3 ms·
The interesting question for me is whether existing mobile hardware can accelerate this in any way - i.e. through the ARM Neon extensions and similar. Otherwise
by andybak 17y ago
The interesting question for me is whether existing mobile hardware can accelerate this in any way - i.e. through the ARM Neon extensions and similar. Otherwise we will have to wait for a whole new generation of hardware if this is going to make any dent in the mobile web.
- mbrubeck 17y agoBoth Mozilla and Google funded work to implement hardware-accelerated Theora on mobile devices[1][2]. While I don't know of any similar efforts for VP8 yet, I'd guess that it is inevitable. [1]: http://google-opensource.blogspot.com/2010/04/interesting-times-for-video-on-web.html http://google-opensource.blogspot.com/2010/04/interesting-ti... [2]: http://blog.mjg.im/2010/04/16/theora-on-n900.html http://blog.mjg.im/2010/04/16/theora-on-n900.html
- zokier 17y agoOn2 was already developing hardware solutions (iirc they already had hw encoder/decoder designs) for their codecs before G acquired them.
- ZeroGravitas 17y agoOn2 announced some stuff with ARM a couple of years ago, apparently they worked together to both design and optimize VP8 for the ARM platform, including NEON. And one of the low cost platforms for Android phones had a mysterious "On2" hardware acceleration block on their schematics at least 6 months ago. A quote from ARM today on the VP8 announcement: The efficient processing is helped by the fact that the VP8 codec is already optimized for Cortex-A/v7A class processors with the NEON SIMD (Single Instruction Multiple Data) engine. NEON allows the ARM CPU to work on multiple bits of data in parallel, and optimizations for ARM11 processors are also available http://blogs.arm.com/software-enablement/support-for-vp8-and-webm-on-arm/ http://blogs.arm.com/software-enablement/support-for-vp8-and...