If we can find a nice way to implement IVideoFrameStep that works with most filters on the presenter side this limitation will be gone and ff/rew are more reliable on higher speeds with MP2.
Currently audio is the problem. Most filters don't allow faster than 4x with an audio filter in the graph, therefore MP1 has to skip ahead, try to play some frames, skip ahead and so on. Ugly, but that's the way ff/rew at higher speeds are currently implemented. We won't change this for MP1 as this means huge rewriting of code, which takes a lot of time that is better spend on writing a new presentation code for MP2.
Currently audio is the problem. Most filters don't allow faster than 4x with an audio filter in the graph, therefore MP1 has to skip ahead, try to play some frames, skip ahead and so on. Ugly, but that's the way ff/rew at higher speeds are currently implemented. We won't change this for MP1 as this means huge rewriting of code, which takes a lot of time that is better spend on writing a new presentation code for MP2.