I mean, Jason Cammisa is not some "YouTube person". He's arguably his generation's Jeremy Clarkson.
e.g.: https://www.youtube.com/playlist?list=PLHKCmmH-x9mIbtnKiNfg2...
He tells the stories, interviews with the people who built these cars (if possible), gets and drives a copy of the said car.
We're talking about a person who painstakingly perfects the pronunciation of the brand and model names just because he feels doing otherwise is not respectful for the brand, model and people involved with the car.
---
Let's talk about specs. The spec[0] (pg 4) states that:
Photometers are placed at fixed locations on the test track to record the visibility and glare illumination of
the test vehicle on each approach. To correct for changes in illumination that are due to changes in
vehicle pitch, multiple photometers are used at each measurement location to capture illuminance
readings at different heights. The illuminance readings are synchronized to the vehicle position and pitch
using a common GPS time signal. The synchronized data are used to produce pitch-corrected illuminance
versus distance curves that are used for the headlight rating. All data are processed using the DIAdem
software package distributed by National Instruments. The processing scripts are available at
https://github.com/iihs-hldi.
This is a fixed receptor, fixed path, multiple approach test and is susceptible to optimization of the illumination map designed by the car manufacturers.
Moreover, test document states that:
Illuminance data are collected with Gamma Scientific photometers (Part # U68401). The photometric
sensors provide a very close match to the spectral response of the human eye. They are fitted with
diffusers to reduce the illuminance measurements for off-axis incidence angles in accordance with
Lambert’s cosine law. The sensors match the targeted cosine response to within 3 percent at angles up to
25 degrees, which is the maximum angle between the test vehicle and sensors on the sharpest curve (at
distances greater than 10 m). The sensor signals are passed through a low-pass filter with a cutoff
frequency of 35 Hz to allow for accurate measurements of pulse width modulated light sources such as
LEDs. Each sensor is connected to its own transimpedance amplifier board that has fixed gains to yield a
fast response while still minimizing linearity errors in the range of illuminance values for which the
headlight ratings are assigned.
Again more places to optimize the headlight.
If the test finds the lights are "in-spec", and people are increasingly unhappy/uncomfortable, then something is wrong.
[0]: https://www.iihs.org/media/0e823704-32d1-4500-b095-15d064d82...