It depends on what "quality" service you want, whether it is simply at the current speed but with dual tracks or if you want semi-express service once you get into Brampton and want upgraded speeds along the route.
I also agree that the speed improvements are definitely worthwhile, not just due to their potential for increased ridership but also because they will reduce operating costs by reducing equipment demand.
With our current 2+ hour schedule, a round trip would probably take 5 hours (including layover), so 5 trainsets would be required to maintain hourly service. But if the travel time is dropped to 1h45 or less, we would only need 4 trainsets - a very significant reduction in fuel/operator/maintenance costs.
Express service from Brampton to Toronto will already be possible in 2015 thanks to the Georgetown south project. The GTS easily gives us all the track capacity and speed that we would need in the foreseeable future, so the only focus of investment is the two other segments of line (GEXR and CN). The lowest capacity of the two is the GEXR segment, so that's where we need to invest. My impression is that the CN segment can already support hourly GO service despite the level of freight traffic (it is mostly triple-tracked) so all this project needs to improve is Georgetown-Kitchener.
Improving speeds significantly increases line capacity for single-track lines, which could actually reduce costs by reducing the need for double tracking. For example, if we reduced the Kitchener-Guelph travel time to 20 minutes, a single-track line west of Guelph would have enough capacity for hourly service, including a 20 minute layover in Kitchener (during which time other trains can use the single-tracked segment). I think reducing the KW-Guelph travel time by 8 minutes is fairly achievable, given that the Guelph slow zone costs nearly 6 minutes on its own.
Getting rid of the slow spot in Geulph alone will probably cost a couple hundred million. Dual tracking the route and buying more trains to serve it is probably all that is qouted in the $400 million cost, you are probably looking at closer to $1 billion if you do track upgrades to reduce travel time. (which is a worthy investment IMO)
Working from
Dunkalunk's map, the rail upgrade project would only need to eliminate the two crossings on Kent Street (Glasgow and Dublin). Grade separating other roads would be nice for Guelph residents, but is not necessary for faster train service so it could be funded independently and on a different time scale.
The equipment demands are not necessarily very high, because the service increase in question is off-peak, when we have surplus trainsets anyway.