust the effect in one place instead of having to adjust for each clip. 
 
So for eaze of tweaking along the way, you could:
- Automate sends to a single effect instance.
- Split the bits that get the effect to a separate track with a single instance of either an insert or send effect.
 
OTOH, if the parms will be different for each part, then clip FX is the way to go.