3rd dither test (With Loudness penalty) of Acustica Audio's "LACE" & Reaper's own dithering at 16bit

Third dither test (With UA-cam Loudness penalty) of Acustica Audio's "LACE" and Reaper's internal dithering at 16bit.
Be aware there is a sine-wave pinging sound in this video that may appear loud.
Test-procedure steps:
01: Created a 96kHz/32bit sine-wave 'ping' (Using Izotope RX9) that fades from 0 dB FS to absolute silence (Followed by a short insertion of absolute silence to make sure there is a period with zero normalized signal amplitude)
This file has zero harmonics by itself.
02: Loaded the ping-clip into Reaper and lowered the channel-gain of the ping to -80 dB FS using the channel-fader in Reaper.
03: Feeding the -80 dB ping into "LACE" while going through some of LACE's dither-settings at 16bit and rendering this from Reaper with the master-fader at nominal gain (With Reaper's dithering turned OFF)
03a: Rendering the -80 dB ping from Reaper while going through both of Reaper's internal dither-settings at 16bit with the master-fader at nominal gain (Also rendered one with all dithering turned OFF)
04: Loaded all the rendered clips into individual channels on Reaper and used an Airwindows bit-shift-gain set to +11 bits (Should raise the volume without creating artifacts) and once again played each clip by itself (The clips placed first on the time-line in Reaper)
05: Loaded all 6 dithered and non-dithered versions (Of the -80 dB ping-clips rendered from Reaper into 96kHz/16bit files) into Izotope RX9 and created a -6 dB and -12 dB gain-reduced version at 96kHz/16bit (Simulating UA-cam reducing the volume of an uploaded file to meet their loudness-target)
06: Loaded all the rendered clips into individual channels on Reaper and used an Airwindows bit-shift-gain set to +11 bits (Should raise the volume without creating artifacts) and once again played each clip by itself (The clips placed second, on the time-line in Reaper, are the -6 dB versions and those placed third are the -12 dB versions)
Conclusion:
In terms of subjectively perceived distortion; the noise-shaped versions survived the gain-reduction better than those only using TPDF without noise-shaping (And the non-dithered versions were the worst in all cases)
In terms of loudness; see the measured values below.
///
Relative comparisons with numbers:
Files that were NOT gain-reduced (So that would be those that were exported from Reaper before UA-cam's loudness-penalty. The original source-files with optimum quality if you will. The songs you would put on a CD)
No dither TruePeak maximum: -13.7 dB FS / integrated -17.0 LUFS / Crest Factor: 3.3 dB
Reaper dither with NO noise-shaping TruePeak maximum: -10.3 dB FS / integrated -18.6 LUFS / Crest Factor: 8.3 dB
Reaper dither WITH noise-shaping TruePeak maximum: -6.4 dB FS / integrated -14.2 LUFS / Crest Factor: 7.8 dB
LACE dither TPDF TruePeak maximum: -10.6 dB FS / integrated -17.7 LUFS / Crest Factor: 7.1 dB
LACE dither Psychoacoustic TruePeak maximum: -7.1 dB FS / integrated -15.1 LUFS / Crest Factor: 8 dB
LACE dither Entropic TruePeak maximum: -2.7 dB FS / integrated -11.1 LUFS / Crest Factor: 8.4 dB
Files that WERE gain-reduced -6 dB (So that would be those that were struck lightly by UA-cam's loudness-penalty and lowered in volume)
No dither TruePeak maximum: -16.9 dB FS / integrated -20.3 LUFS / Crest Factor: 3.4 dB
Reaper dither with NO noise-shaping TruePeak maximum: -15.3 dB FS / integrated -21.3 LUFS / Crest Factor: 6 dB
Reaper dither WITH noise-shaping TruePeak maximum: -11.1 dB FS / integrated -18.1 LUFS / Crest Factor: 7 dB
LACE dither TPDF TruePeak maximum: -15.3 dB FS / integrated -20.1 LUFS / Crest Factor: 4.8 dB
LACE dither Psychoacoustic TruePeak maximum: -12.0 dB FS / integrated -18.2 LUFS / Crest Factor: 6.2 dB
LACE dither Entropic TruePeak maximum: -7.9 dB FS / integrated -15.4 LUFS / Crest Factor: 7.5 dB
Files that WERE gain-reduced -12 dB (So that would be those that were struck hard by UA-cam's loudness-penalty and lowered even more in volume)
No dither TruePeak maximum: -22.0 dB FS / integrated -23.5 LUFS / Crest Factor: 1.5 dB
Reaper dither with NO noise-shaping TruePeak maximum: -19.2 dB FS / integrated -23.9 LUFS / Crest Factor: 4.7 dB
Reaper dither WITH noise-shaping TruePeak maximum: -16.3 dB FS / integrated -23.1 LUFS / Crest Factor: 6.8 dB
LACE dither TPDF TruePeak maximum: -19.2 dB FS / integrated -23.6 LUFS / Crest Factor: 4.4 dB
LACE dither Psychoacoustic TruePeak maximum: -17.2 dB FS / integrated -24.0 LUFS / Crest Factor: 6.8 dB
LACE dither Entropic TruePeak maximum: -14.1 dB FS / integrated -20.6 LUFS / Crest Factor: 6.5 dB

[@element64]

Hi .You disabled comments on your ai video .I wanted to ask you a question?
.
I created a fun video on generative tools & was finalising a dj mix with randomised generations. It consisted of , drum breaks , esoteric textures , female voices & funky soundtrack themes from the late 60s to the 70s .I never specified any persons in a prompt, I now know several weeks later that many of the voices in the generations have passed away.
.
The tech giants & devs won't provide transparency & declare the human the person , the voice model source after several requests.
This changes the essence of my efforts. As the voices are unauthorised which is very uncomfortable revelation.
.
You have content on platforms & had to provide credit to the generative tools.Yet the devs are not disclosing vocal sources. Are you aware or comfortable that unconsenting deceased , unauthorised persons, can be hallucinating on your instrumental tracks?
I can provide examples.

[@JELmusic]

I didn't turn off comments, so that must've been a youtube-glitch. I've just turned them back on and checked they're on, so you can go back and try adding your comment on that video again (So people with relevant knowledge may read it and reply)
I will check back on the video from time to time in the coming days to see if youtube turns the comments off again. If they do, then I will have to find out why that happens so I can keep them on.
You're very welcome to comment on that video, as open and honest debate is always good :)
And if you have examples of Udio doing anything wrong, feel free to post them there.