while fixing the issue:
http://code.google.com/p/sympy/issues/detail?id=1385
I recorded my terminal and I was also commenting what I was doing, so
that you can get an idea how to work with git, squash patches, switch
branches etc. And also how to write and run tests, doctests with
sympy.
It took me 23 minutes to record the video (e.g. to fix the issue).
Now it's more than an hour and I still can get the final video out.
Here are two files:
http://sage.math.washington.edu/home/ondrej/scratch/video/
audio and video, the just have to be combined together. Can someone
please help me with that? I am afraid it is not possible to do it on
my Ubuntu...
It took me about a weekend to try all video editing programs I found
on linux to realize that I am not able to force any of them to
actually record a terminal+sound, combine it together and not to loose
quality. So I wrote this script:
http://github.com/certik/record
that uses "recordmydesktop" that works pretty well, but even though it
can also record sound, in practice it can't (it's very very low
quality), so this python script also uses gstreamer to record sound,
it calls (alsasrc ! audioconvert ! flacenc ! filesink location=%s)
and that works. So I endup with two files, now I need to combine them
together. On smaller files mencoder works quite well, but on the above
files mencoder skips lots of frames and the result is unusable
(mplayer can play it properly). The trick is to convert the flac to
ogg (I used quality 10, so given my low quality microphones, this
doesn't loose anyting), then mencoder can produce a video with sound,
but the sound is about 0.5s (or 1s) delayed. Mencoder has an option
"-delay", but it doesn't seem to work (I didn't notice any difference
when played with mplayer afterwards).
So it's very sad, but I have to admit I can't even put the audio and
video together so that it is in sync. And originally I wanted to
record some short intro with my camera and then attach those two
things together --- well, this is totally impossible. I managed to get
something working with cinelerra, but it can't read my sound and video
for some reason, that my script above produced. It works with
sound+video coming from my camera though (e.g. allows me to do some
fancy transformations etc.). But since it doesn't work with camera and
video from the terminal, it's totally useless.
I think the problem may be in the format I am using for videos and/or
audios. But I am giving up. If anyone on this list knows how to get
this working on linux, I would be very grateful.
Ondrej
Also the sound volume needs to be increased, the audio is very quiet.
Ondrej
It gives me some duplicate frames several times per second:
...
Pos: 235.2s 3531f ( 0%) 21.11fps Trem: 0min 0mb A-V:0.024 [1027:143]
1 duplicate frame(s)!
Pos: 235.8s 3540f ( 0%) 21.11fps Trem: 0min 0mb A-V:0.022 [1025:143]
1 duplicate frame(s)!
Pos: 236.5s 3551f ( 0%) 21.11fps Trem: 0min 0mb A-V:0.021 [1026:143]
1 duplicate frame(s)!
Pos: 238.9s 3587f ( 0%) 21.11fps Trem: 0min 0mb A-V:0.022 [1022:143]
1 duplicate frame(s)!
Pos: 239.1s 3589f ( 0%) 21.11fps Trem: 0min 0mb A-V:0.024 [1021:143]
1 duplicate frame(s)!
Pos: 239.2s 3591f ( 0%) 21.11fps Trem: 0min 0mb A-V:0.025 [1021:143]
1 duplicate frame(s)!
Pos: 239.8s 3600f ( 0%) 21.11fps Trem: 0min 0mb A-V:0.020 [1019:143]
...
And when played with mplayer, it looks like the attached screenshot (I
scaled the image down, but you get the idea).
I suspect my mencoder doesn't work properly. My version is:
MEncoder 2:1.0~rc2-0ubuntu19+medibuntu1 (C) 2000-2007 MPlayer Team
which version do you have?
Ondrej
I tried that one. It can't read the .ogv file (ogg theora file) on my
system (it's not among supported formats), so I set the open dialog to
all files, opened it anyways, but it shows it, but it can't play it...
But I just found, that you can find BSD licensed decoders and encoders here:
and they are in most distributions now anyways. I didn't find python
wrappers for it yet, but I'll probably just write cython wrappers
myself. Then hopefully I should be able to read the stream in Python
(possibly playback with pyglet) and join the audio myself in a simple
python script.
Being able to read, decode and encode video from Python itself would
help too, I actually need to create some nice videos that result from
my animations, so far I use mencoder, but having a small nice library
in Python would be better.
I am a little worried it will not work, because this program (that
should work for vorbis+theora) utterly fails:
$ ogmmerge -o out.ogm audio.ogg video.ogv
Using OGG/OGM demultiplexer for audio.ogg.
+-> Using Vorbis audio output module for stream 1.
Using OGG/OGM demultiplexer for video.ogv.
Error: the reader for video.ogv did not produce a header page.
It may be that the recorddesktop produces nonworking theora file.
Ondrej
animations -> simulations.
O.
So you have the same version as I have.
>
>
> I get duplicate frames too, but I think the output file is OK. Run the
> script for about a minute and view the output file.
I did exactly that and you can see the output is totally garbled.
Ondrej
Send me your public ssh key offlist, I'll create you an account on my
server, so that you can upload it.
Thanks,
Ondrej
Upload it to youtube please.
Thanks a lot,
Ondrej
You already figured out how to create the key, so to send a file, just do:
scp your_file.mp4 bro...@li21-74.members.linode.com:
and that's it.
Ondrej
Vinzent converted the video here:
http://sage.math.washington.edu/home/ondrej/scratch/video/
I think he said the quality is not so high, but it works.
Unfortunately, I can't play it with any player on my ubuntu.
Let me know if it works for you.
Ondrej
The mp4 file works fine for me using VLC on OS X, but it doesn't work
using any QuickTime based player.
Thanks!
Cheers,
Tim.
---
Tim Lahey
PhD Candidate, Systems Design Engineering
University of Waterloo
http://www.linkedin.com/in/timlahey
Alan, yes, please upload your video the same way.
I can't play either of them, but I am going to uplaod it to vimeo and
we'll see. If they can play it, that's all that matters.
Thanks very much for the help,
Ondrej
I uploaded it to vimeo:
but it says "This video will start converting in a moment" and it say
that for 5 hours... So if you want to try and give it a shot too, just
go ahead.
Ondrej
It took them 10 hours to process it. Yes, the sound is too quiet, but
the video itself is excellent, very nice quality of the images. So
next time I'll fix the sound too and it will be usable.
Ondrej