Re: [FFmpeg-user] Delay between the first packet and last packed in the muxing queue

2014-12-08 Thread Matej Mailing
2014-12-07 13:33 GMT+01:00 Matej Mailing mail...@tam.si:
 2014-12-07 6:06 GMT+01:00 Roger Pack rogerdpa...@gmail.com:
 On Sat, Dec 6, 2014 at 5:58 AM, Matej Mailing mail...@tam.si wrote:

 Yes, that is the moment when the input drops due to some network
 issues, but it is back after a second or so.



 so what you want is an http connection that auto reconnects when the
 connection drops? I wonder if there's some timeout option that could be set
 on it [if it's timing out--it might not be...]
 what do you mean is back after a second or so ffmpeg doesn't seem to
 think it's back...


 Hi,

 this is exactly what I would like to have - an http connection that
 auto reconnects after the connection drop. With back after a second
 or so I mean that due to network issues any network input can be
 stopped for some time and then becomes available agan. In this case
 output show resume as normal.

 This is a very practical situation with any http input format.

 Thanks,
 Matej




 Since there is no
 synchronization between video and sound required, I would like that
 ffmpeg continues playing mp3 stream. In current situation those errors
 keep showing up all the time and the output is never normal again -
 I have to manually restart ffmpef process (until the next input's drop
 occurs...)

 I think this is a very common situation with all the input streams
 since they can drop for some time due to many possibly network issues
 and resuming when the stream is available again seems to be the option
 we want.

 Thanks,
 Matej

 2014-12-05 23:50 GMT+01:00 Roger Pack rogerdpa...@gmail.com:
  On Fri, Dec 5, 2014 at 12:20 AM, Matej Mailing mail...@tam.si wrote:
 
  Hi,
 
  when muxing the screen capture with the live mp3 stream, it sometimes
  happens that live mp3 stream becomes unavailable for a short moment
  (due to routing issues etc.) and when this happens, the error
  http://mp3.rtvslo.si/val202: Unknown error pops up.
 
 
 
  Sounds to me like the input is dropped at that point [?]
  ___
  ffmpeg-user mailing list
  ffmpeg-user@ffmpeg.org
  http://ffmpeg.org/mailman/listinfo/ffmpeg-user
 ___
 ffmpeg-user mailing list
 ffmpeg-user@ffmpeg.org
 http://ffmpeg.org/mailman/listinfo/ffmpeg-user

 ___
 ffmpeg-user mailing list
 ffmpeg-user@ffmpeg.org
 http://ffmpeg.org/mailman/listinfo/ffmpeg-user


Hi,

right now from my perspective it seems that the best option is to use
-shortest output option and have a cronjob checking if ffmpeg is
running and if it's not, restart the process ..
However this doesn't sound as a perfect solution in this case - how do
you guys deal with input streams failing in such cases?

Thanks!
___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user


Re: [FFmpeg-user] Delay between the first packet and last packed in the muxing queue

2014-12-08 Thread Sinan Alyuruk

On 12/08/2014 11:52 AM, Matej Mailing wrote:



Hi,

right now from my perspective it seems that the best option is to use
-shortest output option and have a cronjob checking if ffmpeg is
running and if it's not, restart the process ..
However this doesn't sound as a perfect solution in this case - how do
you guys deal with input streams failing in such cases?

Thanks!
___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user


I use an upstart script on centos6.6 with respawn option, but you may 
prefer any reliable supervisor daemon.


Regards,
Sinan
___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user


Re: [FFmpeg-user] concat loses 2 frames or freezes up video

2014-12-08 Thread S Andreason


 Never-the-less, trying to concat certain files that appear whole and 
intact with all I/P/B frames, does cause a couple problems, with several 
files, about 10% of the ones I tested.


 The input files used for these tests, can be found at:
  
http://seahorseCorral.org/videos/tests/20141107_4horses-scatter-cows_mini1355.mov 



After some needed sleep I thought of some new tests.

Loading #1 in avidemux looks correct. An I frame every 15 frames. Ah, 
the 2 frames Are there after all, but they are skipped over in mplayer 
because the timing is not 30 fps or something about the joint location.
In avidemux, I frame# 375 is at 12.537 , 390 is at 13.039 but should be 
13.000. Each I frame drifts in time by 0.001 or 0.002 seconds. Using the 
inquiry function of ffmpeg reveals the problem.


 ffmpeg -i 20141107_test1-loses2frames.mov
ffmpeg version N-68279-g72c9844 Copyright (c) 2000-2014 the FFmpeg 
developers

 built on Dec  7 2014 14:41:06 with gcc 4.9.1 (GCC)
 configuration: --enable-gpl --enable-libx264 --disable-doc 
--disable-htmlpages --disable-podpages --disable-ffserver 
--disable-network --enable-libvpx

 libavutil  54. 15.100 / 54. 15.100
 libavcodec 56. 14.100 / 56. 14.100
 libavformat56. 15.102 / 56. 15.102
 libavdevice56.  3.100 / 56.  3.100
 libavfilter 5.  2.103 /  5.  2.103
 libswscale  3.  1.101 /  3.  1.101
 libswresample   1.  1.100 /  1.  1.100
 libpostproc53.  3.100 / 53.  3.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '20141107_test1-loses2frames.mov':
 Metadata:
   major_brand : qt
   minor_version   : 512
   compatible_brands: qt
   encoder : Lavf56.15.102
 Duration: 00:00:22.08, start: 0.00, bitrate: 7661 kb/s
   Stream #0:0(eng): Video: h264 (Main) (avc1 / 0x31637661), yuv420p, 
1280x720, 7149 kb/s, 29.91 fps, 30 tbr, 15360 tbn, 30720 tbc (default)

   Metadata:
 handler_name: DataHandler
   Stream #0:1(eng): Audio: pcm_s16le (sowt / 0x74776F73), 32000 Hz, 
mono, s16, 512 kb/s (default)

   Metadata:
 handler_name: DataHandler

It thinks the frame rate should be 29.91 fps. So why does it correct 
itself only at the joint?


Looking back in the output from when this file was created, it DOES say 
it knew the output file should be 30.0 fps.


Loading #2 in avidemux shows one I frame every 10 seconds. If I ignore 
the warped time stamp, All I-frames look correct. Note the length says 
it is  21:21.13 and fps is 1.5

 ffmpeg -i 20141107_test2-freezes.mov
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '20141107_test2-freezes.mov':
 Metadata:
   major_brand : qt
   minor_version   : 512
   compatible_brands: qt
   encoder : Lavf56.15.102
 Duration: 00:21:21.13, start: 0.00, bitrate: 393 kb/s
   Stream #0:0(eng): Video: h264 (Main) (avc1 / 0x31637661), yuv420p, 
1280x720, 367 kb/s, 1.50 fps, 30 tbr, 15360 tbn, 30720 tbc (default)

   Metadata:
 handler_name: DataHandler
   Stream #0:1(eng): Audio: pcm_s16le (sowt / 0x74776F73), 32000 Hz, 
mono, s16, 512 kb/s (default)

   Metadata:
 handler_name: DataHandler

Trying ffplay for playback, has same problem as mplayer, so can't blame 
mplayer.


Loading #3 in avidemux shows each I-frame to be every 0.0282 seconds
 ffmpeg -i 20141107_test3-fails.mov
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '20141107_test3-fails.mov':
   encoder : Lavf56.15.102
 Duration: 00:00:54.00, start: 0.00, bitrate: 7985 kb/s
   Stream #0:0(eng): Video: h264 (Main) (avc1 / 0x31637661), yuv420p, 
1280x720, 132398 kb/s, 531.60 fps, 30 tbr, 19200 tbn,
   Stream #0:1(eng): Audio: pcm_s16le (sowt / 0x74776F73), 32000 Hz, 
mono, s16, 512 kb/s (default)


531 fps? but this time the duration correctly says 54 seconds. Is there 
a memory leak? The fps and length are appearing more like random than a 
pattern.


So how do I force it to write the output at 30 fps? If I add -r 30 , the 
new console output now does show a change for tbc to 30
 ffmpeg -f concat -i concat-list3.txt -r 30 -c copy 
20141107_test3-rate30.mov

Output #0, mov, to '20141107_test3-rate30.mov':
 Metadata:
   encoder : Lavf56.15.102
   Stream #0:0: Video: h264 (avc1 / 0x31637661), yuv420p, 1280x720, 
q=2-31, 7453 kb/s, 30 fps, 15360 tbn, 30 tbc
   Stream #0:1: Audio: pcm_s16le (sowt / 0x74776F73), 32000 Hz, mono, 
512 kb/s


but playing the resulting file still has the problem, and using identify 
also shows no effective change.

 ffmpeg -i 20141107_test3-rate30.mov
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '20141107_test3-rate30.mov':
   encoder : Lavf56.15.102
 Duration: 00:00:54.00, start: 0.00, bitrate: 7985 kb/s
   Stream #0:0(eng): Video: h264 (Main) (avc1 / 0x31637661), yuv420p, 
1280x720, 131538 kb/s, 528.15 fps, 30 tbr, 15360 tbn,
   Stream #0:1(eng): Audio: pcm_s16le (sowt / 0x74776F73), 32000 Hz, 
mono, s16, 512 kb/s (default)


Now that I've narrowed down the problem further, this looks very similar 
to another problem I started having last month when I started switching 
my scripts over 

[FFmpeg-user] AVCHD files

2014-12-08 Thread emaggi...@libero.it
Hi,

Could someone give me any hints to deal with AVCHD (*.MTS) files with free 
software?

I've tried mplayer (MPlayer SVN-r37150-4.8.3) with no result.
ffmpeg (version 2.1.5) seems go better: it'a problem on my hw resource, also, I 
think (Athlon 64X2@2.7Ghz, with 4GB RAM).


And these files are on some DVD UDF that are damaged, too.
GNU ddrescue (v.1.17) is a excellent tool for recovering a disk image, then I'd 
carve it with foremost. 
Foremost (version 1.5.7) is also a excellent tool, but needs instructions on 
what types of file are to be found.
File (v.5.19) enigmatically say that *.MTS files are data.

With a little research on the net I cannot found the signature (magic number) 
for these files: it's depends on copyright reason?

I've tried also empirically with hexdump and strings on the head of some valid 
MTS file, looking for it.
Then on the basis of some MTS file head comparison I've write these lines on my 
foremost.conf
# euristic AVCHD MTS from head confontation:
MTSy20 \x00 \x00 \xb0 \x11 \x00 \x00 \xc1 \x00 \x00 \x00 
\x00 \xe0 \x1f \x00 \x01 \xe1 \x00 \x23 \x5a \xab \x82 

Is this a reliable test?

Thanks,
 enrico
___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user


[FFmpeg-user] Image sequence naming convention

2014-12-08 Thread Andreas Weller
Hi.

I did a bunch of camera shots with my raspberry pi camera. In my script
I use the following naming scheme for the files created: `date
--iso-8601=seconds`.jpg

Now I would like to further process them with ffmpeg.

What's the correct way to call ffmpeg with these files? Most examples on
the net simply do it like -f image2 output_%05d.jpg - but this
obviously doesn't work.


Thank you!


Regards,
  Andreas

___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user


Re: [FFmpeg-user] Image sequence naming convention

2014-12-08 Thread S Andreason

Andreas Weller wrote:

I did a bunch of camera shots with my raspberry pi camera. In my script
I use the following naming scheme for the files created: `date
--iso-8601=seconds`.jpg

Now I would like to further process them with ffmpeg.

What's the correct way to call ffmpeg with these files? Most examples on
the net simply do it like -f image2 output_%05d.jpg - but this
obviously doesn't work.

  


Hi Andreas,

If you are looking to turn those images into an animation or movie, then 
this is the format of the command you're looking for:


 ffmpeg -r 30 -i frame-%d.ppm -c:v h264 -s 640x480 out.avi

You'll need each frame numbered in sequential order. The date sequence 
won't work as is.

frame-1.ppm frame-2.ppm frame-3.ppm etc...

and unless there is a new filter for loading non-RAW files, I'd suggest 
a script to batch

 convert frame-1.jpg frame-1.ppm

Stewart
___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user


Re: [FFmpeg-user] Image sequence naming convention

2014-12-08 Thread Carl Eugen Hoyos
S Andreason sandreas41 at gmail.com writes:

   ffmpeg -r 30 -i frame-%d.ppm -c:v h264 -s 640x480 out.avi
 
 You'll need each frame numbered in sequential order. 
 The date sequence won't work as is.
 frame-1.ppm frame-2.ppm frame-3.ppm etc...
 
 and unless there is a new filter for loading non-RAW 
 files, I'd suggest a script to batch
   convert frame-1.jpg frame-1.ppm

This seems to be a very bad idea since different 
colour spaces are involved and FFmpeg should be 
able to read the jpg file.

If renaming the source files is impossible, using 
a glob pattern could be an alternative.

Carl Eugen

___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user