As I was struggling with all of this, I was perplexed about a couple of issues. Still am.
1) I don't see how a revolver's use of recoil (barrel rise, etc.) can be the same as semi-auto. Both weapon types MUST be affected and have barrel rise, but there OUGHT to be a noticeable difference. I noticed in a Starline Brass video that those folks, seemingly expert, seem to same the two weapons are affected in the same way -- but they don't directly say they are affected.
I also don't understand how barrel length affects recoil (as it, in turn, affects point of impact. (I do know that powder can be changed to ignite more quickly -- typical .22 WMR rounds designed for long guns and fired in a revolver or semi-auto perform differently than then using .22 WMR rounds designed for shorter barrels: less of the powder ignites OUTSIDE the barrel (i.e., less is wasted) when its matched to the barrel length. Barrel length and a given load could, and in theory, that would take longer and force the bullet to travel a greater as distance as recoil raises the barrel higher. But, barrel length could also allow more powder to be burned, which could, in turn, incease the bullet's velocity and speed the bullet down the barrel, shortening the time needed for it to exit. Same load but different performance. Too damned many variables.
2) Others tell us that most semi-autos, with an action design that delays recoil somewhat, there is can be a VERY SMALL amount of vertical movement during the early part of recoil. Most of the recoil-induced barrel movement isn't seen until the heavy barrel and slide slam back against the frame (stop area) and causes the assembly to tilt. With a semi-auto there is an extra step in the recoil process, evidenced in the entire barrel and slide moving to the rear before the bullet exits the barrel. But it happens quickly -- with the barrel moving to the rear by about 1/10th of an inch before the bullet is gone. That amount of movement, pressing against a spring that will apply only trivial amounts of force against the frame, and will generally apply it BELOW the bore axis (but above the hand holding the gun) -- its still there, but argulably with less effect.
3) The later effects of recoil can be somewhat less obvious in a semi-auto than when using similar loads in a revolver. -- both because there is higher bore axis in the revolver and nothing DELAYING or redirecting the force of bullet's movement as in semi-autos. In either type of weapon, some work is being done BEFORE the bullet leaves the barrel, so there should be evidence of that work being done -- if not VISIBLE muzzle rise, at least with a visible change in points of impact. Bore axis complicates the comparison. Then, too, some revolver designs have a very low bore axis (ala the Mateba design, with the barrel aligned with the bottom of the cylinder.) I wonder how muzzle rise affects point of impact with a low-bore axis revolver compared to one with a high bore axis.
If we could perform the same test in a Ransom Rest, one series of shots done normally, and another series done with the slide locked, we might see some interesting differences. I'd like to see Ransom Rest tests of semi-autos with similar barrel lenghts firing similar loads, too. Nobody, thus far has done that to addressed any of these "difference" question, showing us evidence with aimed targets shot from a rest. I'd really like to see that.
4) Revolvers are a little different, if only because of bore axis. Some also say that revolver recoil has two phases -- one in the cylinder, and a second when the round is in the barrel. I don't know that recoil in revolvers is a two phase process -- as, to me, it just seem like a slight reduction of the forces put into effect the primer is ignited and the bullet passes the cylinder gap. All we're really concerned about is the bullet's velocity as it leaves the barrel and the angle of the barrel if it's affected by recoil. The gases lost in the barrel/cylinder gap might be the equivalent of work done like gases transferred via a gas piston in a semi-auto that uses gas pressure to cycle the action. The gap is also (in terms of force redirected) a bit like a compensator, but it's doing its job a bit earlier in the bullet's travel.
As for the sight alignment in revolver and their higher front sights -- sometimes offered as evidence of a revolvers innate tendency to shoot higher than a semi-auto: it
is different, but some part of that height is necessary because the rear of most revolvers becomes increasingly BEEFY as calibers go up --
as extra material is needed in the cylinder and the frame to keep things together. That extra material raises the rear sight, and the front sight has to be raised, too.
5) What seems to be a confusing factor in this discussion is that a number of bullets are aerodynamically different, and because we're not shooting our weapons in a vacuum that difference may be greater than we might expect. Limnophile gave us an example of 9x18 rounds that seemed identical, but performed differently -- and bullet structure seemed to be the difference.
I found PMC .357 magnum loads (C and F in the chart below) where the heavier round (158 gr) and the lighter round (125 gr) had the
same muzzle velocity (1194), but the 125 gr. round hit at .62" above level at 25 yds, while the 158 gr round hit at .61" (i.e., lower) at the same distance. The heavier bullet dropped less than the lighter one at all measured distances which is NOT the normal performance of heavier bullets with similar or lower velocity.. The bullet designs are quite different, the aerodynamics of the rounds might be the cause. Here's a base of PMC data that gives performance at distances.
It's hard to find comparison data for round performance. In the chart above, things are GENERALLY as theory says they should be -- except that round F, mentioned earlier doesn't behave in the expected manner when compared to round C. (The unexpected performance may be explained by inertia, and the heavier bullet's better aerodynamics), with the resistance of the air having less effect on the heavier and more aerodynamic round.) It's obvious I'm not as familiar with the PHYSICS involved in this discussion as some.
I have yet to find a high-speed video of a revolver that clearly shows the barrel position as the bullet leaves the barrel, and soon after. Nor have I found a video showing how a semi-auto's recoil and assumed muzzle rise, when the delay is removed) is different: there ought to be a different point of impact. And bore axis shouldn't be a big factor as the bullet leaves the barrel.
Ransom Rest tests addressing any of these unanswered questions or issues would be helpful; that would require that someone align the sights with each shot -- typically NOT the way Ransom Rest tests are done. For at least one of the tests, you'd probably have to be able to lock the semi-auto's slide, and that's not easy with all semi-autos. (Revolvers with 4" and 5" barrels can be found that shoot 9mm and 40mm and 10mm and .45 ACP rounds. I'd like to see the performance of those rounds differ, if at all, with all using the same loads (and barrel lengths).