Converting video formats

Dafydd · 09-28-2012, 03:56 PM

Using the script below, I have converted the format of 20 videos. Took 119 hours 14 minutes.

Code:

#! /bin/bash
## convert video for use on website.

for filename in *.ogg;
	do
	  if [ -f "$filename" ]
    then
		 outfile=${filename%.*} ## strip off the extension
			ffmpeg -i "$filename" "$outfile".webm
	  fi
	done

I still have the 'mp4' format and the 'swf' format to go.

Is there a faster way to do this? Something simular to this?

Code:

 
ffmpeg -i "$filename" -o "$outfile".mp4 -o "$outfile".swf

OR!!!

Even a 10% reduction in run time over the 2 formats would mount to a whole day.

Dave

Dman58 · 09-28-2012, 04:31 PM

Overclock your machine, upgrade your cpu & ram. Lol

Maybe since they have been converted already you can use those new files to convert to the next desired formats. Instead of converting them all over to a new format.

Does this make sense?

dugan · 09-28-2012, 05:01 PM

Are you just changing the container format? You can add the -acodec copy and -vcodec copy flags to ffmpeg, and it won't recompress the audio and video.

porphyry5 · 09-28-2012, 05:08 PM

Quote:

Originally Posted by Dafydd

Using the script below, I have converted the format of 20 videos. Took 119 hours 14 minutes.

Code:

#! /bin/bash
## convert video for use on website.

for filename in *.ogg;
	do
	  if [ -f "$filename" ]
    then
		 outfile=${filename%.*} ## strip off the extension
			ffmpeg -i "$filename" "$outfile".webm
	  fi
	done

I still have the 'mp4' format and the 'swf' format to go.

Is there a faster way to do this? Something simular to this?

Code:

 
ffmpeg -i "$filename" -o "$outfile".mp4 -o "$outfile".swf

OR!!!

Even a 10% reduction in run time over the 2 formats would mount to a whole day.

Dave

Assuming there are no directory names ending in '.ogg', you could rewrite it as

Code:

for fl in *.ogg; do ffmpeg -i "$fl" -o "${fl%.*}".webm; done

but it would make no appreciable difference in the time

I've never done web videos myself, nor used ffmpeg, but I do a lot of dvd encoding with both mencoder and HandBrake. They both average about 2 hours per dvd for a 2 pass process. Is ffmpeg perhaps being thrown off by your using a music encoding suffix (ogg) for files that actually are videos? As in trying every format that it will handle to determine what your ogg files actually are. What does its log file have to say?

TobiSGD · 09-28-2012, 05:13 PM

Quote:

Originally Posted by porphyry5

Is ffmpeg perhaps being thrown off by your using a music encoding suffix (ogg) for files that actually are videos? As in trying every format that it will handle to determine what your ogg files actually are. What does its log file have to say?

Ogg is not a music encoding format, it is a container format, like avi is. The actual audio codec is named Vorbis, the video codec is name Theora: http://www.theora.org/faq/#13

porphyry5 · 09-29-2012, 10:32 AM

Quote:

Originally Posted by TobiSGD

Ogg is not a music encoding format, it is a container format, like avi is. The actual audio codec is named Vorbis, the video codec is name Theora: http://www.theora.org/faq/#13

Thank you for that clarification, I've only ever encountered ogg music files before.

Dafydd · 09-30-2012, 06:14 PM

Quote:

Originally Posted by dugan

Are you just changing the container format? You can add the -acodec copy and -vcodec copy flags to ffmpeg, and it won't recompress the audio and video.

You have ask a question for which I don't have an answer.

I used only the 'ogg' in a tread recently posted. Deeper research has produced this information

HTML Supported.

Code:

Browser			MP4	WebM	Ogg
___________________________________________________
Internet Explorer 9	YES	NO		NO
Firefox 4.0		NO	YES		YES
Google Chrome 6		YES	YES		YES
Apple Safari 5		YES	NO		NO
Opera 10.6		NO	YES		YES

I was trying to get a modified version of this code snippet from W3School to work.

Code:

<!DOCTYPE html>
<html>
<body>

<video width="640" height="480" controls="controls" autoplay="autoplay">
  <source src="movie.ogg" type="video/ogg">
  <source src="movie.mp4" type="video/mp4">
  <source src="movie.webm" type="video/webm">
<!-- Object tag deleted -->
</video>

</body>
</html>

Will 'iPad' accept one of these containers?
This VIDEO stuff is a new learning curve for me. Thanks for you help and feed back.
Dave

dugan · 09-30-2012, 07:26 PM

You're encoding for the web? Don't even bother with ogg containers or Theora video. There is no browser whose currently supported version supports theora video in an ogg container, but not vp8 video in a webm container.

Just do one patent-encumbered h264/mp3/mp4 encode and one patent-unencumbered vp8/vorbis/webm encode. Something like this should cover you:

Code:

ffmpeg -i input.avi -vcodec libx264 -acodec mp3 movie.mp4
ffmpeg -i input.avi -vcodec libvpx_vp8 -acodec vorbis movie.webm

This combination will work in every current browser.

Code:

<video>
  <source src="movie.mp4" type="video/mp4">
  <source src="movie.webm" type="video/webm">
</video>

BenCollver · 09-30-2012, 09:56 PM

Quote:

Originally Posted by porphyry5

Assuming there are no directory names ending in '.ogg', you could rewrite it as

Code:

for fl in *.ogg; do ffmpeg -i "$fl" -o "${fl%.*}".webm; done

but it would make no appreciable difference in the time

Assuming you have a processor with 2 cores, you could speed it up by transcoding 2 videos in parallel. Off the cuff, it could look like the following:

Code:

cores=2
i=0
for fl in *.ogg
do
    ffmpeg -i "$fl" -o "${fl%.*}.webm" >/dev/null 2>&1 &
    i=$(($i + 1))
    if [ $i -ge $cores ]
    then
        wait
        i=0
    fi
done

-Ben

porphyry5 · 10-01-2012, 06:40 AM

Quote:

Originally Posted by BenCollver2

Assuming you have a processor with 2 cores, you could speed it up by transcoding 2 videos in parallel. Off the cuff, it could look like the following:

Code:

cores=2
i=0
for fl in *.ogg
do
    ffmpeg -i "$fl" -o "${fl%.*}.webm" >/dev/null 2>&1 &
    i=$(($i + 1))
    if [ $i -ge $cores ]
    then
        wait
        i=0
    fi
done

-Ben

Thank you, I'd just naively assumed multi-core processors used their capabilities automatically. And presumably this technique can be adapted to any number of cores, for any lengthy process that has a series of inputs.

BenCollver · 10-01-2012, 07:33 AM

Quote:

Originally Posted by porphyry5

Thank you, I'd just naively assumed multi-core processors used their capabilities automatically. And presumably this technique can be adapted to any number of cores, for any lengthy process that has a series of inputs.

You are welcome. If you are careful, you can use "xargs -P 2" and get a little better performance than my first example.

Code:

$ cat >transcode.sh <<__EOF__
#!/bin/sh
fl="$1"
ffmpeg -i "$fl" -o "${fl%.*}.webm"
exit 0
__EOF__
$ chmod +x transcode.sh
$ find . -type f -name '*.ogg' -print0 | xargs -0 -n 1 -P 2 ./transcode.sh

porphyry5 · 10-02-2012, 07:14 PM

Quote:

Originally Posted by BenCollver2

You are welcome. If you are careful, you can use "xargs -P 2" and get a little better performance than my first example.

Code:

$ cat >transcode.sh <<__EOF__
#!/bin/sh
fl="$1"
ffmpeg -i "$fl" -o "${fl%.*}.webm"
exit 0
__EOF__
$ chmod +x transcode.sh
$ find . -type f -name '*.ogg' -print0 | xargs -0 -n 1 -P 2 ./transcode.sh

And thank you for these techniques too. So one can indicate to an app that it should read from standard input until a designated line. But is there a particular reason why you feed the output of find through a pipe to xargs, rather than use find's -exec option?

BenCollver · 10-02-2012, 11:37 PM

The "xargs -P 2" option runs 2 commands in parallel. The "find -exec" option doesn't.

-Ben

Dafydd · 10-03-2012, 12:13 AM

Thanks to all for the informaton, instructions and patients. I have put the script to running. I will be back after Sunday, Renaissance Festival is starting this weekend and I have things to do to get ready.

porphyry5 · 10-04-2012, 03:39 PM

Quote:

Originally Posted by BenCollver2

The "xargs -P 2" option runs 2 commands in parallel. The "find -exec" option doesn't.

-Ben

I might be misreading the 'find' manpage, but I thought that the

Code:

 -exec command {} +

variant would allow that simultaneous processing, because it will pass multiple files to a single evocation of 'command', while

Code:

 -exec command {} \;

does them sequentially, just one file per evocation of 'command'.