This German study is much more convincing. It's hard to interpret, because it would take a lot of work to really check how they've constructed their "synthetic control group". However, they seem to have tried multiple checks against other regions, doing "placebos" to validate their predictions (i.e. if we check the control Vs. regions with no mandatory face masks. They also did some time shifted placebo checks too. These attempts to disprove their hypothesis make me trust the study a lot more.
There's a lot of variation in COVID case rates, so you can't be 100% sure the low rates in one region are due to face masks and not random chance. I'd still like to see similar assessments of other datasets from other regions. But it looks pretty damn likely that mandatory masks helped.
I think they were trying to say that it's hard to match a synthetic control group per city because each city has different trends for their Covid-19 outbreaks, and each city has different policies centered around mask-wearing (In Jena's case, mask-wearing was introduced as an intervention, quite early on during the outbreak. Other jurisdictions may not fare as well, or maybe adherence to wearing masks is significantly weakened). So this helps for creating a weighted average, but makes it more difficult to trace mask-wearing protocols with other cities and the correlate it to reduction in cases, as you mentioned. Even so, the researchers still found that masks had a robust effect on other cities. As you also mentioned, the researchers went to great lengths to minimize as many biases as possible with the various implementation of control groups. This is why I wanted to highlight this study since there seems to be an anti-mask sentiment that is not rooted at all in rationality, and we have essentially the best evidence at the moment to say that there is reason to believe that masks contribute positively.
If it weren't tedious to tailor synthetic control groups per city, it would be interesting and more convincing evidence (beyond simply showing a model of transmission as seen in other studies) to show that masks do have a clear effect in reducing transmission of Covid-19. I hope we get to that point, but it seems problematic now since we're in a different landscape now in mid-June vs. in March/April.
Having said that, I think another question that would help significantly for a study that's doing multi-region tracking, would be to have an understanding of why there's some variance in Covid-19 responses (particularly, why do people react so harshly and have their immune system go to overload to the point of organ failure, while others may experience just mild symptoms).