for a week, after I had set up my CO2, but before I had built my diffuser, I ran my CO2 directly into my canister intake. The CO2 would puddle up and get trapped in the top of my canister, then when the airpocket inside my canister got big enough, some would get chopped up by the impeller, spewing a cloud of tiny bubbles into my tank. The noise wasn't an issue, it sounded like static or white noise for a second every few minutes or so.
I had concerns with this though. First, to me, it seemed as if there was always a CO2 bubble in my canister, and with much probablility that CO2 bubble was in constant contact with a part of the gasket surrounding the motorhead. This can't be a good thing for the gasket over an extended period of time. If the CO2 bubble wasn't in constant contact with the gasket, it was, in the very least, in constant contact with the plastic housing of the motorhead. I imagine, over time, this is not good for the filter, as CO2 will degrade the plastic the canister is made of.
Also, I don't think I got as much diffusion as I figured I would get. The bubbles that would burp out of the canister every few minutes weren't all micro bubbles like you would see with a skimmer, they varied in size, from micro to pea size. Also, since my outflow from the canister is near the surface the larger bubbles quickly broke on the surface, causing me to lose that CO2.
I built an inline reactor. It was very easy, very cheap, and now I get 100% (or very close to it) diffusion.