Finally finished my randomized flicker test - thoughts

I ran something similar on my WS2812B strips last winter for a haunted porch display, where the strobe patterns were firing dozens of times per second for hours at a stretch. My experience lined up with yours. I checked the power injection points with a cheap IR thermometer every half hour or so and never saw anything beyond the usual warmth you get from a long run at full white.

The one thing I did notice was flicker-related burnout on a couple of individual pixels near the start of the strip, but I'm fairly sure that was a solder joint issue from my own sloppy work, not the rapid switching itself. Swapped those pixels and the replacements have held up through two seasons of abuse.

What I did change was my brightness cap. Running strobes at full white every cycle pulled serious current, and my 5V supply would occasionally brown out mid-sequence. Dropping the max brightness in WLED to around 70 percent smoothed that out completely, and honestly the effect looked better anyway since the flashes had more contrast against the dimmer baseline.

So from my corner, the drivers are fine with rapid cycling, but watch your total current draw when lots of pixels flash white at once. That was the real bottleneck in my setup, not the switching.

My run didn't go quite as clean as yours. I had a strobe-heavy sequence running on a strip behind the living room window through most of last December, a couple hours a night, and around week three I started getting random dropouts on one section. When I finally pulled the corner back, the joint at my first power injection point had gone crusty and half lifted off the pad. My best guess is the constant heating and cooling from all those on-off cycles worked the connection loose faster than a steady burn would have. The strip itself and the cheap 5V supply I was using never acted up, so I'd agree the LEDs and drivers take flicker in stride, but the connections are the weak link in my experiance. I redid that joint with a proper soldered connector and a dab of hot glue over it and haven't had a single dropout since. One other thing I noticed, my power supply has a little fan and it got noticeably twitchy with the constant load swings, almost like it couldn't settle. Your hour of testing is a solid start, but I'd let it run a few weeks of real use before calling it fully fine, since my problem only showed up after the pattern had cycled a few hundred times.

I tried something similar last month using a really fast palette swap in WLED, instead of just rapid on/off. I found that when the sequence changes color or brightness so quickly, the microchip itself seemed to struggle to keep up with the data stream. It felt like the controller was getting overloaded by the sheer volume of state changes, not the current draw. Even with good power, those ultra-short, complex transitions felt more demanding on the processor side of things. I ended up slowing the transition time down a bit to about 20 milliseconds to keep everything running smooth.