What about HTML5 video tests between the two browsers? As youtube being the one of the largest (if not _the_) video providers, it'd be interesting to see that comparison as a benchmark.

Just made a short (about 3 minutes each) youtube test on a MB Air 2014 (HTML5, full HD, fullscreen).

Result: There was a 3 Watts (roughly) difference favoring safari.

Awesome! Thank you for that!

That benchmark is more interesting to me.

On my MacBook Pro (2.66GHz i7, early 2010), just starting Chrome causes the graphics to switch from the integrated Intel HD Graphics to the NVIDIA GeForce GT 330M, which incurs a substantial battery life penalty.

(I discovered this quite by accident when investigating why my display would flicker when starting/quitting Chrome if I was also using f.lux.)

How did you figure this out? What tools did you use?

Just goes to show how unoptimized video playback on Adobe Flash Player apparently is after all those years.

The brightness appears to be higher on the right. dr.dk uses a Flash Player with variable bitrates on Chrome instead of HTML5. Other things like a low battery can skew the results.

The screen definitely looks brighter, but the camera is also pointed further to the right (notice what keys are within the frame). IPS definitely helps with viewing angles but as you can see by [0] you still lose brightness as you move to the side.

Since it is a single tweet, we cannot really take it as seriously as we could if they explained their methodology and showed that it was repeatable.

[0] "Conventional IPS" slide: http://www.j-display.com/english/technology/jdilcd/pictureq....

Isn't this really just a Flash vs. H.264 power drain test?

And don't most Mac portables ship with h.264 hardware acceleration?