[FFmpeg-user] does flv support the metadata of rotation as mov ?
Hi, I have read the source code of movenc.c, which supports the metadata of rotation. Does flv support 'rotation' like mov? Thanks! Regards Andrew ___ 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".
[FFmpeg-user] Saving h264 RTP stream into an mp4 container
Hi, I'm trying to save incoming RTP stream into an MP4 container -- without transcoding it. For the most part, it works well, except two problems. First problem is mostly cosmetic: at the start of every file there's a warning logged: "Timestamps are unset in a packet for stream ...". I don't see how it is possible to set both PTS and DTS, given only RTP timestamps. The second problem is related, but is functional rather than cosmetic: if the RTP stream contains B-frames, whenever timestamp order decreases -- which is normal for that stream -- the packet fails to be written, with an error of "Application provided invalid, non monotonically increasing dts to muxer ..." So, the question is -- what is the correct way to assign timestamps when muxing raw h264 frames? ___ 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] ffmpeg -f avfoundation HELL - please help!
Well, it seems I need to be Einstein to grab the fucking screen with audio using ffmpeg, no one in this world can help me either. The examples aout there are way too simple. There are no guides either. This is depressing. On Tue, Jan 3, 2017 at 3:58 PM Demián Andrés Rodriguezwrote: > Nearly with vp8/vp9. With h264/aac it certainly does, but audio still > sucks... > > frame= 95 fps= 31 q=28.0 size= 277kB time=00:00:03.15 bitrate= > 719.4kbits/s dup=2 drop=0 speed=1.03x > frame= 141 fps= 31 q=28.0 size= 307kB time=00:00:04.68 bitrate= > 536.2kbits/s dup=2 drop=0 speed=1.02x > frame= 157 fps= 31 q=28.0 size= 336kB time=00:00:05.20 bitrate= > 528.2kbits/s dup=2 drop=0 speed=1.02x > frame= 203 fps= 30 q=28.0 size= 379kB time=00:00:06.75 bitrate= > 459.5kbits/s dup=2 drop=0 speed=1.01x > frame= 234 fps= 30 q=28.0 size= 413kB time=00:00:07.76 bitrate= > 435.5kbits/s dup=2 drop=0 speed=1.01x > frame= 264 fps= 30 q=28.0 size= 449kB time=00:00:08.77 bitrate= > 419.0kbits/s dup=2 drop=0 speed=1.01x > > What other options do I have? Please tell me which is the fastest way to > encode using vp8/vp9 and I will use it at least to discard if cpu is the > problem. > > On Tue, Jan 3, 2017 at 3:54 PM Reindl Harald > wrote: > > > > Am 03.01.2017 um 19:48 schrieb Demián Andrés Rodriguez: > > Recording just audio either with vorbis, opus or aac/mp4 works fine, the > > problem is video & audio together > > what about anser the question "Does it encode at realtime, i.e. speed >= > 1.0" asked multiple times? > ___ > 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". > > ___ 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] Generate/Playback Interlaced (1080i) video with lavfi.
On Wed, 4 Jan 2017, Matthias, Thomas wrote: Thanks for the help Carl. Seeing the difference isn’t that important, really what I’m trying to accomplish is outputting 1080i video using the FFMPEG wrapper for the DeckLink SDK. I’ve tried numerous combinations of filter settings with either 1080i or 1080p source content, but I think there I just a limitation/bug in the DeckLink wrapper/device code that’s preventing me from outputting interlaced video. Any clues on how to fix this would be amazing! Just a wild guess, but if I remember correctly, the decklink device advertises both progressive and interlaced 1080 formats. You can check this with -list_formats. If the progressive one is the first, then I think the decklink device in ffmpeg will use that, no matter what you do. If that is really the case, then this has to be fixed in ffmpeg/libavdevice. Regards, Marton ___ 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] OpenEXR input with cropped data window produces black output
Hello, FFmpeg EXR decoder, have some trouble with display window != data window And if i remember correctly, if top left doesn't start at 0/0 But the way to manage the display window, is not the same, if data window is bigger than display window or the opposite. I begin to work on this, but not enough time now to finish this patch Martin ___ 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] Generate/Playback Interlaced (1080i) video with lavfi.
Thanks for the help Carl. Seeing the difference isn’t that important, really what I’m trying to accomplish is outputting 1080i video using the FFMPEG wrapper for the DeckLink SDK. I’ve tried numerous combinations of filter settings with either 1080i or 1080p source content, but I think there I just a limitation/bug in the DeckLink wrapper/device code that’s preventing me from outputting interlaced video. Any clues on how to fix this would be amazing! On 12/21/16, 12:31 AM, "ffmpeg-user on behalf of Carl Eugen Hoyos"wrote: 2016-12-20 21:11 GMT+01:00 Matthias, Thomas : > Hi, > > I’m trying to generate some smtpebars with a script, and it works fine for 1080p video. > However, I’ve tried numerous combinations for generating 1080i video using the > –vf tinterlace filter. How do you expect to see the difference between interlaced and progressive smtpebars? I believe the tinterlace filter works fine but without horizontal movement you will not be able to see a difference. Note that the following should simplify your command line: $ ffmpeg -f lavfi -i filter1,filter2, ..., filtern 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". ___ 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] [FFmpeg-devel] vstats info options
Hello, I have been able to get accurate bitrate stats by using the OutputStream's data_size structure in the print_report method. total_size += ost->data_size; // For enc->codec_type AVMEDIA_TYPE_VIDEO and AVMEDIA_TYPE_AUDIO. Progress reports now have valid size and bitrate for non-seekable media. This is a fallback setting on total_size if it is still not valid after avio_tell(). Does this seem ok? Is it preferrable to file a bug report or a commit for review? Thanks On Tue, Jan 3, 2017 at 8:13 PM, Kamaldeep Tumkurwrote: > Hello, > > It appears that the progress information is more valid for 'seek'able > media. But for live capture/encode, the frame_size and bitrate values > are not valid. If there was access to the AVPacket structure or frame > data at the time of sending the progress, we may have had reliable > size and bitrate stats. > > Also lines 1560-1562 (ffmpeg.c v3.1.1) > > total_size = avio_size(oc->pb); > if (total_size <= 0) // FIXME improve avio_size() so it works with > non seekable output too > total_size = avio_tell(oc->pb); > > Is there another possiblility here? Could you shed some light on why > the size is calculated based on seek lengths. > > Thanks > Kamal > > On Mon, Jan 2, 2017 at 5:28 PM, Kamaldeep Tumkur > wrote: >> Thanks for your suggestion on the use of progress. This is definitely >> a user question. Though some further questions are: >> >> frame=4866 >> fps=24.1 >> stream_0_0_q=32.0 >> bitrate=N/A >> total_size=N/A >> out_time_ms=202688000 >> out_time=00:03:22.688000 >> dup_frames=0 >> drop_frames=0 >> speed= 1x >> progress=continue >> >> >> There is no valid bitrate information. vstats always returns a valid >> bitrate-per-frame and an avg-bitrate-per-sec value. Is there a way to >> get the current bitrate consistently? >> >> Are the assessments in the following ticket accurate? >> https://trac.ffmpeg.org/ticket/5288 >> >> Going to look at having my application reliably use progress data. >> >> Thanks >> Kamal >> >> On Mon, Jan 2, 2017 at 1:44 PM, Marton Balint wrote: >>> >>> On Mon, 2 Jan 2017, Marton Balint wrote: >>> On Mon, 2 Jan 2017, Kamaldeep Tumkur wrote: > Hello, > > Is there a possibility to provide the option to write vstats info from > an encoding session remotely, say to another DB or to memory from > where it can be handled by another process. > > The only possibility for consuming vstats info in real-time seems to > be to have some kind of watch+notifier that will act upon > modifications to the vstats log file. I want to be able to use these > stats from a live encoding session for other purposes. Please suggest. > Would you accept patches to ffmpeg.c so that it would provide more > options to export the vstats info? Thanks for the time. Have you checked the -progress option? >>> >>> Also please respond on ffmpeg-user mailing list, as this question clearly >>> belongs there. >>> >>> Thanks, >>> >>> Marton >>> ___ >>> ffmpeg-devel mailing list >>> ffmpeg-de...@ffmpeg.org >>> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel ___ 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".