The Pixel’s camera shows a black stripe across both the live preview and the final image:
That’s under the high-intensity LED lamp on my desk, which must have a high-frequency flicker. I’m amazed the camera remains in absolutely stable sync with the flicker for as long as I’m willing to aim it.
The stripe covers only the moth and greenery, not the LCD monitor in the background, so it’s caused by the overhead lamp, not something internal to the Pixel or its camera.
A closer look shows shading on either side of the deepest black (clicky for more dots):
The stripe location and width differ based on the image zoom level, although in no predictable way:
The Pixel camera definitely doesn’t have optical zoom, so it’s surely related to the scaling applied to convert the physical sensor array into the final image. Even though all images have 4048×3036 pixels (or the other way around, at least for these portrait-layout pix), zoomed images get made-up (pronounced “interpolated”) data in their pixels.
Not a problem under any other illumination I’ve encountered so far, so it’s likely something to do with this specific and relatively old LED lamp.
The Smell of Molten Projects in the Morning 
And it’s the reflected light as background reflections appear okay (unless that’s a monitor). That’s neat. I have some led ring lights that cause moving bands in the image when I take a photo through the binocular of the microscope. Not like that, though.
The pale gray strip in the background is the bottom of the portrait monitor, allegedly refreshing at 60 Hz without a trace of banding.
Dunno if there’s room in the lamp base for a bridge rectifier & a honkin’ big filter cap …
I had some early LED floods in the bathroom that would kill the signal to the Icom IC-F14 that I use to get NOAA weather radio. Another one screwed up FM reception in the kitchen. I don’t have any RF test gear, so I don’t know what they were putting out, but it was fairly potent. Doubt they would do well against EMI specs… Newer ones don’t seem to have this issue.
AFAICT, the FCC actually started insisting on better EMI compliance a while back, although the low-budget / no-name crap from Amazon / eBay probably still has empty spots on the PCBs where RF suppression components lived during the emission tests. [mutter]