ffmpeg loudnorm results in ... not very loud

Question

I'm mixing together 27 separate video (and audio) streams, and I'm trying to use the loudnorm filter to set my output audio to a reasonable level.

I run the loudnorm filter in two passes using information I read here, for the first pass the filters look like this:

[0:a]volume=0.90,aresample=async=1:first_pts=0[aud0]
[1:a]volume=2.08,aresample=async=1:first_pts=0[aud1]
[2:a]volume=1.48,aresample=async=1:first_pts=0[aud2]
[3:a]volume=2.53,aresample=async=1:first_pts=0[aud3]
[4:a]volume=1.34,aresample=async=1:first_pts=0[aud4]
[5:a]volume=1.81,aresample=async=1:first_pts=0[aud5]
[6:a]volume=1.19,aresample=async=1:first_pts=0[aud6]
[7:a]volume=3.01,aresample=async=1:first_pts=0[aud7]
[8:a]volume=1.61,aresample=async=1:first_pts=0[aud8]
[9:a]volume=1.85,aresample=async=1:first_pts=0[aud9]
[10:a]volume=1.92,aresample=async=1:first_pts=0[aud10]
[11:a]volume=1.36,aresample=async=1:first_pts=0[aud11]
[12:a]volume=3.05,aresample=async=1:first_pts=0[aud12]
[13:a]volume=1.97,aresample=async=1:first_pts=0[aud13]
[14:a]volume=2.93,aresample=async=1:first_pts=0[aud14]
[15:a]volume=5.03,aresample=async=1:first_pts=0[aud15]
[16:a]volume=4.03,aresample=async=1:first_pts=0[aud16]
[17:a]volume=7.84,aresample=async=1:first_pts=0[aud17]
[18:a]volume=29.81,aresample=async=1:first_pts=0[aud18]
[19:a]volume=31.59,aresample=async=1:first_pts=0[aud19]
[20:a]volume=6.00,aresample=async=1:first_pts=0[aud20]
[21:a]volume=1.45,aresample=async=1:first_pts=0[aud21]
[22:a]volume=1.75,aresample=async=1:first_pts=0[aud22]
[23:a]volume=4.17,aresample=async=1:first_pts=0[aud23]
[24:a]volume=2.83,aresample=async=1:first_pts=0[aud24]
[25:a]volume=4.55,aresample=async=1:first_pts=0[aud25]
[26:a]volume=2.66,aresample=async=1:first_pts=0[aud26]
[aud0][aud1][aud2][aud3][aud4][aud5][aud6][aud7][aud8][aud9][aud10][aud11][aud12][aud13][aud14][aud15][aud16][aud17][aud18][aud19][aud20][aud21][aud22][aud23][aud24][aud25][aud26]amix=inputs=27[a]
[a]loudnorm=I=-24:LRA=11:tp=-2:print_format=json

And loudnorm reports this:

"input_i": "-29.92"
"input_tp": "-15.22"
"input_lra": "17.10"
"input_thresh": "-41.11"
"output_i": "-22.93"
"output_tp": "-7.56"
"output_lra": "12.20"
"output_thresh": "-33.57"
"normalization_type": "dynamic"
"target_offset": "-1.07"

so I run it again, using the values from the first pass of loudnorm:

[a]adelay=delays=200:all=1[b]
[b]loudnorm=linear=true:I=-24:LRA=11:tp=-2:measured_I=-29.92:measured_LRA=17.10:
   measured_tp=-15.22:measured_thresh=-41.11:offset=-1.07:print_format=summary[level]
[level]asplit[a1][a2]

I then re-attach the video (which has its own interesting processing) and audio:

-map "[video]" -map "[a1]" -c:v libx264 "output.mp4"  -map "[a2]" "output.mp3""

and the output video just seems really quiet to me. I was trying to figure out a way to quantify that, and the only thing I've got is that when I drop it in Sony Movie Studio, the audio is clearly nowhere near as loud as it could be:

Am I calling loudnorm wrong somehow?