Re: [FFmpeg-user] Autolaunch FFMPEG when network traffic detected
On Mon, Jul 23, 2018 at 9:34 AM, yannickb
wrote:
> I wish I could monitor each of those ports and launch the corresponding
> command automatically as soon as the incoming stream is detected (and stop
> when the stream ends)
>
>
>
One advantage of using a Linux distro like Debian as you mention, is that
you can run each instance of a program in a separate account i.e. user1,
user2 etc for each instance of ffmpeg in this example.
That way its easier to keep of udp ports open and running processes by
userid / group id. You can also use bash lock files per user to maintain
only one open ffmpeg instance.
UDP has a few quirks in my experience. You can use this command to see the
processes using udp.
lsof -i udp
For keeping track of open ffmpeg processes and UDP ports, here is a quick
few lines of bash code ...
#!/bin/bash
uid=`id -u $shell_user `
gid=`id -g $shell_user `
process_count=`ps -ef --Group $gid --User $uid | grep -v grep | grep ffmpeg
| wc -l `
if [[ $process_count -gt 0 ]]; then
printf "\n\nffmpeg is in progress\n\n"
else
printf "\n\nffmpeg is not in progress\n\n"
fi
port=999
serviceIsRunning=false
openPorts=$(/bin/netstat -tulpn | grep -vE '^Active|Proto' | awk '{ print
$4}' | awk -F: '{print $NF}' | sed '/^$/d' | sort -u)
for openPort in $openPorts
do
if [[ "$port" == "$openPort" ]]; then
serviceIsRunning=true
echo "service is running."
break
else
printf "\n\nfound open UDP: $openPort that is not port: $port "
fi
done
if [ $serviceIsRunning == false ]
then
echo "service is not running."
fi
Best regards,
Robert
___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user
To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] issue with monitoring script
On Fri, Jul 20, 2018 at 5:37 PM, Pedro Daniel Costa < portaln...@outlook.com.br> wrote: > Hi guys > > I need some guidance > > I have setup a crontab > > To check every minute if a live transcoding channel has stopped. > > The script gets pulled via crontab and I can see the instance starting up, > but then after couple of seconds it gets shutdown, > > > If I run the script manually it starts both ffserver and ffmpeg input > couple of seconds later and it run for hours/days with no problem.. > > If I close the terminal where I have executed the script and let the > crontab execute it, it start and shutdown couple of seconds after > > I have correct permissions on the script files and config files.. > > Here is the ouput of the script > > Can someone help me on this as I want to make a monitoring script to check > all the live transmission transcodings being used > > > And if it crashes or stops I want the crontab to init ffserver and ffmpeg > shortly after again > > > Here is the output of the script, can someone please tell me what I am > doing wrong? I use the same similar script to my middlware application to > check the live multicast transmission and tunners setup from the PCI > devices.. > > > ## > # check channel transcoding > # ffserver channel1.sh > ## > > #!/bin/bash > > > if ps x |grep -v grep | grep channel1.cfg ; > then > echo "FFserver runing " > else > echo "FFserver down" > ffserver -d -f channel1.cfg > fi > sleep 10 > if ps x |grep -v grep | grep channel1.ffm ; > then > echo "Transcoder running " > else > echo "transcoder down" > ffmpeg -i udp://@239.106.3.0:4002 http://170.80.123.234:8890/channel1.ffm > fi > > > > > Any hints or tips on how to setup a correct check status script will be > appreciated. > > > > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". Try editing your crontab so stdout / stderr go to a log file, to see the specific error - see example below. Also /var/log etc may have cron logs. Another thing is I would use an absolute path to ffmpeg, you can find the correct path by the command 'which ffmpeg' in the shell . You may also try putting the process in the background with an '&' at the end of your commands. sh channel1.sh > /home/username/xxx.log 2>&1 & You can also get some extra logging via these lines at the top of your file (after '#!/bin/bash') , and an exit on the error line. set -e set -x set -o errexit Best regards, Robert ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] kmsgrab on Intel 8th generation aka h/w accelerated screen capture
On Sun, Jun 24, 2018 at 7:10 PM, Kai Hendry wrote: > Hi ffmpeg users, > > Sorry I initially posted my question here: > https://www.reddit.com/r/ffmpeg/comments/8tgejw/kmsgrab_on_intel_8th_ > generation_aka_hw/ > > Via https://trac.ffmpeg.org/wiki/Hardware/VAAPI#ScreenCapture I > discovered kmsgrab which astonishingly doesn't seem to drop frames or > over heat my T480s when recording my Xorg display! > Unlike when I record with x11grab & VAAPI: > https://github.com/kaihendry/recordmydesktop2.0/blob/vaapi/x11capture#L55 > Also see: http://lists.ffmpeg.org/pipermail/ffmpeg-user/2018- > March/039227.html > Same issue on ffmpeg 4.x btw: > https://s.natalian.org/2018-06-25/152978.mp4.log > > However since kmsgrab requires CAP_SYS_ADMIN as it points out to its > friendly documentation > https://www.ffmpeg.org/ffmpeg-devices.html#kmsgrab I believe I need to > run it with `sudo`. However when I run it as `sudo` I hit a issue > accessing my pulseaudio microphone. > https://stackoverflow.com/questions/51007952/recording- > pulse-audio-as-cap-sys-admin-with-ffmpeg > > Any tips how to get this all working? > https://github.com/kaihendry/recordmydesktop2.0/blob/kms/x11capture#L54 > is my branch of kmsgrab based screen capture cli. > > > Kind regards, > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". That command worked for me though of course without your external dep. Do you mean run ffmpeg as sudo? You need the proper entry in /etc/sudoers and the ffmpeg command per sudo rules must have an absolute path like /bin/ffmpeg. Regards, Robert ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] x11grab segfault when display is closed
On Fri, Jun 22, 2018 at 9:03 AM, Matt Garman wrote: > I'm using a static git build, version N-46353-g40b7e6071-static[1] on > CentOS 7.3. This server is used as a remote login launch point (or "jump > box"). What we'd like to do, for security/auditing purposes, is create > videos of user sessions. > > I cannot answer the ffmpeg specific question though as a CentOS 7.x user I use Xvfb a lot. I have it doing important high volume prod stuff at my day job and its been a workhorse that has rarely been a problem. Seems like its possible to make videos via Xvfb. I just do screenshots myself. https://github.com/lightsofapollo/x-recorder Regards, Robert ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] Unable to stream SAP to remote pulseaudio
On Tue, Jun 19, 2018 at 1:58 PM, Antoine Vacher < antoine.vac...@tigre-bleu.net> wrote: > Any idea someone? > > Thanks, > > Antoine > > I suggest posting your question to the Linux Audio Users list (LAU) . There are daily email threads on that list about Pulse Audio and at least a few of the developers there use FFMPEG in their tools. I am no help myself as I do my signal processing outside of computers, I just process the captured audio and video via FFMPEG on Linux via Bash. Regards, Robert ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] How to make ProRes files with SMPTE DTV audio for 5.1
On Mon, Jun 18, 2018 at 4:33 PM, Mark Scott wrote: > Hello, > I'm very new to ffmpeg. I work at a production company and I need to make > Broadcast ProResHQ files with 5.1 audio plus Stereo on Ch.7&8. > > I haven't been able to figure out the command line inputs to get the > correct audio configuration. I'm attaching a screen grab of the inspector > windows for the Master file, and for the broadcast file that I have made > successfully with other software. The resulting audio needs to be as > follows: 24-bit Integer (little Endian) SMPTE DTV (L R C LFE LS RS LT RT) > 48kHz. Our master files are exported with 6 mono tracks and 1 stereo track. > > The following command line creates a ProResHQ file with Stereo audio. > > ffmpeg -i TEST.mov -c:v prores -profile:v 3 -acodec pcm_s24le > TEST_ProResHQ.mov > > > How can I change the command line to get the SMPTE DTV audio? > > > MASTER FILE: > > > > BROADCAST FILE WITH CORRECT AUDIO CONFIG: > > > > Thank you, > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". On Linux you can read the LTC from a wav file via the ltcdump command, no idea on other platforms. I am doing that right now actually. Change the output from mov to wav: ffmpeg -y -v error -nostdin -i my.mov -vn -c:a pcm_s24le -ar 48000 -ac 2 my.wav ltcdump -f 29.97 my.wav 2> /dev/null > my.ltc.txt I recently discussed timecode with ffmpeg here: http://ffmpeg.org/pipermail/ffmpeg-user/2018-June/040060.html ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] Audio mapping
On Thu, Jun 7, 2018 at 3:18 AM, Kasper Folman wrote: > > But what if I had a file with 4 mono tracks, and I wanted to merge those 4 > tracks into 2 stereo tracks. It does do what I’m asking it to, but again, > only the first audio track is enabled. > > ffmpeg -t 1 -i "/Volumes/Storage/_MoveToBacku > p/formatSamples/1x16-chAudioTest.mov" -filter_complex > "[0:0]copy,fps=25,setfield=prog,setdar=dar=16/9,setsar=sar=1 > /1[vOut];[0:1]pan=1c|c0=c0[a1];[0:2]pan=1c|c0=c0[a2];[0:3] > pan=1c|c0=c0[a3];[0:4]pan=1c|c0=c0[a4];[a1][a2]amerge=input > s=2[aOut0];[a3][a4]amerge=inputs=2[aOut1]" -c:a pcm_s24le -map [aOut0] > -map [aOut1] -map [vOut] -c:v prores_ks -r 25 -pix_fmt yuv422p10le > -profile:v 1 -field_order progressive -y -strict -2 > "/Users/kf/Desktop/testfile.mov" > I am using separate wav and mov files together however I am on Linux and do not use quicktime as a player. And I am on 3.3.4. Anyways using an alternative approach, a BWF works for me as a way to get 8 tracks into Ch1 / Ch2 audio of a mov file. Joining mono WAV files into a single BWF may or not work for you but thought I'd mention it. cuts.mov in this example is combined footage from two Zoom Q8 cameras, f8t.wav is a BWF produced by an 8 track Zoom F8. Only showing first pass of 2pass. ffmpeg -v error -i cuts.mov -i f8t.wav -pass 1 -passlogfile export.log -c:a pcm_s16le -c:v libx264 -profile:v main -level 4.0 -refs 1 -x264-params "keyint=8:b-pyramid=0:no-scenecut:nal-hrd=cbr:force-cfr=1" -r 3/1001 -s 1920x1080 -b:v: 24M -bufsize 24M -maxrate 24M -b:a 48k -muxrate 24M -sample_fmt s16 -ac 2 -ar 48000 -af "aresample=async=1:min_hard_comp=0.000100:first_pts=0" -crf 18 -metadata comment="Created with parseLTC.sh" -map 0:0 -map 1:0 -y export.mov Regards, Robert ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] FFMPEG output to append to a text file without overwriting the content
On Mon, Jun 4, 2018 at 6:36 PM, robertlazarski wrote: > > > On Mon, Jun 4, 2018 at 6:06 PM, Sana Tafleen > wrote: > > I would not expect ffmpeg to create a new process in that loop but I have > no experience with mpegts. I would try -stdin since I have seen unexpected > behavior without it in loops. > > Meant to say -nostdin . ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] FFMPEG output to append to a text file without overwriting the content
On Mon, Jun 4, 2018 at 6:06 PM, Sana Tafleen wrote: > Hello, > > I am sending a UDP stream to the destination and saving the contents of the > FFMPEG output to a text file. I run the ffmpeg command in a loop as > follows, > > > while : > do > echo `ffmpeg -hide_banner -f v4l2 -i /dev/video0 -c:v libx264 -f > mpegts tcp://ip:port -c:v libx264 /path/to/.mp4 2> out.txt -y` > done > > > > When I disconnect the cable connected the destination, the above ffmpeg > process stops and a new one starts oevrwriting the content of the output > that has been saved. And when I reconnect the cable, a new ffmpeg process > starts and its output is what is displayed in the out.txt file. > > I need the output of each ffmpeg process that runs to an output file. Can > anyone please suggest me a way to do it? > This part of the command, '2> out.txt ' says (a) redirect stderr and not stdout, to out.txt. And (b) overwrite the previous contents. You can append instead of overwrite by using '2>> out.txt ' . I would not expect ffmpeg to create a new process in that loop but I have no experience with mpegts. I would try -stdin since I have seen unexpected behavior without it in loops. A long shot would be using nohup, if its a hangup of some sort. I would try that if I was still stuck. Then strace on the command to see why it was creating the extra process. Kind regards, Robert > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] Using segment and concat with multiple videos containing LTC audio on CH1
On Fri, May 4, 2018 at 8:31 AM, robertlazarski
wrote:
> Hello all, first post.
>
> I am trying to segment mov videos from multiple Zoom Q8 cameras with
> exactly the same specs, by using the Linux command ltcdump, bash and
> ffmpeg.
>
To answer my own question and after many weeks later of experimentation,
using segment was the wrong tool for the job.
Each LTC audio frame at 3/1001 in this case is 1600/48000 or around
.034 seconds and the closest I could come to that with segment while still
being able to read the LTC, was '-force_key_frames
"expr:gte(t,n_forced*0.07)" ' , which usually meant segments of 64ms but
often ranged to 85ms ... didn't work for me.
I had much better luck calculating the -ss and -to times using the sample
position of the LTC. I was still somehow off by about a frame or .034
seconds on each video transition between cameras as the cut accuracy wasn't
good enough. The vocals of the result video clearly had drift.
See below, using 2pass and lots of params because debugging the transcoded
videos is easier the closer they match the source. 16 bit depth instead of
24 on source videos for debugging purposes (long story).
ffmpeg -y -v error -nostdin -i ../q81.mov -pass 1 -passlogfile q81.log -ss
5.506292 -to 25.5264 -c:a pcm_s16le -c:v libx264 -profile:v baseline -level
3.0 -x264-params keyint=8:b-pyramid=0:no-scenecut:nal-hrd=cbr:force-cfr=1
-refs 1 -r 3/1001 -s 1920x1080 -b:v: 24M -bufsize 24M -maxrate 24M -b:a
48k -muxrate 24M -sample_fmt s16 -ac 2 -ar 48000 -af
aresample=async=1:min_hard_comp=0.000100:first_pts=0 -crf 18 q81_cut_0.mov
ffmpeg -y -v error -nostdin -i ../q81.mov -pass 2 -passlogfile q81.log -ss
5.506292 -to 25.5264 -c:a pcm_s16le -c:v libx264 -profile:v baseline -level
3.0 -x264-params keyint=8:b-pyramid=0:no-scenecut:nal-hrd=cbr:force-cfr=1
-refs 1 -r 3/1001 -s 1920x1080 -b:v: 24M -bufsize 24M -maxrate 24M -b:a
48k -muxrate 24M -sample_fmt s16 -ac 2 -ar 48000 -af
aresample=async=1:min_hard_comp=0.000100:first_pts=0 -crf 18 q81_cut_0.mov
So, 25.5264 - 5.506292 = 20.020108
However:
ffprobe -v error -show_entries format=duration -of
default=noprint_wrappers=1:nokey=1 q81_cut_0.mov
20.054000
Each video in the script below is off by the same amount. Strangely using
an offset of 0.033892 from 20.054000 - 20.020108 didn't work as it cut too
much though splitting the value in half at 0.017 works splendidly. By the
end of the video I was able to come under .002 of drift.
See below for the script, magicNumberOffset="0.017" is as described above.
Might be generally useful as an example how to use bash with ffmpeg.
#!/bin/bash
# This script generates a single video from parts of videos from 2 Zoom Q8
cameras
# by using ltc on ch1 of each video. The priority is as little ltc audio
drift as possible,
# script performance and extra transcoding are not an issue. Little audio
drift in this case
# means the start and end times match the source videos closely, with no
dropped video frames
# nor dropped ltc frames. I was able to get within 88/48000 or 0.00183
seconds by the end of the video,
# and while I have no immediate ideas how to do better I will continue to
try.
# The ltc comes from 2 Tentacle Sync units, that were 'jammed' to a
# Zoom F8. The main idea is to sync the final export.mov video with a BWF 8
track audio file generated on
# a Zoom F8, via the ffmpeg command to generate the export.mov file.
# This is possible by using the BWF timecode metadata, and ltc on ch1 of
each video.
# ch2 contains a scratch track.
# A fixed duration such as 20 seconds is set, indicating a camera
# transition. Videos are mov at 1920x1080 with ltc audio at 48KHZ / 16 bits
# usage:
# /home/myuser/input> ls
# f8.wav output parseLTC.sh q81.mov q82.mov
# /home/myuser/input> cd output/
# /home/myuser/input/output> sh ../parseLTC.sh
# clear any files from previous processing
rm -f *.txt
rm -f *cut*
rm -f *.csv
# read LTC from wav files, first camera is q81 and the second camera is q82
ffmpeg -y -v error -nostdin -i ../q81.mov -vn -c:a pcm_s16le -ar 48000 -ac
2 q81.wav
ffmpeg -y -v error -nostdin -i ../q82.mov -vn -c:a pcm_s16le -ar 48000 -ac
2 q82.wav
ltcdump -f 29.97 q81.wav 2> /dev/null > q81.ltc.txt
ltcdump -f 29.97 q82.wav 2> /dev/null > q82.ltc.txt
# some errors I found while playing option bingo with ffmpeg
validate() {
lastTimecodeInSegment=0
lastTimecodeInSegment=`tail -n 1 ltc.txt | awk '{ print $2 }' | sed
's/\(.*\):/\1./' `
# can sometimes receive unparsable dates
if [[ ${#lastTimecodeInSegment} -ne 11 ]]; then
lastTimecodeInSegment=0
printf "\nError: timecode number of chars is not correct: "
printf " %s" "${#lastTimecodeInSegment}"
return
fi
# can receive invalid time such as: 0680 00:13:61:10 |
22 1609
if ! date -d "$lastTimecod
[FFmpeg-user] Using segment and concat with multiple videos containing LTC audio on CH1
Hello all, first post.
I am trying to segment mov videos from multiple Zoom Q8 cameras with
exactly the same specs, by using the Linux command ltcdump, bash and
ffmpeg.
The LTC is being generated by a Tentacle Sync for each camera, that was
"jammed" to a Zoom F8 which is master audio I use to sync the final video
via LTC ... I have this working fine on each camera individually using
Ardour.
The goal is to concat the segments from these cameras into one video. The
main idea is the videos switch cameras at a fixed duration determined at
runtime by a simple constant like 20 seconds, determined by reading the
output from ltcdump on each video.
I have this very close to working. The last step is fixing an audio drift
problem caused during the transition from one camera to another. This drift
is because I cannot figure out how to exactly match an LTC audio frame to
each segment. I tried lots of numbers for the segment and was unsuccessful.
Here is an example of a transition and the problem I am trying to fix. Q81
is camera 1, Q82 is camera 2. I am missing frames 28 and 29 - the last .xx
digits are the frame and not milliseconds. And the LTC frame of somewhere
between 1597 and 1601 samples at 48KHZ does not match the segment exactly.
### ltcdump output from file q81_out280.mov.wav , last two .xx digits are
frames
#User bits Timecode |Pos. (samples)
#DISCONTINUITY
00:26:34:27 | 1221 2821
### ltcdump output from file q82_out305.mov.wav , last two .xx digits are
frames
#User bits Timecode |Pos. (samples)
#DISCONTINUITY
00:26:35:00 | 818 2418
What I believe I need to do is get the segment to match the LTC frame as
closely as possible. In these examples the segment I believe I need is 1600
/ 48000. However I tried and could not get the segments to match.
Here's my code to show what I am doing. Its about 300 lines.
#!/bin/bash
# This script generates a single video from parts of videos from 2 cameras
by using ltc.
# Videos are mov at 1280x740 with audio at 48KHZ and 16 bits
#
# usage:
# /home/myuser/input> ls
# output parseLTC.sh q81.mov q82.mov
# /home/myuser/input> cd output/
# /home/myuser/input/output> sh ../parseLTC.sh
# clear any files from previous processing
rm -f *.txt
rm -f *all*
rm -f *out*
# Trim first camera so the LTC frame starts as closely as possible to zero
# 00:26:15:00 |2 1615
ffmpeg -y -v error -ss 00:00:06.067697 -t 00:30:15.365 -i ../q81.mov -c
copy q81t.mov
mv q81t.mov q81.mov
mv ../q82.mov q82.mov
# debug
ffmpeg -y -v error -i q81.mov -vn -c:a pcm_s16le -ar 48000 -ac 2 q81.wav
ffmpeg -y -v error -i q82.mov -vn -c:a pcm_s16le -ar 48000 -ac 2 q82.wav
# split file by key frames in a .07 second duration, need a couple entries
that
# ltcdump can find in each file - ideally just a single entry.
# first Q8
ffmpeg -y -v error -fflags +genpts -i q81.mov -vsync 1 -c:a pcm_s16le -c:v
libx264 -refs 1 -x264opts b-pyramid=0 -r 3/1001 -threads 0 -s 1280x720
-b:v: 1024k -bufsize 1216k -maxrate 1280k -b:a 192k -sample_fmt s16 -ac 2
-ar 48000 -af "aresample=async=1:min_hard_comp=0.10:first_pts=0"
-segment_time 0.07 -segment_list q81t.ffcat -sc_threshold 0
-force_key_frames "expr:gte(t,n_forced*0.07)" -f segment -reset_timestamps
1 -crf 0 q81_out%03d.mov
# second Q8
ffmpeg -y -v error -fflags +genpts -i ../q82.mov -vsync 1 -c:a pcm_s16le
-c:v libx264 -refs 1 -x264opts b-pyramid=0 -r 3/1001 -threads 0 -s
1280x720 -b:v: 1024k -bufsize 1216k -maxrate 1280k -b:a 192k -sample_fmt
s16 -ac 2 -ar 48000 -af
"aresample=async=1:min_hard_comp=0.10:first_pts=0" -segment_time 0.07
-segment_list q82t.ffcat -sc_threshold 0 -force_key_frames
"expr:gte(t,n_forced*0.07)" -f segment -reset_timestamps 1 -crf 0
q82_out%03d.mov
# save the ffcat output to run again if need be while debugging
cp q81t.ffcat sav_q81t.ffcat
cp q82t.ffcat sav_q82t.ffcat
# some errors I found while playing option bingo with ffmpeg
validate() {
fileToProcess=''
fileToProcess=$1
epoch_secs_ltc_to_process=0
ltc_to_process_formatted=''
isLTCDumpOutputValid=0
found_stop_time=0
# test for shell error
printf "\nExecuting: ltcdump -f 29.97 "$fileToProcess" "
# adding frame rate removes this DISCONTINUITY
# 00:26:53:24 | 7224 8824
#DISCONTINUITY
ltcdump -f 29.97 "${fileToProcess}" > /dev/null 2>&1
if [ $? -ne 0 ] ; then
printf "\nError processing file %s\n" "${fileToProcess}"
printf "\nshell returned non zero value on validity check"
return
fi
printf "\nExecuting: ltcdump -f 29.97 "${fileToProcess}" 2>/dev/null |
grep -v '#' | grep ':' | wc -w "
hasLTC=0
hasLTC=`ltcdump -f 29.97 "${fileToProcess}" 2>/dev/null | grep -v '#' |
grep ':' | wc -w`
if [[ $hasLTC -eq 0 ]]; then
printf "\nError processing file %s\n" "${fileToProcess}"
printf "\ncannot parse timecode\n"
return
fi
rm -f ltc.txt
# find timecode in f
