Mk.82

"The reason this may not show up (as much) in stills is because you don't need to have the sensor active as long for stills as you do for video." Problem with that assumption is that mirror less cameras does full sensor readout all the time when powered up. Everything is recorded and processed down sampling to keep details etc. When you hit record, the CCPU starts to process high bandwidth requirement data compression and battery will be consumed more, both heating up. The sensor doesn't heat up. You can confirm this by using a FLIR to see how E-M5 sensor stay cold at long recording times (60min) but the CPU and battery warms up. That heat is directed to aluminum body at back. There was few years ago videos of tests E-M5 being in Sauna at 75° temperature IBIS enabled and recording 3x20min videos continually. The video didn't look worse after thethe first 2x20 session but was same as the first seconds when the camera was brought to in sauna. After last 20min the noise started to be more visible but difference was like shooting ISO 3200 vs 800. Still very usable. The sensor heating isn't caused by itself capturing and processing ADC. All the talks about sensor heating up because IBIS are nicely said ignorance. The mirrorbox needs to be shielded from CPU and battery, otherwise that heat gets quickly transmitted to mirrorbox (in this case sensor box).