I'm really tempted to not bring this up as I know it's going to sound like I'm trying to downplay what you've done. It's only the possibility that some handloader could assume that your experiment resolves the issue for all loads and all powders and have a serious pressure excursion that leads me to add to the discussion. You've done a fine job of providing data that the powders tested, in the loads tested, did not suffer visible damage in the certainly extreme length of tumbling you tested at. There are some factors though that you at least didn't mention evaluating.
Was the powder at all compressed in the loads tested as earlier replies have asked? If so, the environment the powder grains experienced would be different than in the case of a looser load. And one would expect the looser load's powder to suffer more damage, leading to more effect from the two factors below.
Did the tumbling generate any very fine powder (usually called "fines") in the cases? This would have to come from the surface of the grains, and your photos do imply that if any "fines" were generated, the total would probably not have been very large. Smokeless powder grains are pretty robust physically. However, it doesn't take very much in the way of very fine smokeless or black powder to significantly steepen the pressure vs time rise after ignition. Similarly, a change in the statistical distribution of grain size toward the smaller grains will increase the burn rate and burn rate exponent. After all, one of the major parameters the powder manufacturers use to adjust burn rate is the physical size and configuration of the powder grains. Steepening the pressure-time curve does in fact lead to higher peak pressures. Sometimes very much higher.
Many smokeless powders (not to my knowledge black powders, but I'm not certain of this) have a burn rate modifier impregnated into the surface of the grain after the grain is formed. In the formulations I'm most familiar with, that modifier is often a burn rate retarder which among other things reduces the rate at which the burn rate increases as the pressure rises after ignition. If that modifier is worn away to any appreciable extent, the burn rate modifier will be compromised, and the burn rate and burn rate exponent will not be what the manufacturer intended. They will also not likely be consistent or repeatable.
What you accomplished is interesting and your photos of the before and after grains are damn good for any level of equipment, there are though some key factors that from your post you may not have addressed.
Pat
ETA, while I was typing, I see you added the 65% fill info, please disregard that first question.