Does not repro on 66.0.3359.158 on sentry chromebook.

Your "about:gpu" says that you have a color profile installed with a pure gamma 2.2. With a true gamma 2.2 monitor, the first few values all produce *very* little light. (The "1" value is only 0.000005 as bright as "255".) Since the "1" value in the video is meant to be brighter than this, chrome will translate the "1" value in the video to the output value that is closest to the right brightness, which is probably "6".

If your monitor really is a gamma 2.2 monitor, this should be correct.
Unfortunately, it's likely that your monitor is not in fact a true gamma 2.2 monitor, but instead uses a gamma function which is close to 2.2, but has a flat region at the bottom. This means that the "6" value that chrome outputs doesn't produce the brightness that chrome expects. (Based on your ICC profile.)

Could you let us know what monitor and color profile you are using?

I'm using an LG 27UD58-B.

I've attached the .icc that I'm using.  It does appear to have a flat region at the bottom, according to the GNOME utility.

Deleted: edid-e3da03028f8abca404ffea6bf3e49950.icc 1.5 KB

edid-e3da03028f8abca404ffea6bf3e49950.icc 1.5 KB Download

Wonder if we're approximating the ICC profile wrong then?

It might be of note that I am using a Dell XPS 13 9350, with this monitor attached as an external monitor.  Should I also upload the 9350 .icc as well?

Also I'm pulling the .icc from what the GNOME 3 Color dialogue says.  Is there a more authoritative place I should be pulling it from?

The gnome color dialogue should be fine I think.
Chrome will read the ICC profile from X properties set on the root window, so you could look there, but there is no easy way to actually extract the data.
The second monitor thing may be part of the problem as until recently chrome would always use the color profile of the primary monitor on linux. (Fixed in version 67.0.3388.0 or later.) (Note that what you think is "primary" might not be what X11 think is "primary" though.)

Does the video look fine when you show it on the LG monitor and bad when you show it on the XPS monitor? (or vice versa?)

BTW, I looked at the ICC profile you uploaded in ICC Profile Inspector, and I don't see a flat region towards the bottom, which would mean that chrome is interpreting it correctly, unless I'm missing something.

I think the solution might be to find a better ICC profile for your monitor. Alternatively, you can go into about:flags in chrome and force an sRGB color profile.

Video looks wrong on both monitors.

I say that there seems to be a flat region towards the low end based on the TRC curve that GNOME Color is displaying.  Am I misinterpreting this graph/is it not actually relevant?

Deleted: Screenshot from 2018-05-17 20-30-30.png 97.3 KB

	Screenshot from 2018-05-17 20-30-30.png 97.3 KB View Download

That graph is weird. I don't know what it actually is trying to show us.
Possibly it's showing the difference between your profile and an sRGB profile.

This link shows how close sRGB and gamma 2.2 curves are:

 http://fooplot.com/#W3sidHlwZSI6MCwiZXEiOiJ4XjIuMiIsImNvbG9yIjoiIzFFMDBGRiJ9LHsidHlwZSI6MCwiZXEiOiIoeDw9MC4wNDA0NSkqeC8xMi45MisoeD4wLjA0MDQ1KSooKHgrMC4wNTUpLzEuMDU1KV4yLjQiLCJjb2xvciI6IiNGRjAwMDAifSx7InR5cGUiOjEwMDAsIndpbmRvdyI6WyItMCIsIjEiLCItMCIsIjEiXX1d

Note that if you zoom in near 0,0 you'll notice that the blue curve is really close to zero at the beginning, which is the effect you're seeing. Unfortunately, many monitor manufactures seem to view this as "close enough" for some reason.

Alright - I've forced an sRGB profile and it works around the issue.

It seems like the ICC profile I'm getting from EDID data (I think that's where it comes from?) is not great.  Once I own calibration tools I'll flip this workaround back off. 

Willing to mark as NOTABUG/whatever the equivalent is here.

WAI, Working as Intended