@Volts - Results including the vanilla ASi M1:
21 Bridges (2019) Handbrake M1 Pro vs M1 Max
Original - Plex DVR Broadcast - MPEG-TS 2.93 GB
M1
H.264
[11:21:59] hb_scan: path=/Users/rob/Desktop/21 Bridges (2019)/21 Bridges (2019).ts, title_index=1
[11:38:11] work: average encoding speed for job is 166.893768 fps
2.01 GB
19min08sec
H.265
[11:00:50] hb_scan: path=/Users/rob/Desktop/21 Bridges (2019)/21 Bridges (2019).ts, title_index=1
[11:14:51] work: average encoding speed for job is 192.731827 fps
1.81 GB
14min01sec
M1Pro
H.264
[19:13:42] hb_scan: path=/Users/rob/Movies/Movies/21 Bridges/21 Bridges (2019).ts, title_index=1
[19:30:49] work: average encoding speed for job is 186.444977 fps
2.00 GB
17min07sec
H.265
[20:27:37] hb_scan: path=/Users/rob/Movies/Movies/21 Bridges/21 Bridges (2019).ts, title_index=1
[20:41:02] work: average encoding speed for job is 201.152725 fps
1.81 GB
13min25sec
M1Max
H.264
[19:33:03] hb_scan: path=/Users/rob/Movies/Movies/21 Bridges (2019)/21 Bridges (2019).ts, title_index=1
[19:45:51] work: average encoding speed for job is 294.679535 fps
2.11 GB
12min48sec
H.265
[20:31:31] hb_scan: path=/Users/rob/Movies/Movies/21 Bridges (2019)/21 Bridges (2019).ts, title_index=1
[20:40:08] work: average encoding speed for job is 313.789276 fps
1.65 GB
8min37sec
[Opinion mode - On; easily dismissed!]
I’m not entirely convinced as to the merit of using vmaf on digital broadcast 1080 material. The original broadcast compression may or may not be good (the example used is from BBC HD where they produce a really good quality for the bitrate achieved - others are not so good) but the qualitative and subjective results are dominated in most cases by the eventual upscaling from 1080p to 2160p for the average large-format 4K display.
I find the quality of upscaling broadcast material varies considerably. A number of years ago I would not have imagined that the upscaling quality of a display would eventually rival and in some cases overtake that of a quality AV amp but LG OLEDs from 2019 onwards did just that. Almost all others (still in my view btw) have failed to come close to LG though until Sony managed it recently whilst using the same LG panel (go figure).
I would not be surprised if a dedicated 1080p display in an A to B comparison would reveal the differences as detected by vmaf (and there are many on the forum who can do that far better than I) but with the dominance of 4K in the market it matters little to most.
My settings produce material indistinguishable from the TS original. Others could probably refine them further to produce a faster and/or smaller file size and I enjoy seeing what others use.
[Opinion mode - Off]
You mentioned the difference between ASi and QuickSync and the comments on quality. This certainly appeared to be the case in the early days of M1 vs the latest QSV but things have changed.
M1Pro and M1Max were excellent from the outset and the M1 itself was improved considerably with later macOS revisions and all 3 SoCs rival QSV now. The only oddity is the M1Ultra which has the quality but not quite the expected performance. Given the market this SoC was targeted at this is somewhat surprising.
I’ll have another look at LATM audio ‘management’ so thanks for the tip as learning is never over given tech developments!
