I have some game clips from Nvidia shadow play that I like to casually shorten and / or turn them into webms or keep them as mp4s. I use the same ffmpeg line for them. I do slightly change the line because of the input file, start time, and output file.
How could I set up something like a batch file (I was thinking maybe node as well) where it just asks for the input file, start time, and output file?
The current ffmpeg command line I use is like this:
ffmpeg -i desktop.mp4 -ss 00:01:50 -b 900000 -vf scale=640:trunc(ow/a/2)*2 output.webm
You can prompt for user input using the following pattern:
SET /P FILENAME=Enter Filename:
ECHO USER ENTERED %FILENAME%
So with your code you'd setup your 3 variables then use:
ffmpeg -i "%INFILE%" -ss %STARTTIME% -b 900000 -vf scale=640:trunc(ow/a/2)*2 "%OUTFILE%"
Related
iam trying to make action on events in FFMPEG.
For example: ffmpeg -i http://domain/index.m3u8 -c copy -f segment -strftime 1 -segment_time 10 %Y-%m-%d-%H-%M-%S.mp4
FFMPEG take live stream, cut by slices and create files. I want to run a script do_with_file.sh after every slice created, without ffmpeg pausing.
Is there any option in ffmpeg to make it?
Ofcource, i can take stdout from ffmpeg and looking for "segment" text:
ffmpeg ....mp4 | grep 'segment #' | do_with_file.sh
First of all, info line about "segment" showed in stdout, before file was created.
It is not working, if i want run ffmpeg in background.
And in my mind, it is not geek way :)
P.S. English is not my native language, sorry for mistakes.
You can ask ffmpeg to tell you when a segment is finished recording:
-loglevel verbose
With this option you'll get the event you're looking for:
[segment # 0x0f0f0f0f0f0f] segment:'filename.ext' count:N ended
But, if you're prefer a "geek" way, you may try inotifywait:
while segment=$(inotifywait --quiet --event close_write --format %w%f path/to/dir); do
do_with_file $segment
done
On Windows I have a simple batch file which I drop video files onto to convert to webms, it saves a lot of time as I prefer to just use the same configuration and don't care much about the names.
#echo off
echo.
ffmpeg -i %1 -c:v libvpx-vp9 -quality good -cpu-used 2 -b:v 5000k -qmin 15 -qmax 45 -maxrate 500k -bufsize 1500k -framerate 60 -threads 8 -vf scale=-1:1080 -c:a libvorbis -b:a 192k -f webm %1.webm
pause
I know the .bat file won't really work in Linux (I'm on pop os so pretty much Ubuntu) so with the other lines removed and the %1 changed to $1 it works. It won't do anything if I try dragging and dropping a video file onto it though.
I can type sudo then drag and drop the .sh file followed by a video into a terminal and press enter and it will have the same effect as dragging a video file onto a bat file in Windows.
Is there a way to recreate dragging and dropping a file directly onto another and it executing in Linux or is the terminal the only way?
So ... for me something like this works:
Create a desktop file, e.g. shell.desktop. That needs to look something like this:
[Desktop Entry]
Name=simple
Exec=/bin/bash -c "echo %F; sleep 10" # obviously this is where your ffmpeg incantation goes
Terminal=true
Type=Application
Icon=gvim
Categories=Utility;
StartupNotify=false
That is now a drop target - voila. Mine just echoes back the path of the file dropped.
I'm interested to try out Termux command line on Android to record microphone audio to mp3. I've tried running different commands but without much effect. Can anyone pinpoint correct example command to start recording of the microphone, to mp3, at a default location, for example downloads folder? (This is on Android Oreo)
termux-microphone-record
-d Start recording w/ defaults
-f Start recording to specific file
-l Start recording w/ specified limit (in seconds, unlimited for 0)
-e Start recording w/ specified encoder (aac, amr_wb, amr_nb)
-b Start recording w/ specified bitrate (in kbps)
-r Start recording w/ specified sampling rate (in Hz)
-c Start recording w/ specified channel count (1, 2, ...)
-i Get info about current recording
-q Quits recording
from https://wiki.termux.com/wiki/Termux-microphone-record
Termux does not (yet) appear to support recording directly to mp3 format. To get an mp3, you'll need to convert your recording using ffmpeg.
AWR Wide format has good quality for speech recording.
# Begin recording
termux-microphone-record -e awr_wide -f filename.amr
# Stop recording
termux-microphone-record -q
# Convert to mp3
ffmpeg -i filename.amr filename.mp3
The Below command will record for 10 sec and save your file.mp3 in your termux Home directory.
termux-microphone-record -d -f filename.mp3 -l 10
You need to install 1. the app for Android Termux.Api (i did from F-droid) and 2. the linux package (pkg install) 3. i had to give permission to microphone for termux.Api
after it records and stops automatically with -l.
I've a some hundreds of video files in a folder structure. All of them have video and audio streams, but some of them don't have any sound, despite having an audio stream. Is there a way to find out those files without having to resort to opening each file individually.
Most ways I know only check if there is an audio stream.
Thanks.
You can run the following in batch mode:
ffmpeg -hide_banner -i file.mp4 -af volumedetect -vn -f null - 2>&1 | grep mean_volume
The output for each file will be of the form
[Parsed_volumedetect_0 # 0000000002b1e800] mean_volume: -17.2 dB
Perfect digital silence will have a value of -91 dB, but anything below, say, -40 dB is probably just tape noise. Test and verify a few inputs manually and set a value.
Hi I a have a question I have openCV and ffmpeg on the Raspberry Pi and I am trying to stream live video from the raspberry pi. At the moment I have the output output of openCV saving as a .avi file and I have a command for ffmpeg
ffmpeg -i out.avi -hls_segment_filename '%03d.ts' stream.m3u8
This Command take the output creates the playlist(.m3u8) and the segments(.ts).
At present I have openCV programmed in C++ (this can not change) I have an executable programmed from this and I have both the executable C++ and the above ffmpeg in a Bash Script.
#!/bin/bash
while true; do
./OpenCV
ffmpeg -i out.avi -hls_segment_filename '%03d.ts' stream.m3u8
done
This does allow me to stream the processed openCV video my issue is as the Bash script is in a while loop it keeps resetting the playlist and the .ts files, so i have to constantly press play on the client connection.
Is there anyway around this?
I tried including a variable that would increment every loop but if i replace '%03d' with this i get an error.
If you insist on using your program (OpenCV) and ffmpeg in a loop then you can specify the initial hls sequence number for stream.m3u8 using start_number. Something like this:
... as before ...
ffmpeg -i out.avi -hls_segment_filename '%03d.ts' --start_number $I stream.m3u8
where I is a variable that you have to increment each time the loop runs.
But this approach is very fragile and will probably result in an incorrect stream because it assumes that ffmpeg will produce only a single segment but in reality it will probably produce multiple segments.
A much better approach is to run OpenCV and ffmpeg in parallel and make them talk to each other. By doing so there will be no need to write to a temporary file out.avi and run OpenCV and ffmpeg in sequences and keep the media sequences synchronized.
I think you can hack it like this. Note that you may need to change OpenCV so that it writes constantly to out.avi and does not return after a while:
./OpenCV &
tail -n +0 -f out.avi | ffmpeg -i pipe:0 -hls_segment_filename '%03d.ts' stream.m3u8
A better approach is change your program to write to stdout or to a named pipe and run it like so:
./OpenCV | ffmpeg -i pipe:0 -hls_segment_filename '%03d.ts' stream.m3u8