more benchmark_xl output

[jonsneyers]

Fixes #3585

Also adds more details when using the flag --print_distance_percentiles.

Also fixes the aggregate value for ssimulacra2: it was ignoring negative values and taking the geomean of just the positive ones; now it just takes the arithmetic mean of all ssimulacra2 values.

Before:

Encoding      kPixels    Bytes          BPP  E MP/s  D MP/s     Max norm  SSIMULACRA2   PSNR        pnorm       BPP*pnorm   QABPP   Bugs
----------------------------------------------------------------------------------------------------------------------------------------
50th/90th percentile distance: 0.74641287  0.89263189
jxl:d0.5        13562  4936672    2.9119624   2.323  34.301   0.76519037  91.52246513  46.23   0.34268783  0.997894059218   2.913      0
50th/90th percentile distance: 1.30287242  1.55675101
jxl:d1          13562  3045965    1.7967034   2.239  34.252   1.35021809  86.95225346  42.34   0.58427450  1.049767991430   2.440      0
50th/90th percentile distance: 2.29967427  2.89556241
jxl:d2          13562  1835797    1.0828695   2.292  32.952   2.41013471  78.73319052  38.83   0.97862361  1.059721707947   2.693      0
50th/90th percentile distance: 3.09834218  3.91219139
jxl:d3          13562  1334015    0.7868867   2.248  33.389   3.23678298  71.66090226  36.88   1.31249357  1.032783716883   2.611      0
50th/90th percentile distance: 15.69093132  17.99648857
jxl:d24         13562   270716    0.1596855   0.605  38.327  16.06720416  -2.15884859  29.27   5.74953159  0.918116671120   2.607      0
50th/90th percentile distance: 16.13576889  18.36326790
jxl:d25         13562   261793    0.1544221   0.606  35.376  16.68272962  -3.80875468  29.17   5.89834294  0.910834629742   2.593      0
Aggregate:      13562  1173373    0.6921298   1.463  34.723   3.59547238  81.85819581  36.57   1.43471342  0.993007961883   2.639      0

After:

46 images
Encoding      kPixels    Bytes          BPP  E MP/s  D MP/s     Max norm  SSIMULACRA2   PSNR        pnorm       BPP*pnorm   QABPP   Bugs
----------------------------------------------------------------------------------------------------------------------------------------
   _________________________________________| median:         0.74641287  91.96166229          0.33922365
  /                                         | p10 (worst):    0.89263189  89.77494049          0.38702947
 /                                          | p5 (worst):     0.93385208  89.31037903          0.39110589
jxl:d0.5        13562  4936672    2.9119624   2.315  31.096   0.76519037  91.52246533  46.23   0.34268782  0.997894047755   2.913      0
   _________________________________________| median:         1.30287242  88.15885162          0.57385707
  /                                         | p10 (worst):    1.55675101  82.56882477          0.70896733
 /                                          | p5 (worst):     1.61269927  82.47370911          0.72702777
jxl:d1          13562  3045965    1.7967034   2.205  34.431   1.35021809  86.95225222  42.34   0.58427450  1.049767990116   2.440      0
   _________________________________________| median:         2.29967427  80.39791870          0.95846683
  /                                         | p10 (worst):    2.89556241  70.31759644          1.26363790
 /                                          | p5 (worst):     2.96536636  68.60237885          1.32267296
jxl:d2          13562  1835797    1.0828695   2.306  34.101   2.41013471  78.73319006  38.83   0.97862363  1.059721729400   2.693      0
   _________________________________________| median:         3.09834218  73.34270477          1.27836907
  /                                         | p10 (worst):    3.91219139  60.53897095          1.67121387
 /                                          | p5 (worst):     4.07968855  55.49774170          1.81831229
jxl:d3          13562  1334015    0.7868867   2.316  32.388   3.23678298  71.66090188  36.88   1.31249357  1.032783719159   2.611      0
   _________________________________________| median:        15.69093132   5.88187265          6.02538586
  /                                         | p10 (worst):   17.99648857 -26.42949295          6.62558889
 /                                          | p5 (worst):    18.10667038 -35.92488098          6.94895744
jxl:d24         13562   270716    0.1596855   0.606  34.924  16.06720416  -2.15884861  29.27   5.74953159  0.918116671543   2.607      0
   _________________________________________| median:        16.13576889   4.86366034          6.11648321
  /                                         | p10 (worst):   18.36326790 -28.11663246          6.71656990
 /                                          | p5 (worst):    18.51855659 -36.31728745          7.22389078
jxl:d25         13562   261793    0.1544221   0.609  37.065  16.68272962  -3.80875474  29.17   5.89834299  0.910834637497   2.593      0
Aggregate:      13562  1173373    0.6921298   1.469  33.948   3.59547238  53.81686769  36.57   1.43471343  0.993007964975   2.639      0