Is it safe to say that "faster" encode is hardware (Apple T2, Intel QuickSync, or AMD VCN) encoding and "slower" encode is CPU encoding? It's a huge difference in speed on both my MacBook Pro 2019 and iMac 2019.
MacBook Pro with Touch Bar
No, that's not it. Faster is a quick one-pass encode. Slower is a two-pass encode. Slower will give better results as the second-pass of analysis allows for better quality transcodes, and usually slightly larger files sizes. Has nothing to do with that hardware/software performs the transcode.
GPU encoding is hardware accelerated. Besides that point, quoting Apple: “T2 also features HEVC video transcoding that’s up to an incredible 30 times faster, enabling pro users to work more quickly with higher resolution video throughout their workflow.” From the announcement of the Mac mini 2018. I think “faster” is hardware (whether GPU or T2) accelerated.
Thanks for that link. I was pretty certain watching the very low CPU utilization during faster encodes compared to the full beans CPU utilization using slower. Also, faster is more than 10x faster than slower which lead me to believe that faster was indeed using hardware acceleration.
It seems that with T2 or M1 chip I must unlearn that slower and multi-pass = better quality (thanks for the link Luis!):
https://larryjordan.com/articles/apple-adds-hardware-to-improve-video-compression-speeds/
If what Larry says is true, 1st I've heard of that, I want to see it documented.
H.264 is the only codec supported for hardware encoding/decoding.
Other encoding formats mostly use GPU, which would not show up in your CPU measurements.
"GPU encoding is hardware accelerated."
No, it's just faster software encoding than on a CPU.
Faster encode vs slower encode for HEVC and H.264
