Between the two scenarios, performance will typically be the same, but Scenario 1 might sometimes be worse.
In scenario 1, all frequencies entered in the single "site" are potential control channels. The scanner will initially scan through those frequencies, starting at the top of the list. When it finds a signal that it can decode as a control channel, that's the one it will use, until that control channel fades. Then it will start scanning the list again until it finds another control channel. As noted, unless you memorize which frequencies are associated to particular sites, and have the setting enabled that shows the control channel frequency on the display, you'll have no idea which site you're scanning.
In scenario 2, the scanner will initially scan frequencies in each site, starting at the top of the list of sites. It will scan until the scanner finds a control channel whose decode rate exceeds the "High" threshold set for the site. It will continue to scan this one site/control channel until the decode rate on that control channel fades below the "Low" threshold. Then it will start scanning other sites for active control channels that exceed the high threshold. Using this method, it won't necessarily find the control channel with the "best" decode rate, but rather the first control channel with a decode rate over the "High" threshold. Assuming you have the setting enabled that displays the site name, you'll know which site you're scanning. If you're in a scenario where there is a lot of overlap among sites (ie, using a good antenna system, or at a high elevation), the High and Low threshold settings for each site become pretty important to how the scanner handles sites.
Additionally, if different sites typically carry different sets of talkgroups, you may want more control over which sites you monitor. Unfortunately, the menu system on these scanners is such that locking out and unlocking systems and sites is a very tedious process (why does the Edit System menu take so long to load?)
Consider a third scenario, where you create different "systems" (TSYS) objects. Each TSYS would have either a single site, or a few sites grouped together. You can then associate talkgroups from each different TSYS with different scan lists. This would give you the best control for scanning a specific site, or a small group of sites. I generally use this method with my Whistler SD Card scanners when monitoring multi-site systems.
For my statewide trunked system, I typically group sites into distinct TSYS objects for each county, or if they're smaller, a group of counties. I might have from 1 to 5 sites in each TSYS. I group the talkgroups for each TSYS into separate scan lists based on agency type (Fire, EMS, Law Enforcement, Local Govt, Utilities, State police, etc), and then group these scan lists into different Scan Sets. So a Scan Set would contain all the scan lists that contain talkgroups from a single TSYS (site or group of sites). This generally works well. Using this method, you can pretty quickly run out of scan lists and scan sets, so I use V-Scanner folders to hold configurations for different regions in the state. It's time consuming to create intially, and it's even more time consuming to update each TSYS object in each V-scanner (There's a LOT of duplication of talkgroups), but it gives the maximum level of control at region/county/agency levels.
