Some m0n0walls traffic shape better than others, it depends on how good your ISP is at providing the "advertised" speed that you have purchased.
The critical part about traffic shaping for VOIP is that you never, never, never (truly) never actually saturate your connection to your ISP in either direction. That means knowing for a fact what your worst case, actual available bandwidth will be. Your ISPs claims aren't worth mentioning. You need to measure it yourself.
In my case I measured using several speed test sites, at different times of the day, then took the worst case results as the basis of my traffic shaper setup. It seems wasteful at first, like there's bandwidth available that you will never use. However, it really does work. I have a complex VOIP setup, a mix of a local IP-PBX and hosted services, and it all works very reliably.
FWIW, consider also using a reduced bandwidth codec. I like G.729a a lot. I use it both on my Asterisk box and on my Polycom & snom phones with OnSIP.
Yes, technically the call quality is a little less than G.711. However, the complaints people commonly make about VOIP call quality are more often than not a reflection of poor bandwidth management than poor codec selection.
By using a low bit rate codec you de-emphasize the requirement for precision traffic management. It gives your some room for error. This is especially true on asymetrical links like DSL or cable modems.
I'm hoping to do a screencast recording this coming weekend to document the process from end-to-end. I'm setting up another m0n0wall on a Comcast cable modem. Thus I have to go through the entire process again, so it's a good opportunity to give a detailed account of how it's done.
Michael