FFMPEG - Accurately cutting video/audio and merging multiple videos together - ffmpeg

Question

This post has been updated since the original post
thanks in advance for any help I get with this...
I am working with 4 video assets....all stereo audio, same a/v spec.
intro -
mainvideo -
midroll_video (advert) -
endboard
I need to add an audio only cross fade between the 2 video elements part2.mp4 and pip.mp4. These 2 videos are made by the code (rather than being 2 of the 4 videos listed above). I have added in the code as kindly instructed by #Баяр Гончикжапов but unfortunately it is still not working. See conversation below for more info/ the part i need help with.
Thanks in advance!
This is the code I am using.
# Input parameters
mainvideo=mezzfile.mp4
endboard=endboard.mp4
intro=sting.mp4
midroll_video=midroll.mp4
tailtime=20
fadelength=0.2
midroll_edit_value=00:11:31.600
# Time calculations to define cut point
duration=$(ffprobe -v error -show_entries format=duration -of default=noprint_wrappers=1:nokey=1 $mainvideo)
endboard_cut_point=$(echo "scale=2;$duration-$tailtime" | bc)
duration=$(ffprobe -v error -show_entries format=duration -of default=noprint_wrappers=1:nokey=1 $mainvideo)
midroll_cut_point=$(echo "scale=2;$duration-$midroll_edit_value" | bc)
# Safety check
tbn_main=$(ffprobe -v error -select_streams v -show_entries stream=time_base -of default=noprint_wrappers=1:nokey=1 $mainvideo)
tbn_main=${tbn_main#*/}
tbn_intro=$(ffprobe -v error -select_streams v -show_entries stream=time_base -of default=noprint_wrappers=1:nokey=1 $intro)
tbn_intro=${tbn_intro#*/}
tbn_end=$(ffprobe -v error -select_streams v -show_entries stream=time_base -of default=noprint_wrappers=1:nokey=1 $endboard)
tbn_end=${tbn_end#*/}
if [[ $(( ($tbn_intro+$tbn_main+$tbn_end)/3 )) -ne $tbn_main ]]; then
echo "WARNING: source video files have the different timebase."
echo "The use of the concat demuxer will produce incorrect output."
echo "Re-encoding is highly recommended."
read -s -k $'?Press any key to exit.\n'
exit 1
fi
# Trim the main part of mainvideo
ffmpeg -hide_banner -y -i $mainvideo -to $midroll_edit_value -c copy part1.mp4
ffmpeg -hide_banner -y -i $mainvideo -ss $midroll_edit_value -to $endboard_cut_point -c copy part2.mp4
ffmpeg -hide_banner -y -i $mainvideo -ss $midroll_edit_value -to $duration -c copy part2av.mp4
# Trim the tail of mainvideo and overlay it onto endboard
ffmpeg -hide_banner -y \
-i $mainvideo \
-i $endboard \
-filter_complex \
"[0:v]select='gt(t,$duration-$tailtime)',scale=w=iw/2.03:h=ih/2.03,setpts=PTS-STARTPTS[v_tail]; \
[0:a]aselect='gt(t,$duration-$tailtime)',asetpts=PTS-STARTPTS[a_out]; \
[1:v][v_tail]overlay=format=auto[v_out]" \
-map "[v_out]" \
-map "[a_out]" \
-video_track_timescale $tbn_main \
pip.mp4
ffmpeg -hide_banner -y -i "part2.mp4" -i "part2av.mp4" -i "pip.mp4" \
-filter_complex \
-map "[0:v];"
-map "[1:a];"
-map "[2:v]"
-c copy "part2pip.mp4" \
# Pass all parts through the concat demuxer
[ -f filelist.txt ] && rm filelist.txt
for f in $intro part1.mp4 $midroll_video part2pip.mp4; do echo "file '$PWD/$f'" >> filelist.txt; done
ffmpeg -hide_banner -y -f concat -safe 0 -i filelist.txt -c copy TEST_FILE.mp4
# Sweep the table
rm pip.mp4 part1.mp4 part2.mp4 filelist.txt```

Answer 1

Insert video in the middle of main video and use -f concat:
# Input parameters.
# All video has same properties: codec_name, width, height, sample_aspect_ratio, r_frame_rate, time_base, pix_fmt
# if one of the video has difference then this video needs to be transcoded
# All audio has same properties: codec_name, channels, sample_rate
# if audio has diff: aresample=async=1,apad,atrim=0:$duration -ac 2
# get durations of audio and video and container: ffprobe -v quiet -show_entries format=duration:stream=duration -of default=nw=1:nk=1 "$f"
intro=VID_20230206_113828.mp4
mainvideo=VID_20230206_113949.mp4
endboard=VID_20230206_114456.mp4
midroll_video=VID_20230206_114248.mp4
#tailtime=20
tailtime=$(ffprobe -v 0 -select_streams v:0 -show_entries format=duration -of default=nw=1:nk=1 "$endboard")
midroll_edit_value=25
#WID=$(ffprobe -v 0 -select_streams v:0 -show_entries stream=width -of default=nw=1:nk=1 "$mainvideo")
#HEI=$(ffprobe -v 0 -select_streams v:0 -show_entries stream=height -of default=nw=1:nk=1 "$mainvideo")
#SAR=$(ffprobe -v 0 -select_streams v:0 -show_entries stream=sample_aspect_ratio -of default=nw=1:nk=1 "$mainvideo")
#if [ "$SAR" = "N/A" ]; then SAR=1; fi
FPS=$(ffprobe -v 0 -select_streams v:0 -show_entries stream=r_frame_rate -of default=nw=1:nk=1 "$mainvideo")
TBN=$(ffprobe -v 0 -select_streams v:0 -show_entries stream=time_base -of default=nw=1:nk=1 "$mainvideo")
TBN=${TBN#*/}
#FMT=$(ffprobe -v 0 -select_streams v:0 -show_entries stream=pix_fmt -of default=nw=1:nk=1 "$mainvideo")
echo $WID $HEI $SAR $FPS $TBN $FMT
#CHL=$(ffprobe -v 0 -select_streams a:0 -show_entries stream=channel_layout -of default=nw=1:nk=1 "$main")
#SRA=$(ffprobe -v 0 -select_streams a:0 -show_entries stream=sample_rate -of default=nw=1:nk=1 "$main")
#echo $CHL $SRA
# Trim the main part of mainvideo
ffmpeg -hide_banner -y -i "$mainvideo" -c copy \
-f segment -segment_times $(($midroll_edit_value-5)),$(($midroll_edit_value+5)) \
-reset_timestamps 1 %d.mp4
# Insert mid
dura_0mp4=$(ffprobe -v error -show_entries format=duration -of default=nw=1:nk=1 0.mp4)
m=$( echo "$midroll_edit_value - $dura_0mp4" | bc -l )
ffmpeg -hide_banner -y -i 1.mp4 -i $midroll_video -filter_complex "
[0:v]trim=0:${m}[v0];
[0:v]trim=${m},setpts=PTS-STARTPTS[v2];
[0:a]atrim=0:${m}[a0];
[0:a]atrim=${m},asetpts=PTS-STARTPTS[a2];
[v0][a0][1:v][1:a][v2][a2]concat=n=3:v=1:a=1" \
-r $FPS \
-video_track_timescale $TBN \
1_mid.mp4
# Time calculations to define cut point
dura_2mp4=$(ffprobe -v error -show_entries format=duration -of default=nw=1:nk=1 2.mp4)
endboard_cut_point=$(echo "$dura_2mp4-$tailtime" | bc -l)
ffmpeg -hide_banner -y -i 2.mp4 -t $endboard_cut_point -c copy part2.mp4
# Trim the tail of mainvideo and overlay it onto endboard
echo "" > fcs.txt
PAD="[1:v]"
NUM=25
SCA=$((3*$NUM))
for (( i=0; i<$NUM; i++ )); do
echo "[0:v]trim=start_frame=$i:end_frame=$(($i+1)),setpts=PTS-STARTPTS+$i*N,scale=iw-iw*$i/$SCA:ih-ih*$i/$SCA[f$i];" >> fcs.txt
echo "$PAD[f$i]overlay=enable='eq(n,$i)'[o$i];" >> fcs.txt
PAD="[o$i]"
done
echo "[0:v]trim=1,setpts=PTS-STARTPTS+1/TB,scale=iw-iw*$i/$SCA:ih-ih*$i/$SCA[f$i];" >> fcs.txt
echo "$PAD[f$i]overlay=enable='gte(t,1)'" >> fcs.txt
ffmpeg -hide_banner -y \
-ss $endboard_cut_point -i 2.mp4 \
-i $endboard \
-filter_complex_script fcs.txt \
-r $FPS \
-video_track_timescale $TBN \
pip.mp4
# Pass all parts through the concat demuxer
[ -f filelist.txt ] && rm filelist.txt
for f in $intro 0.mp4 1_mid.mp4 part2.mp4 pip.mp4; do echo "file '$PWD/$f'" >> filelist.txt; done
ffmpeg -hide_banner -y -f concat -safe 0 -i filelist.txt -c copy TEST_FILE.mp4
mpv TEST_FILE.mp4
If you cut video by this code and check:
ffmpeg -i $mainvideo -ss 30 -c copy part2.mp4
ffprobe -select_streams v:0 -show_entries packet=pts_time,flags \
-of csv=print_section=0 "part2.mp4" | awk -F',' '/K/ {print $1}'
you can see something like this start:0.016000 and keyframes didn't starts from 0.000000, starts from 2.266016, for example. Because first keyframe is before beginning. If you concat that file without encoding you can get desync.