Yeah, its better to just buy blunt props that are 1 - 2 mm smaller than the diameter of the cowling. This prevents prop striking the cowling but still allows for minimal tip vortices. Beyond that, stick with even numbered blades. This way the props point of greatest lift and lowest occure at the same time and lesson vibration.

I'm not sure what you guys mean. I tried some dal3045 bullnose props but there wasn't any 3" props i had that fit the ducts without cutting them down. I think no matter what you have to cut them down. I used a dremmel and the 3d printed jig that cuts them all perfectly. I plan on trying some of the h1 4x4.3x3 that is recommended by nurk and andy shen. But i just felt like a 4 blade prop would be more balanced because the weight is more evenly distributed throughout the prop, and between the 3 types of props i tried the 4 blades worked the best imo

Yeah, i get that finding the perfect 3" is hard.

Now as for the rest of what i ment it goes like this. The props blades leading edge when facing into the direction of flight produces more lift. Conversly when it is moving in the direction opposite it produces less. For most Quads, this means that the front props produce more thus when the blades leading edge is on the outside (furtherset away from the copter). While the rear props produce more as the blade is closer to the frame. For props having odd numbers of blades, this can produce all sorts of thrust variance as the props spin. This variance can cause the copter to vibrate as can be seen in the Gyro trace if you look closely enough. Now with props having even numbers of blades, both the highest and lowest points of lift occur simultaniuosly, balencing each other out. This produces less thrust variance and vibrations as the props spin. However all of this is moot once the propeller is shrouded so long as the props are kept farly close to the shroud wall. (< 1-2 mm)

Only issue I see is that most Ducts / Shrouds are set up for low RPM propellers and lose any benifit of extra lift the very moment the Copter is no longer level or throttles up. This is because when not level, the Lip of the Duct / Shroud receives the incomming air unevenly. Thus producing extra lift only on it's leading edge. With this as the speed increase, so does both the drag and prop cavitation. In flight you can often hear a Shrould oscillate depending what it's made from and how thick. Instead, what should be used is a "decelerating duct". (think jet-ski water pump) While this doesn't produce extra lift during hovering, it's less prone to drag / cavitation at higher notor speeds that are typlical for multirotors.