If it were possible to post a clip some where it could help us see the problem. Otherwise, you would have to assume some sort of change occurred with the exact encoding method, though I am no expert on this issue. Perhaps if one could find a system where it works and were it does not, it could be helpful.
Thank You,
PS, that chart seemed more understandable after drinking a beer I noticed. Perhaps you might try that as well. lol!
caf4926, I looked at that sketch in your first post, and it does not make much sense to me, unless you are testing across multiple boot partitions (or PCs).
Are you saying the following ?
video encoded today with ffmpeg ( ? ) can’t be played by smplayer/mplayer of today (version please ? )
video encoded yesterday with ffmpeg ( ? ) CAN be played by smplayer/mplayer of today
video encoded today with ffmpeg ( ? ) CAN be played by an older smplayer/player (version please ?)
video encoded yesterday with ffmpeg ( ? ) CAN be played by smplayer/mplayer of yesterday
At the risk of stating the obvious which unfortunately may not even be relevant, I note this:
I think “Lavf”, is “LibAVFormat” which is the ffmpeg file parser. I don’t think its an encoder.
Ergo could it be that there is something with the file parsing ? ie the wrapper/packaging, where today’s mplayer/smplayer is a bit fussier than yesterday’s mplayer/smplayer, and the change in LibAVFormat has modified the wrapper/packaging (call it the file parsing).
My apologies for my horrible non-technical description and bad terminology.
Ergo to summarize my rather speculative look : today’s smplayer/mplayer has a problem with videos wrapped/parsed with Lavf53.2.0 ?
So the link Carl sent plays just fine on two up to date openSUSE 11.4 64 bit PC’s, but I have updated to mplayer2 on both computers and no longer use the old mplayer2. Of course, from the link I got, I used Totem to actually play the video, if that makes any difference which was my only choice in Firefox.
Given mplayer2 does NOT include mencoder, I consider mplayer2 more an inferior software ‘fork’ than an update, although I can’t deny it may play some formats that mplayer fails to play.
Edit : I note the video (with audio) plays on my PC with xine, vlc and ffplay.
Given mplayer2 does NOT include mencoder, I consider mplayer2 more an inferior software ‘fork’ than an update, although I can’t deny it may play some formats that mplayer fails to play.
Edit : I note the video (with audio) plays on my PC with xine, vlc and ffplay.
So I do not know what the issue is, but with so many choices that do play, I am not sure how big a problem this might be. As for updating to mplayer2 from mplayer, I have had no issues with playing content online or offline but I did note the differences in a message oldcpu had posted elsewhere. I am thinking mplayer2 will at some point be the only real choice, but perhaps I don’t understand the issues for the fork.
if and when mencoder is included with mplayer2 I may update to mplayer2. But with no mencoder, I can’t honestly even consider updating to mplayer2. I use mencoder a fair amount.
Mencoder or no mencoder. (Edit - I’ve read the developers on the team of the mplayer2 fork are unhappy with the mencoder code - so much so they refuse to use it).
That file doesn’t play in SMPlayer (well the audio doesn’t), on my updated 11.4 system.
Let me check what updates are in the system that hasn’t been updated very recently.
Hmmm … looking at the changelog between 1.0rc4_r33321 and r33574
* Sun Jun 12 2011 pascal.bleser-at-opensuse.org
- update to 33574:
* fix segmentation fault when pressing U (stop playing) in GUI
* a GUI related option which corresponds to a MPlayer option '-optname' will
be named '-gui-optname' now
* rename option '-guiwid' '-gui-wid'
* add option '-gui-include'
* constrain libcdparanoia's caching which badly breaks playback with -nocache
* improve seeking for files where start_time is not (properly) set
* setup default speex modes, allows decoding of speex in flv which does not
contain a header packet
* fix #1442: do not remove video filters: If a video filter option isn't set,
it does not mean to remove the filter, but such option (if set) should only
cause adding the filter (if not already present). That way filters can be
specified in MPlayer's config or on the command line, too.
* support Etymonix MPEG-2 I-frame Video Codec
* Command Line Options: Support FFmpeg per-codec AVOptions: this change
extends the existing support for arbitrary FFmpeg options to also include
per-codec AVOptions. You can now pass per-codec options in the same way as
global options are passed: mplayer -lavdopts o=<option>
* increase the maximum value of the DVB timeout to 240 seconds: some devices
may need more time for the initial tune (e.g. firmware loading). Let the
user specify higher timeout value if there is need to. The default remains
30 seconds.
* fix #574: instead of exit_player(), exit with appropriate return code
* support up to 20 mouse buttons, there really seem to be input devices with
at least 12
* fix bug introduced in r25726 that would cause us to write uninitialized
data into buffer
* support displaying of 9- and 10-bit pixel formats, as used by v210 and H264
* support playback for more ffv1 pixel formats
* support S302M (with -demuxer lavf)
* support S302M in the native ts demuxer
* support FFmpeg r10k decoder
* support Digital Voodoo SD 8 Bit (DVOO)
* add MNG output support
* stream dump: print progress information
* allow directory selection with middle mouse button (single click)
* start playlist selection in same directory the file selector starts in
* support 9- and 10-bit YUV input for OpenGL vos
* fix clear/border color of chroma texture for 9- and 10-bit formats: avoids
pink borders for those formats
* workaround embedded ssa/ass never hiding with -noass
* only accept regular files as .bin files for .cue files
* remove a duplicated open() call that could lead to a file-descriptor leak
in some cases
* avoid possible crash if cue filename is very short
* take notice of MPlayer option '-display': if option '-display' is given,
initialize GTK in a way that the GUI will run on that display
* change -udp-slave code to temporarily fall back to normal timing after 30s
network timeout
- bump ffmpeg to d127d26997ff046a9f112643eebd1007590b4fa2
* Sun Apr 24 2011 pascal.bleser-at-opensuse.org
- update to 33321:
............
Its hard to say what caused it … there are many changes … some of them being (that maybe could be relevant):
bump ffmpeg to d127d26997ff046a9f112643eebd1007590b4fa2
Command Line Options: Support FFmpeg per-codec AVOptions: this change extends the existing support for arbitrary FFmpeg options to also include per-codec AVOptions. You can now pass per-codec options in the same way as global options are passed: mplayer -lavdopts o=<option>
… I’m guessing … possibly one of them.
My guess is this is an upstream bug. I’m in above my head in some other stuff right now, and don’t have time to search for upstream bug reports on this.
On my main Desktop, I had to use a solution because it was just ticking me off.
So I switched to a backup packman repo I had rsync’d to my HD
It’s all working fine now. And I’ll probably stick with it like this