Configurando a distância do quadro-chave no ffmpeg conforme o Adobe Media Encoder

Question

Estou trabalhando com vídeo em um site que exige que o material seja procurado tanto para frente quanto para trás. É semelhante ao efeito que a Apple emprega em sua página para o novo MacBook Pro. Se você observar a maior parte da imagem do Laptop, ela terá um vídeo dinâmico que será reproduzido dependendo da sua posição de rolagem.

Agora, consegui criar um vídeo mp4 que funciona tanto para frente quanto para trás com pouca ou nenhuma falha, empregando a Distância do quadro-chave definida como 1 no Adobe Media Encoder:

Eu ainda não repliquei usando o CLI e o ffmpeg. Eu tentei executar este comando:

ffmpeg -i Vulkano-HD_1.mp4 -vcodec libx264 -x264-params keyint=96:min-keyint=96:scenecut=-1 -an out2.mp4

A fonte é 4s e estou tentando forçar 96 quadros-chave (4 * 24fps), mas não consigo fazer funcionar. O arquivo criado com o Adobe Media Encoder sai em 4.7MB, enquanto o cli é 1.6MB, então algo está definitivamente em alta.

Alguma idéia do que estou fazendo errado aqui?

Log completo

ffmpeg version 3.0.2 Copyright (c) 2000-2016 the FFmpeg developers
  built with Apple LLVM version 7.3.0 (clang-703.0.29)
  configuration: --prefix=/usr/local/Cellar/ffmpeg/3.0.2 --enable-shared --enable-pthreads --enable-gpl --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-opencl --enable-libx264 --enable-libmp3lame --enable-libxvid --enable-vda
  libavutil      55. 17.103 / 55. 17.103
  libavcodec     57. 24.102 / 57. 24.102
  libavformat    57. 25.100 / 57. 25.100
  libavdevice    57.  0.101 / 57.  0.101
  libavfilter     6. 31.100 /  6. 31.100
  libavresample   3.  0.  0 /  3.  0.  0
  libswscale      4.  0.100 /  4.  0.100
  libswresample   2.  0.101 /  2.  0.101
  libpostproc    54.  0.100 / 54.  0.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'Vulkano-HD_1.mp4':
  Metadata:
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: mp42mp41
    creation_time   : 2016-12-19 13:33:58
  Duration: 00:00:04.04, start: 0.000000, bitrate: 9382 kb/s
    Stream #0:0(eng): Video: h264 (Main) (avc1 / 0x31637661), yuv420p(tv, bt709), 1920x816, 9358 kb/s, 24 fps, 24 tbr, 24k tbn, 48 tbc (default)
    Metadata:
      creation_time   : 2016-12-19 13:33:58
      handler_name    : Alias Data Handler
      encoder         : AVC Coding
[libx264 @ 0x7fc67d807000] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 AVX2 LZCNT BMI2
[libx264 @ 0x7fc67d807000] profile High, level 4.0
[libx264 @ 0x7fc67d807000] 264 - core 148 r2668 fd2c324 - H.264/MPEG-4 AVC codec - Copyleft 2003-2016 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=12 lookahead_threads=2 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=96 keyint_min=49 scenecut=0 intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to 'out2.mp4':
  Metadata:
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: mp42mp41
    encoder         : Lavf57.25.100
    Stream #0:0(eng): Video: h264 (libx264) ([33][0][0][0] / 0x0021), yuv420p, 1920x816, q=-1--1, 24 fps, 12288 tbn, 24 tbc (default)
    Metadata:
      creation_time   : 2016-12-19 13:33:58
      handler_name    : Alias Data Handler
      encoder         : Lavc57.24.102 libx264
    Side data:
      unknown side data type 10 (24 bytes)
Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> h264 (libx264))
Press [q] to stop, [?] for help
frame=   57 fps=0.0 q=28.0 size=      27kB time=-00:00:00.04 bitrate=N/A speed=Nframe=   80 fps= 78 q=28.0 size=     343kB time=00:00:00.91 bitrate=3062.6kbits/frame=   97 fps= 63 q=28.0 size=     638kB time=00:00:01.62 bitrate=3215.4kbits/frame=   97 fps= 35 q=-1.0 Lsize=    1598kB time=00:00:03.95 bitrate=3306.2kbits/s speed=1.43x    
video:1596kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.086879%
[libx264 @ 0x7fc67d807000] frame I:2     Avg QP:21.50  size: 28347
[libx264 @ 0x7fc67d807000] frame P:86    Avg QP:22.75  size: 17547
[libx264 @ 0x7fc67d807000] frame B:9     Avg QP:22.97  size:  7561
[libx264 @ 0x7fc67d807000] consecutive B-frames: 81.4% 18.6%  0.0%  0.0%
[libx264 @ 0x7fc67d807000] mb I  I16..4: 26.4% 69.2%  4.5%
[libx264 @ 0x7fc67d807000] mb P  I16..4: 19.2% 26.8%  0.8%  P16..4: 29.9%  4.8%  1.0%  0.0%  0.0%    skip:17.6%
[libx264 @ 0x7fc67d807000] mb B  I16..4:  2.8%  3.3%  0.0%  B16..8: 37.6%  3.9%  0.2%  direct: 1.9%  skip:50.2%  L0:43.2% L1:53.1% BI: 3.6%
[libx264 @ 0x7fc67d807000] 8x8 transform intra:57.8% inter:94.6%
[libx264 @ 0x7fc67d807000] coded y,uvDC,uvAC intra: 24.4% 40.6% 2.3% inter: 13.1% 24.4% 0.0%
[libx264 @ 0x7fc67d807000] i16 v,h,dc,p: 32% 17% 15% 36%
[libx264 @ 0x7fc67d807000] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 14% 27%  4%  4%  6%  3%  4%  2%
[libx264 @ 0x7fc67d807000] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 36% 13% 16%  4% 10% 10%  5%  4%  2%
[libx264 @ 0x7fc67d807000] i8c dc,h,v,p: 56% 17% 23%  4%
[libx264 @ 0x7fc67d807000] Weighted P-Frames: Y:0.0% UV:0.0%
[libx264 @ 0x7fc67d807000] ref P L0: 74.1% 10.4% 11.4%  4.1%
[libx264 @ 0x7fc67d807000] ref B L0: 83.1% 16.9%
[libx264 @ 0x7fc67d807000] kb/s:3233.87