My family generally doesn't appreciate differences in quality, thus I favor an approach that is "very good" in both audio and video quality, but saves on disk space. If I want the full HD audio and unmolested video experience, I'll view the source disc instead. I use Handbrake for everything (generally nightlies) unless it's 1080i VC-1 source, which requires something else like RipBot264 instead.
My settings for each type of source (good balance between encode performance and reasonable quality):
- Common to all: MKV using H.264, keep original source size, but with loose anamorphic, and auto-cropping. Pass-thru AC3, or transcode DTS to 640kbps AC3, no down-mix from 5.1. Keep all English subtitles (especially forced) w/o burning in. for PGS subtitles, post-encode, I use MKVMerge since WDTV doesn't handle Handbrake's PGS approach. Vobsub subtitles pass through just fine. Preserve all chapters. Advanced encoding options at defaults except "b-adapt=2:rc-lookahead=50"
- 1080p Blu-Ray, CQ at RF22, no filters. Results in a Profile 4.0 file at roughly 2.2GB/hour or about 5-9Mbps
- 1080i Blu-Ray, CQ at RF22, Detecline and Decomb at "Default". Results in a Profile 4.0 file at roughly 2.2GB/hour or about 5-9Mbps
- 480i DVD, CQ at RF19, Detecline and Decomb at "Default". Results in a Profile 3.1 file at roughly 800MB/hour or roughly 2Mbps
Avoiding HD Audio codecs entirely, and transcoding 1.5Mbps DTS down to 640kbps AC3 (preserving 5.1 channel) makes for a noticeable reduction in file size w/o any appreciable loss in fidelity.
For Blu-Ray, higher qualities (i.e. RF19 on Blu-Ray) look even sharper and truer to the source, however the encode times are just too long on my aging Phenom IIx4 955 @ 3.2ghz machine for "marginal" improvement, and file size reduction isn't significant enough. RF22 is as good or better than a "cable" HD experience, while encoding at roughly 10fps and yielding "acceptable" file sizes. If your available storage space is significant, then RF19 could work.
For DVD, with fewer pixels overall, anything higher than RF19 is noticeably poorer, and could only be done satisfactorily by adjusting some advanced settings. The encode performance increases substantially if you do, thus it's just easier to go w/ RF19. These net a performance of ~45fps encodes on average on my hardware.
File size estimates are completely dependent upon the type of material - with CG material for kids taking up less space than film-based 1080p content.
In the end, it's up to you to experiment with the results of each settings. For some, one setting may not fit all, but when you're aiming to transcode a library of 100s of discs for the sake of convenience for non-technical family to enjoy, you may opt to compromise on ultimate quality to keep your workflow sane.