Brian Puhl:
Some folks here are mentioning that filter changes add time to your sequence due to settling, I'd argue you need to adjust your settings for better effectiveness. My swaps are around 5 seconds, maybe sometimes 10 seconds at most. Also, you don't need to constantly change filters every sub, I usually run about 10 per filter then switch.
The issue is quite dependent on the focuser. I have a nitecrawler on two OTA's. On those, it is possible to switch filters (and apply filter offset) without the guide star moving. For those, I find the most efficient thing is shoot something like R, R, G, G, B, B, <dither> and repeat. The dither requires a settle, but on these OTA's the filter change does not.
On another with an optec focuser, switching filters moves the guide star. It is necessary to stop/star guiding, otherwise the beginning of the next exposure is messed up as PHD2 chases the moved guide star back. On stop/start the restart time and settle is typically longer than the dither settle time. I find the best for it is R, R, dither, R, R, dither, etc. through a lot, then switch to B, B, dither, etc. This means the stop/start is only on filter changes.
There are two reasons you need to stop/start with less precise focusers, one is there is no way to just say "settle", but the other is that if the focuser moves a lot (e.g. the direction involves a large backlash compensation), PHD2 might actually lose the guide star. If it tries to re-acquire it can be slow, and might get the wrong star. When I tried doing this without stopping, I was finding every first exposure after filter change had a higher FWHM, which came from a short period of chasing the guide start to get it back.