Re: [FFmpeg-user] on frame changes in recode

[Moritz Barsnick]

You broke threading. :-/ (SCNR)

On Fri, Apr 29, 2016 at 12:34:07 +0200, Carl Eugen Hoyos wrote:

> > ffmpeg version 2.8.6-1ubuntu2
> Generally, on this mailing list, only current FFmpeg git head is supported.

BTW, where did Xen's peculiar "WARNING: library configuration mismatch
avcodec" come from? I don't see a mismatch.

> Feel free to provide an input sample.

Command line and how to obtain the input file were provided in
Message-ID: <5a6c03dccfa6ddcb311c5d53f4d1c...@dds.nl>

As mentioned, I couldn't reproduce, but I'm not sure my methods were
correct. (Using mplayer to check doesn't avoid using its demuxer. I
hope my demuxing to image2 using ffmpeg was okay.)

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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Xen]

Carl Eugen Hoyos schreef op 29-04-2016 12:34:

Hi!


ffmpeg version 2.8.6-1ubuntu2


Generally, on this mailing list, only current FFmpeg git head is 
supported.


Well there you get it, more nitpicking. Important reason not to give 
this data to people.


Example: if you go to #ubuntu (before release of 16.04) and ask 
something about, say, nvidia driver or issue with game, and you say that 
you use 16.04, you get directed to #ubuntu+1 because "unreleased 
versions are not supported". But #ubuntu+1 doesn't care about issues 
that don't have anything to do with #ubuntu+1 so not really feel like 
helping you with stuff that is equivalent across versions.


In the end it turned out my issue was with nVidia 361.28 driver which 
was also supported in 15.10 so my place indeed was in #ubuntu and not 
#ubuntu+1.


It was a steam issue and in #ubuntu-steam they also did not know but 
there was some prick directing me to #ubuntu+1 instead of to 
#ubuntu-steam. Which was pointless. You see?


And, as it turns out, in this case the issue is also not related to 
ffmpeg! Apparently not in any case. It seems to be related to players 
instead of encoders.


So you would then ask me to install and compile ffmpeg from git head, 
which is ludicrous and costs so much time and energy when the benefits 
are minimal. Sometimes you can just look at whether the issue has 
anything to do with version, first, you know.



(Among the reasons are that our man-power is limited and that we can 
only
fix issues in current FFmpeg git head, non-security patches are 
normally

not backported.)


No one is asking you to fix issues in older versions, come on.


There is also a bug tracker for FFmpeg on Ubuntu but you would likely 
be

asked to test current FFmpeg there as well.


And there's be no difference in this issue.


Another reason for using current FFmpeg git head is that we now have a
better default aac encoder.


It's not about audio.



frame=  666 fps=8.0 q=-1.0 Lsize=2496kB time=00:00:22.20
bitrate=  920.6kbits/s


This status line indicates that no frame was dropped or duplicated 
while

transcoding.


Thank you, I'm sorry for leaving it out.

(Although I realize you didn't post your command line, it is possible 
to
create command lines that drop and / or duplicate frames without 
showing

anything suspicious in the status line.)


Apologies, there have been multiple messages and I have posted my 
command line already in one of them and it was very simple:


ffmpeg -i input.mp4 -ss 00:5:50.800 -vcodec libx264 output.mp4


Feel free to provide an input sample.


Both output and input were already provided, to my dismay.

In any case, the situation is that:

(as I said in the other mail).

* input sees no duplicated frame in VLC and gstreamer
* output sees no duplicated frame in mplayer / ffmpeg
* output sees duplicated frame in VLC and gstreamer.

* Main difference would be new keyframe right before duplicated frame.

I left the status line out because it was garbled from ctrl-z followed 
by other stuff followed by fg Apologies.

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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Xen]

Moritz Barsnick schreef op 29-04-2016 11:02:

On Thu, Apr 28, 2016 at 23:31:57 +0200, Xen wrote:

Basically there is no other information unless you want the number of
spaces typed on average every day on my system.


That would be very kind of you. I can recommend some keyloggers, but
ask you to do the math for getting the average.


That's too hard on me, I need help.

Actually I do need help but not with this :p.


I stepped through both the input video (downloaded using youtube-dl
with your command) and "output.mp4" in two ways:
- mplayer and the '.' key;
- dumping frames to images using ffmpeg and "vsync -0" (I hope that's
  correct for dumping exactly one output frame for every input frame).


Well I am just slightly amazed but I have the same result (with 
mplayer).


VLC still duplicates the frame.

Pitivi (probably using gstreamer to play it) also duplicates the frame.

So the issue is not in the rendering but in the player :@.

Alright that's stupid isn't it.

I'm not sure whether I've been stupid or the programs

Thank you for checking in any case I'm sorry. If I'm an ass 
sometimes :p.


:).


I'm happy just shutting my mouth and letting unresolved problems stay
unresolved.


You're good that way :D.


That is of course a quite valid issue, be it a bug or just incorrect
use of ffmpeg.


Yeah...

What other players should I try. DragonPlayer is too stupid to exist and 
I cannot access any Windows tools right now, and possibly I would use 
VLC there anyway.


If your player starts inserting frames it means its actually framerate 
is either going to be higher, or the video will play longer.


They said films were often recorded in 24fps, when played on 25fps 
devices I think they would stretch it and the speed would go up and the 
pitch of the audio, shortening the film. But that requires actual 
playback to be of a higher fps.


In that case there would be no frame duplication or anything, just 
frames playing too fast. To do it right they would have to duplicate 
every 24th frame.


But that's probably impossible with audio or something, don't know.


BASICALLY:

input is 264, from YT, doesn't show artifact in VLC
output is 264, does show artefact in VLC
difference is new keyframe right before the duplicated frame?


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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Carl Eugen Hoyos]

Hi!

> ffmpeg version 2.8.6-1ubuntu2

Generally, on this mailing list, only current FFmpeg git head is supported. 
(Among the reasons are that our man-power is limited and that we can only 
fix issues in current FFmpeg git head, non-security patches are normally 
not backported.)
There is also a bug tracker for FFmpeg on Ubuntu but you would likely be 
asked to test current FFmpeg there as well.

>   Stream #0:1 -> #0:1 (aac (native) -> aac (libvo_aacenc))

Another reason for using current FFmpeg git head is that we now have a 
better default aac encoder.

> frame=  666 fps=8.0 q=-1.0 Lsize=2496kB time=00:00:22.20 
> bitrate=  920.6kbits/s

This status line indicates that no frame was dropped or duplicated while 
transcoding.
(Although I realize you didn't post your command line, it is possible to 
create command lines that drop and / or duplicate frames without showing 
anything suspicious in the status line.)

Feel free to provide an input sample.

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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Xen]

Cley Faye schreef op 29-04-2016 2:50:

2016-04-28 23:31 GMT+02:00 Xen :



All of this is unnecessary, but if it makes you happy, here you have 
it ;P.



​I was curious about your issue, and I too was eagerly waiting for that
log. Sadly it didn't come.


Well I am sorry, I had CTRL-Z-ed in between and so the output was a mess 
and I had to reconstruct it in that sense.


I guess I should just provide everything next time (if there'll be a 
next time) and not give anyone any reason to complain :-/.



ffmpeg version 2.8.6-1ubuntu2 Copyright (c) 2000-2016 the FFmpeg 
developers

  built with gcc 5.3.1 (Ubuntu 5.3.1-11ubuntu1) 20160311
  configuration: --prefix=/usr --extra-version=1ubuntu2 
--build-suffix=-ffmpeg --toolchain=hardened 
--libdir=/usr/lib/x86_64-linux-gnu 
--incdir=/usr/include/x86_64-linux-gnu --cc=cc --cxx=g++ --enable-gpl 
--enable-shared --disable-stripping --disable-decoder=libopenjpeg 
--disable-decoder=libschroedinger --enable-avresample --enable-avisynth 
--enable-gnutls --enable-ladspa --enable-libass --enable-libbluray 
--enable-libbs2b --enable-libcaca --enable-libcdio --enable-libflite 
--enable-libfontconfig --enable-libfreetype --enable-libfribidi 
--enable-libgme --enable-libgsm --enable-libmodplug --enable-libmp3lame 
--enable-libopenjpeg --enable-libopus --enable-libpulse --enable-librtmp 
--enable-libschroedinger --enable-libshine --enable-libsnappy 
--enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora 
--enable-libtwolame --enable-libvorbis --enable-libvpx 
--enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxvid 
--enable-libzvbi --enable-openal --enable-opengl --enable-x11grab 
--enable-libdc1394 --enable-libiec61883 --enable-libzmq --enable-frei0r 
--enable-libx264 --enable-libopencv

  WARNING: library configuration mismatch
  avcodec configuration: --prefix=/usr --extra-version=1ubuntu2 
--build-suffix=-ffmpeg --toolchain=hardened 
--libdir=/usr/lib/x86_64-linux-gnu 
--incdir=/usr/include/x86_64-linux-gnu --cc=cc --cxx=g++ --enable-gpl 
--enable-shared --disable-stripping --disable-decoder=libopenjpeg 
--disable-decoder=libschroedinger --enable-avresample --enable-avisynth 
--enable-gnutls --enable-ladspa --enable-libass --enable-libbluray 
--enable-libbs2b --enable-libcaca --enable-libcdio --enable-libflite 
--enable-libfontconfig --enable-libfreetype --enable-libfribidi 
--enable-libgme --enable-libgsm --enable-libmodplug --enable-libmp3lame 
--enable-libopenjpeg --enable-libopus --enable-libpulse --enable-librtmp 
--enable-libschroedinger --enable-libshine --enable-libsnappy 
--enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora 
--enable-libtwolame --enable-libvorbis --enable-libvpx 
--enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxvid 
--enable-libzvbi --enable-openal --enable-opengl --enable-x11grab 
--enable-libdc1394 --enable-libiec61883 --enable-libzmq --enable-frei0r 
--enable-libx264 --enable-libopencv --enable-version3 --disable-doc 
--disable-programs --disable-avdevice --disable-avfilter 
--disable-avformat --disable-avresample --disable-postproc 
--disable-swscale --enable-libopencore_amrnb --enable-libopencore_amrwb 
--enable-libvo_aacenc --enable-libvo_amrwbenc

  libavutil  54. 31.100 / 54. 31.100
  libavcodec 56. 60.100 / 56. 60.100
  libavformat56. 40.101 / 56. 40.101
  libavdevice56.  4.100 / 56.  4.100
  libavfilter 5. 40.101 /  5. 40.101
  libavresample   2.  1.  0 /  2.  1.  0
  libswscale  3.  1.101 /  3.  1.101
  libswresample   1.  2.101 /  1.  2.101
  libpostproc53.  3.100 / 53.  3.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'input.mp4':
  Metadata:
major_brand : mp42
minor_version   : 0
compatible_brands: isommp42
creation_time   : 2015-12-03 11:44:40
  Duration: 00:06:13.01, start: 0.00, bitrate: 1355 kb/s
Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 
1280x720 [SAR 1:1 DAR 16:9], 1208 kb/s, 30 fps, 30 tbr, 30 tbn, 60 tbc 
(default)

Metadata:
  handler_name: VideoHandler
Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 44100 Hz, 
mono, fltp, 143 kb/s (default)

Metadata:
  creation_time   : 2015-12-03 11:44:42
  handler_name: IsoMedia File Produced by Google, 5-11-2011
[libx264 @ 0x975340] using SAR=1/1
[libx264 @ 0x975340] using cpu capabilities: MMX2 SSE2Slow SlowCTZ
[libx264 @ 0x975340] profile High, level 3.1
[libx264 @ 0x975340] 264 - core 148 r2643 5c65704 - H.264/MPEG-4 AVC 
codec - Copyleft 2003-2015 - http://www.videolan.org/x264.html - 
options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 
psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 
8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=3 
lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0 
bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 
b_bias=0 direct=1 weightb=1 open_gop=0 

Re: [FFmpeg-user] on frame changes in recode

[Moritz Barsnick]

On Thu, Apr 28, 2016 at 23:31:57 +0200, Xen wrote:
> Basically there is no other information unless you want the number of 
> spaces typed on average every day on my system.

That would be very kind of you. I can recommend some keyloggers, but
ask you to do the math for getting the average.

> * whether I use gstreamer (pitivi) or avidemux or ffmpeg
> * the output always has the 2nd frame duplicated, and the "scene" 
> prolonged from 11 frames to 12 frames.

I stepped through both the input video (downloaded using youtube-dl
with your command) and "output.mp4" in two ways:
- mplayer and the '.' key;
- dumping frames to images using ffmpeg and "vsync -0" (I hope that's
  correct for dumping exactly one output frame for every input frame).

I both cases, the transitions look identical, frame by frame.

> You can easily reproduce it given above command.

Sometimes, the answer can already be in the log (see my other answer).
Admitted, not here.

> All of this is unnecessary, but if it makes you happy, here you have it 
> ;P.

I'm happy just shutting my mouth and letting unresolved problems stay
unresolved.

> Maybe you think I'm nitpicking but I'm just concerned that the simplest 
> of edit operations result in this corruption.

That is of course a quite valid issue, be it a bug or just incorrect
use of ffmpeg.

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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Moritz Barsnick]

On Fri, Apr 29, 2016 at 03:38:34 +, Phil Rhodes wrote:
> >There's hardly anything left I think you are nitpicking now.
> You're right, they're nitpicking. You can play this game with them for days 
> and they'll always want one more thing, very few of which have any real 
> purpose. I wouldn't waste your time.

If it's so easy to answer for you with the originally given
information, why don't you do so? ;-)

I know Carl Eugen always want the complete output. But certainly, some
things can be deduced without it, and I try my best sometimes.

But this issue isn't easy for me. I'm just a user, so I usually shut up
if I have no idea, instead of saying "no idea" or "must be a bug". I'm
certain a developer should try to say "it's a bug" or "please post a
bug ticket so someone can check", but nobody forces them to join here.
;-) Anyway I thought more information could be useful:

Xen omitted two pieces of information. While the information s/he gave
*may* have been enough, s/he omitted:
- The banner, telling us whether the ffmpeg version may be outdated. Or
  release vs. git version. (You don't believe how many people report
  issues fixed recently. Or even ages ago.)
- The footer line. Xen explicitly dropped it. But you know, it can
  contain one piece of information relevant to the issue:
  $ grep -B1 -Fw drop ffmpeg.c
  if (nb_frames_dup || nb_frames_drop)
  snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " dup=%d 
drop=%d",
  That's the number of duplicated or dropped frames, tadaa.
  That ffmpeg doesn't report such in *this* given issue I had to
  actually try myself, instead of just getting one extra line of
  copy/paste ASCII in Xen's email.
- Xen refuses to install a keylogger which would allow us to count the
  number of spaces typed in the 24 hours before and after the ffmpeg
  command.

You don't have to try to make it hard for everyone.

That said, I too would appreciate if someone here would step forward
once in a while to say: "Sorry, we have no idea what is going there."
As said, I won't do that, because nobody cares.

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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Phil Rhodes]

>There's hardly anything left I think you are nitpicking now.
You're right, they're nitpicking. You can play this game with them for days and 
they'll always want one more thing, very few of which have any real purpose. I 
wouldn't waste your time.

P



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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Cley Faye]

2016-04-28 23:31 GMT+02:00 Xen :

>
> All of this is unnecessary, but if it makes you happy, here you have it ;P.


​I was curious about your issue, and I too was eagerly waiting for that
log. Sadly it didn't come.
___
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Xen]

Moritz Barsnick schreef op 28-04-2016 13:20:

On Wed, Apr 27, 2016 at 20:17:56 +, Xen wrote:

> Only if you want support on this mailing list;-)
That sounds horrendous.


Why? The log contains valuable information, such as details about the
input, messages regarding what ffmpeg is doing in particular, warnings
and error messages. And, e.g. in your case, whether ffmpeg is dropping
or adding frames.


I was joking. The horrendous part referred to getting support.


Well I think this is about the output of the
first attempt (to mp4):


You shouldn't remove anything, neither your actual command line nor the
banner. (Ton of repetetive messages in the middle could be omitted if
too long.


There's hardly anything left I think you are nitpicking now.

I think I already supplied the command line :-/.

ffmpeg -i input.mp4 -ss 00:5:50.800 -vcodec libx264 output.mp4

Basically there is no other information unless you want the number of 
spaces typed on average every day on my system.


I don't have that data :p.

It's very simple:

* whether I use gstreamer (pitivi) or avidemux or ffmpeg
* the output always has the 2nd frame duplicated, and the "scene" 
prolonged from 11 frames to 12 frames.


I mean I can supply the files as well but I can probably suffice with 2 
youtube-dl commands since they are both on youtube now.


This is the download:

youtube-dl -f 22 https://www.youtube.com/watch?v=N0wdB-Vt-ZY

Here is the finished product:

(rather unfinished :p);

youtube-dl -f 22 https://www.youtube.com/watch?v=DFhFD-UlwRc

You can use any player (e.g. VLC) that can do frame-by-frame to see how 
it is constructed.


I mean, what more do you want?

You can easily reproduce it given above command.

All of this is unnecessary, but if it makes you happy, here you have it 
;P.


Maybe you think I'm nitpicking but I'm just concerned that the simplest 
of edit operations result in this corruption.


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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-user] on frame changes in recode

[Moritz Barsnick]

On Wed, Apr 27, 2016 at 20:17:56 +, Xen wrote:
> > Only if you want support on this mailing list;-)
> That sounds horrendous.

Why? The log contains valuable information, such as details about the
input, messages regarding what ffmpeg is doing in particular, warnings
and error messages. And, e.g. in your case, whether ffmpeg is dropping
or adding frames.

> Well I think this is about the output of the 
> first attempt (to mp4):

You shouldn't remove anything, neither your actual command line nor the
banner. (Ton of repetetive messages in the middle could be omitted if
too long.

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

To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".