I found this very recent ( 9 April 2019 ) CERN paper (pdf) that may be of interest to you..
It is 28 pages, but the last 20 pages are all references!
It is a very interesting read. One takeaway is the extreme complexity and signal processing that goes into making these measurements. I would argue that the complexity is at least comparable with the complexity behind the measurement of the Cosmic Backgroud Radiation, even including mapping the anisotropy.
I have to wonder then, why you seem willing to accept these measurements done by CERN yet have such deep reservations about the measurement of the CMB BB curve?
I tend to think that measuring and plotting the CMB BB curve is much simpler than the observation of the light-by-light scattering process, γγ → γγ, in Pb+Pb collisions at 5.02 TeV.
My answer might be poor and unconvincing, but I'll try to make it.
First at all, from my point of view, the complexity of calculations, algorithms and second and third order theories behind the data analysis of the whole CBR has
increased exponentially from COBE to WMAP to PLANCK satellites, as the databases increased their size with each generation. It take YEARS to compute and
refine raw information from Planck (it took almost FOUR YEARS for the COBE dataset).
There is no comparison in complexities.
My answer is: I do believe that photons have an electromagnetic mass. So CERN results are aligned with my thought on this matter.
But I don't believe that the Universe behave as a black body cavity, because it violates the origin of everything: Kirchoff's law for
thermal radiation, which led to solve his challenge to find the spectral distribution of energy WITHIN a black body cavity.
The Universe has no depth, so absorptivitty can't be applied. Also, I believe that radiation sources are anisotropic and can't be
correctly evaluated in strength and depth. Not to mention that our Milky Way emits random radiation in the same bands, which is
millions of times much more powerful than background RANDOM noise.
And, as an electrical engineer with many, many years working in telecommunications (radiowaves, optical signals) the first thing
that you understand and after can measure is THIS: You can't recover a random signal deeply buried into random noise.
All the theorethical physicists, for decades, have written volumes about this fact, related to the Information Theory and Noise.
The only way it can work (used by the military and civilian mobile techniques) is to pseudo-randomize your signal with a known
pseudo-random encoding sequence. Then, you known what to expect and can send a signal below the noise level OR, as in
mobile telephony using CDMA, you can send a thousand digital voice channels in the same bandwidth. Each channel has its
own pseudo-random encoder and decoder so, even when the other channels appear as noise to your receptor and at higher
level than your expected signal, you can recover it and hear perfectly.
As you can see, I believe in one thing and don't believe in the other. As simple as that (or complex, if I have to proof the CBR