Upload & Sell: Off
The logistics of bouncing is simpler than two flashes on bracket and stand but will not offer precise control of key light placement (pattern of highlight / shadow) or lighting ratio (fill power controls shadows tone and mood).
In terms of key light placement on faces you want it on the front of the face where the eyes and mouth are highlighting the two most important focal points. One criteria for flattering key lighting is that it reaches both eyes. That's done by keeping it no higher than the angle it can get in past the brow from above, about 45 degrees. As you move the key light off-center it changes the shadows and it also changes where the specular highight clues to where the highest points are fall on the face. Below 45 or more than 45 degrees off center the nose shadow begins to starts to be noticed and unflattering as it hangs down over the lip and and out sideways if the light is moved off the centerline of the nose. The darker it is in contrast to the highlights it overlap the more it gets noticed and distract attention from eye and mouth, making and holding "eye contract" with the faces in the photo rather than seeing the shadow clues and thinking, "Wow what a big nose". There are no rules where to put the key light in relation to the face, but in terms of goals and cause and effect some wind up looking more natural and flattering when rendered with a 2D contrast pattern in a photo.
Flattering strategies get light in both eyes. That's relatively easy to do even when the face is moving if the key light stays centered over the camera. The caveat when putting it to the side, besides fall off over groups, is that any face turned in the opposite direction winds up half shaded. With centered lighting there are very few shadows in the full face view and if the people turn a key light raised about 45 degrees hides the nose shadow from camera view behind the nose. That's the rationale behind the use of a flash bracket or centering key light on a stand: creating natural downward modeling angle and getting light past brow into eyes predictably, and hiding the nose shadow.
It is often possible to get bounced light with all the same qualities but it's more of cat and mouse exercise of taking a shot, seeing how the light models and fills the shadows then trying to adjust point the light is hitting. The advantage of bounce vs. direct centered on bracket or stand is it falls off from the ceiling rather than front>back which which makes the background brighter.
As the foundation for my lighting I carry a 580ex on a camera flip Stroboframe bracket with a DIY scoop reflector with a moveable top flap. That allows predictable flattering (light in eyes no harsh shadows on face) modeling when I can't bounce, and the option to open the flap and bounce most of the light off the ceiling when ceiling conditions are favorable.
On that foundation of flattering single flash lighting I add my second off axis 580ex. If I can't control precise placement on the front of faces I don't put it in front, I wheel it around (it's on a modified IV stand) behind the action as back rim light. Rim light eliminates the fall off "shot in a cave" look the frontal lighting on the bracket and creates a more realistic sense of 3D space in a 2D photo than the "overcast day" look of a light bounced overhead in the center of the room. It's a good strategy for many wedding activities where the couple and crowd around them are looking various directions such as the cake cutting, flower / garter toss, dance floor action, etc. Just ask yourself "Can I get flattering light in both eyes?" If not the more flattering strategy is keep the second flash off the faces.
When I can control the position of the key light to the face, such as when the best man faces the couple and offers a toast, it's possible to predict well in advance where his nose will be pointing and set up the off camera flash so it will be 45 degrees to the side of his nose and about 45 degrees above the eyes, creating the "mask" of highlights that the brain recognizes as a 3D face in a 2D photo. Then as he is speaking I'll walk around his face from full length full-face (snap), waist level crop oblique (snap), and tight H&S profile at the peak of the action (snap) recording a nice storytelling sequence for the album in great lighting effortlessly.
The lesson there, learned from an expert many years ago, is that once a key light is set at 45/45 to the face it's just a matter of changing camera position to get flattering full, oblique, or profile shots. When the camera sees the face full the key is 45 from the lens axis, 90 when shadow side oblique is seen, and 135 from lens axis in profile. Once you are consciously aware of that relationship your eye will automatically find those flattering facial angles. The lighting becomes a "no brainer" because it's set in advance based on where the nose will be pointing at the "decisive moment".
For the "formal" posed shots of couple and small immediate family groups you'll have better control of the lighting with them on the stands and camera and flash in M mode. Whether to use the umbrella or not vs. smaller diffusers on bracketed fill and stand mounted key is a judgement call you'll need to make by testing them both beforehand. The smaller diffusers will be easier logistically so the question becomes whether you see enough difference in the results with the umbrella to make it worth the extra effort logistically.
Having shot weddings with direct flash and using it for 40 years I've never been concerned with the fact a smaller source creates distinct shadows because I learned how to make them seem more "normal" and not be noticed in the photos; just either hide the shadows (centered strategy) or place them naturally via precise control of nose-to-keyl light relationship AND have control of fill and make their tone on the face normal looking or lighter than normal whenever the key light is placed to side and creates a nose shadow. I think of the nose shadow in terms of it being a distraction from the eyes and mouth. The lighter it's shadow is relative to all others on the face the less it get noticed. Centered fill winds up closer to the nose than the ears, so just the fall off of fill create a light>dark modeling gradient from nose, over cheeks to darker ears..
Try this before the wedding:
Light a face as described for the best man toasting with one of your direct 430ex slaves as key 45/45 from a subject with your 580ex on camera as Master / fill. Shoot in ETTL for this test so you can easily change the ratio. Start with A:B = 1:8 (very dark shadows) and highlights just under clipping (solid white highlights (Zone 9) = 250.250.250). Then change one variable, the A:B ratio from to 1:4, 1:2, and 1:1 adjusting FEC as needed to keep the highlights the same. Make a montage of the four shots for comparison. Note how changing the ratio changes the perception of the light from being "hard" to "softer" as key and fill strengths approach being equal incident strength.
Then repeat the same test with your umbrella on the stand and compare results with the small diffusers. Next instead of the umbrella try to bounce the key light off the ceiling with the same fill strategy and ratio sequence. You'll have a better idea which strategy will work best and the results each produces and experience using all of them. It's not good idea to use unpracticed techniques at wedding. At worst it can result in missed important shot and even at best you'll spend more time futzing with the gear which makes subjects impatient.
What causes the difference in appearance between key light modifiers at the same A:B Fill:Key ratio. One factor is the larger source creating more penumbra, the gradient edge of shadows. The other is that the bigger sources create bigger overall footprints and more light bounces around the room as wrap-around "spill fill", the same cause and effect that gives bounce a more omni-directional even "overcast day" look. Key light wrap will change the gradient in the transitions, spilled fill the tone of the core "umbra" shadows.
If you are really curious repeat the test same test direct flash, small diffuser and umbrella at night outdoors where there isn't any bouncing of the footprint off walls and ceiling. You probably will not see as much difference between the modifiers as indoors. The lesson there? Regardless of what modifier you decide to use the space you use it in will be a variable. I seldom use umbrellas with my speedlights because in most situations there's a ceiling creating a lot of spill fill. The exception is when shooting in a large room with high ceilings. For bridal and communion dresses I'll sometimes use a silver umbrella because it creates larger catchlights more specular highlight modeling on the beadwork, lace and satin of the dress. I start with the goal, then pick the modifier strategy according to how it will work in the space.
Finally, unlike most situations where you try to tell the story in one shot a wedding album or slide show is more like a movie. Think in terms of story line in the final album and when changing location remember establish the new scene with a wide shot, take medium shots to show the couple and family in the environment, and close-ups of the action. You don't need to shoot them in that order. Often I capture some unexpected action first, then back-up and put the action into context with the medium and wide views I put before the action in the wide-med-close order in the slideshow. For things like the flower toss change the POV from wide > close-up as the action unfolds and you wind up with a more interesting story that makes the viewer feel they were there watching.
Since flash can't cover a big room in a wide shot I use ambient rather than flash for those, or gell the flashes with 1/2 CTO to match tungsen and bounce them off the ceiling to match direction of the ambient; adding light but not changing the ambience. Matching the flash color with the tungsten allows the ISO to be raised without any mixed flash / ambient color issues. Shooting at higher ISO also extends the effective range of the flash.