Try to get the most length out of the room that you can -- so keep the 22 foot rather than the 20 foot dimension. That allows for a lower standing wave problem area (17 to 308 Hz for 22 ft vs. 19 to 340 Hz for 20 ft). May not be the end of the world either way -- let us know if it is a lot harder to change the width or the length into storage space. Allowing for an 8.5 foot ceiling and a 22 foot length, here are the effects of changing the width:
1) Using the golden mean (height x 1.618 = width and width x 1.618 = length -- by the way, that means that height plus width = length), your width should be about 13.75 feet. At that width, the standing waves don't reinforce each other significantly except at 205 Hz (which is really quite good since you don't notice it as much there). There are a few relatively weak spots at 100-120 Hz, at 155 Hz and at 230 Hz where there is at least a 20db gap to the next standing wave.
2) Using the 10x16x25 ratio gets you to a 8.5x13.6x21.25 dimension, which gives an artificial boost (where standing waves reinforce each other) at 135 Hz and 265 Hz. The relative weak spots are at 105, 160, 185 and 240 Hz. Probably not quite as good as option 1. But the ratios are very similar in both cases and the work would be greater since you also need to shorten the room.
3) Using the modified Bufus suggestion yields a better answer. The low end artificial boost is lower overall and only hits the 3db hurdle once at 285 Hz, which will probably not be noticeable. The relative weak spots are somewhat improved as well: 100 Hz, 125 Hz, 175 Hz, 200 Hz and 240 Hz. On average, this performs better overall than does the golden mean approach. It may also allow for better furniture placement and more room to place speakers for optimal reverb.
Note that there are real dips and valleys in between the 13.75 and 16 foot widths -- please don't assume that anything in between is just as good. 13.833, 15.5, 15.833, 15.917 are also fairly similar to the characteristics described above (but 16.083 is much worse).
By the way, Craigs suggestion is also close to the golden mean ratio, but the 14x22 with your ceiling height leads to results not as optimal as the ones described (though with a 7.5 foot ceiling, the only low end boost is at about 80 Hz).
All of these calculations consider only the standing waves calculated from the width, height and length -- none of the tangential waves are considered.
Have fun -- there are lots of great design options. While I've studied this quite a bit recently, I'm sure there are others who are significantly more qualified to address this and I look forward to their comments. email me if you would like the spreadsheets for these calculations.