EC2 VT1 Instances for Live Multi-Stream Video Transcoding
aws.amazon.com
aws.amazon.com
i'd love a quality comparison. GPUs often have hardware encoders, but my understanding is that it's super super common for video-game streamers to have a second box, doing CPU encoding, which produces much better output for the same bitrate.
to come back around to realtime vs batch, xilinx alveo u30 vs google vcu, i've always wanted to know how much if any extra latency CPU encoding adds versus GPU encoding. does moving from on-system CPU encoding to having a separate CPU encoder box add signficiantly more?