CRF of VP8 vs. AVC/h264

[smitbret]

I am running some test conversions of vp8 encoded media and wow it seems like vp8 taxes my CPU (Phenom II x4 955) a lot more than h264 at a similar CRF setting. The Phenom is just the workstation that is running the tests.

I am running through some beta testing of some software that transcodes video on the fly for streaming and their defaut for vp8 is 4. They suggested running it up to 10 to see if that speeds things up. Those settings for an h264 encode would be ridiculous for streaming media. Is the scale different for vp8?

I am running some test encodes through Handbrake right now. I am converting the first 10 minutes of a 1080p digital copy of The Incredibles and down-rezzing to 720p at with a CRF of 22. I am getting around 18-20 fps. Normal/Abnormal?

[jagabo]

Is the scale different for vp8?

Of course. Judge by bitrate (file size) of the resulting files.

I am running some test encodes through Handbrake right now. I am converting the first 10 minutes of a 1080p digital copy of The Incredibles and down-rezzing to 720p at with a CRF of 22. I am getting around 18-20 fps. Normal/Abnormal?

Nobody can tell you for sure because there is a 100 fold difference in speed between x264's fastest (ultrafast) and slowest (placebo) presets and you haven't said what x264 settings you're using. But that sounds about right for the middle presets.