# flux from 2.74E-6 sr ( 8.61E-8 sr/4px) is  0.00388 lx;
 # tilewise luminance  31.3 cd/m2 with rel. SD  695 % (points with 
  luminances over  6.0E2 )
# flux from 1.20E-5 sr ( 8.61E-8 sr/4px) is  0.00231 lx;
 # tilewise luminance  6.00 cd/m2 with rel. SD  553 % (points with 
  luminances over   50    and below  6.0E2 )
# flux from 3.52E-5 sr ( 8.61E-8 sr/4px) is  2.90E-4 lx;
 # tilewise luminance  0.500 cd/m2 with rel. SD  444 % (points with 
  luminances over  1.5    and below   50   )
  
 # interim sum                               0.00648 lx
  
# flux from 1.92E-4 sr ( 8.61E-8 sr/4px) is  7.67E-5 lx;
 # tilewise luminance  0.0324 cd/m2 with rel. SD  362 % (points with 
  luminances over  0.20   and below  1.5   )
# flux from 0.0677 sr ( 8.61E-8 sr/4px) is  2.56E-4 lx;
 # tilewise luminance  0.00349 cd/m2 with rel. SD  214 % (points with 
  luminances below  0.20  )

 # sum of <1.5 cd/m2                         3.33E-4 lx, 5 % of the above sum 

Sums for the new lighting:

File name    : 10.jpeg
# flux from 0.0580*0.59 sr ( 7.57E-8 sr/4px) is  2.23E-4 lx ( 3.43E-4 lx tilewise
# tilewise luminance  0.00552 cd/m2 with rel. SD  224 % 
(points with luminances
 below  1.5    and with non-zero green: there is 59.3 % of them)

                                       About the same as before.
 
 File name    : 11.jpeg
# flux from 1.10E-5 sr ( 7.57E-8 sr/4px) is  6.22E-5 lx ( 6.80E-5 lx tilewise);
# tilewise luminance  0.249 cd/m2 with rel. SD  497 % (points with 
luminances
 over  1.5    and below   50   )
 
 File name    : 12.jpeg
 # flux from 4.54E-7 sr ( 7.57E-8 sr/4px) is  8.38E-5 lx ( 1.02E-4 lx  tilewise);
# tilewise luminance  1.36 cd/m2 with rel. SD 1.28E3 % (points with 
luminances
 over   50   )
But the single overexposed pixel brings about another 1E-4 lx...
 
So, there is about 2.6E-4 from strong sources over 1.5 cd/m2: 
 25x less then before.