Compress any video to 8mb similar to 8mb.video. Usage: "./to8mb.sh input.mp4"

to8mb.sh

#!/usr/bin/env bash
set -Eeuo pipefail
export LC_ALL=C

usage() {
  printf 'Usage: %s INPUT [OUTPUT]\n' "$(basename "$0")" >&2
  exit 2
}

die() {
  printf 'Error: %s\n' "$*" >&2
  exit 1
}

[[ $# -ge 1 && $# -le 2 ]] || usage

input=$1
output=${2:-}
limit_bytes=${LIMIT_BYTES:-8000000}
target_bytes=$((limit_bytes - 100000))

[[ -f "$input" ]] || die "input file not found: $input"
(( limit_bytes > 500000 )) || die "LIMIT_BYTES is too small"

for cmd in ffmpeg ffprobe awk stat mktemp; do
  command -v "$cmd" >/dev/null || die "missing dependency: $cmd"
done

if [[ -z "$output" ]]; then
  dir=$(dirname -- "$input")
  file=$(basename -- "$input")
  stem=${file%.*}
  output="$dir/${stem}_8mb.mp4"
fi

[[ "$input" != "$output" ]] || die "input and output paths must differ"

video_stream=$(ffprobe -v error -select_streams V:0 \
  -show_entries stream=index -of csv=p=0 "$input" | head -n1)
[[ -n "$video_stream" ]] || die "no video stream found"

duration=$(ffprobe -v error -show_entries format=duration \
  -of default=noprint_wrappers=1:nokey=1 "$input" | head -n1)
awk -v d="$duration" 'BEGIN { exit !(d > 0) }' || die "could not determine video duration"

has_audio=$(ffprobe -v error -select_streams a:0 \
  -show_entries stream=index -of csv=p=0 "$input" | head -n1 || true)

# Reserve about 1.5% for MP4/container overhead.
total_kbps=$(awk -v b="$target_bytes" -v d="$duration" \
  'BEGIN { printf "%d", (b * 8 / d / 1000) * 0.985 }')

if [[ -n "$has_audio" ]]; then
  if   (( total_kbps >= 1200 )); then audio_kbps=128
  elif (( total_kbps >= 600  )); then audio_kbps=96
  elif (( total_kbps >= 300  )); then audio_kbps=64
  elif (( total_kbps >= 180  )); then audio_kbps=48
  else                              audio_kbps=32
  fi
else
  audio_kbps=0
fi

video_kbps=$((total_kbps - audio_kbps))

if (( video_kbps < 40 && audio_kbps > 0 )); then
  audio_kbps=0
  video_kbps=$total_kbps
fi
(( video_kbps >= 10 )) || die "video is too long to fit meaningfully within 8 MB"

if   (( video_kbps >= 2200 )); then max_dim=1920
elif (( video_kbps >= 900  )); then max_dim=1280
elif (( video_kbps >= 500  )); then max_dim=960
elif (( video_kbps >= 300  )); then max_dim=854
elif (( video_kbps >= 160  )); then max_dim=640
else                              max_dim=426
fi

vf="scale='min(iw,${max_dim})':'min(ih,${max_dim})':force_original_aspect_ratio=decrease:force_divisible_by=2,setsar=1"

tmpdir=$(mktemp -d)
trap 'rm -rf "$tmpdir"' EXIT
passlog="$tmpdir/pass"
tmp_output="$tmpdir/output.mp4"

encode() {
  local bitrate=$1
  rm -f "${passlog}"* "$tmp_output"

  ffmpeg -hide_banner -loglevel warning -stats -y -i "$input" \
    -map 0:V:0 -vf "$vf" -fpsmax 30 \
    -c:v libx264 -preset slow -b:v "${bitrate}k" -pix_fmt yuv420p \
    -pass 1 -passlogfile "$passlog" \
    -an -sn -dn -f null /dev/null

  audio_args=(-an)
  if (( audio_kbps > 0 )); then
    audio_args=(-map '0:a:0?' -c:a aac -b:a "${audio_kbps}k" -ac 2)
    (( audio_kbps <= 48 )) && audio_args=(-map '0:a:0?' -c:a aac -b:a "${audio_kbps}k" -ac 1)
  fi

  ffmpeg -hide_banner -loglevel warning -stats -y -i "$input" \
    -map 0:V:0 "${audio_args[@]}" -vf "$vf" -fpsmax 30 \
    -c:v libx264 -preset slow -b:v "${bitrate}k" -pix_fmt yuv420p \
    -pass 2 -passlogfile "$passlog" \
    -sn -dn -map_metadata -1 -map_chapters -1 \
    -movflags +faststart "$tmp_output"
}

printf 'Duration: %.1f s\n' "$duration"
printf 'Video: %d kbps, audio: %d kbps, max dimension: %d px\n' \
  "$video_kbps" "$audio_kbps" "$max_dim"

for attempt in 1 2 3; do
  printf 'Encoding pass set %d/3...\n' "$attempt"
  encode "$video_kbps"
  actual_bytes=$(stat -c %s "$tmp_output")

  if (( actual_bytes <= limit_bytes )); then
    break
  fi

  video_kbps=$(awk -v r="$video_kbps" -v max="$limit_bytes" -v got="$actual_bytes" \
    'BEGIN { n = int(r * (max * 0.98 / got)); if (n < 10) n = 10; print n }')
done

actual_bytes=$(stat -c %s "$tmp_output")
(( actual_bytes <= limit_bytes )) || die "could not get output below 8 MB after 3 attempts"

mv -f -- "$tmp_output" "$output"
printf 'Created: %s (%.2f MB)\n' "$output" "$(awk -v b="$actual_bytes" 'BEGIN { print b / 1000000 }')"