I want my website to join some webcam recordings in FLV files (like this one). This needs to be done on Linux without user input. How do I do this? For simplicity's sake, I'll use the same flv as both inputs in hope of getting a flv that plays the same thing twice in a row.
That should be easy enough, right? There's even a full code example in the ffmpeg FAQ.
Well, pipes seem to be giving me problems (both on my mac running Leopard and on Ubuntu 8.04) so let's keep it simple and use normal files. Also, if I don't specify a rate of 15 fps, the visual part plays extremely fast. The example script thus becomes:
ffmpeg -i input.flv -vn -f u16le -acodec pcm_s16le -ac 2 -ar 44100 \
- > temp.a < /dev/null
ffmpeg -i input.flv -an -f yuv4mpegpipe - > temp.v < /dev/null
cat temp.v temp.v > all.v
cat temp.a temp.a > all.a
ffmpeg -f u16le -acodec pcm_s16le -ac 2 -ar 44100 -i all.a \
-f yuv4mpegpipe -i all.v -sameq -y output.flv
Well, using this will work for the audio, but I only get the video the first time around. This seems to be the case for any flv I throw as input.flv, including the movie teasers that come with red5.
a) Why doesn't the example script work as advertised, in particular why do I not get all the video I'm expecting?
b) Why do I have to specify a framerate while Wimpy player can play the flv at the right speed?
The only way I found to join two flvs was to use mencoder. Problem is, mencoder doesn't seem to join flvs:
mencoder input.flv input.flv -o output.flv -of lavf -oac copy \
-ovc lavc -lavcopts vcodec=flv
I get a Floating point exception...
MEncoder 1.0rc2-4.0.1 (C) 2000-2007 MPlayer Team
CPU: Intel(R) Xeon(R) CPU 5150 # 2.66GHz (Family: 6, Model: 15, Stepping: 6)
CPUflags: Type: 6 MMX: 1 MMX2: 1 3DNow: 0 3DNow2: 0 SSE: 1 SSE2: 1
Compiled for x86 CPU with extensions: MMX MMX2 SSE SSE2
success: format: 0 data: 0x0 - 0x45b2f
libavformat file format detected.
[flv # 0x697160]Unsupported audio codec (6)
[flv # 0x697160]Could not find codec parameters (Audio: 0x0006, 22050 Hz, mono)
[lavf] Video stream found, -vid 0
[lavf] Audio stream found, -aid 1
VIDEO: [FLV1] 240x180 0bpp 1000.000 fps 0.0 kbps ( 0.0 kbyte/s)
[V] filefmt:44 fourcc:0x31564C46 size:240x180 fps:1000.00 ftime:=0.0010
** MUXER_LAVF *****************************************************************
REMEMBER: MEncoder's libavformat muxing is presently broken and can generate
INCORRECT files in the presence of B frames. Moreover, due to bugs MPlayer
will play these INCORRECT files as if nothing were wrong!
*******************************************************************************
OK, exit
Opening video filter: [expand osd=1]
Expand: -1 x -1, -1 ; -1, osd: 1, aspect: 0.000000, round: 1
==========================================================================
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
Selected video codec: [ffflv] vfm: ffmpeg (FFmpeg Flash video)
==========================================================================
audiocodec: framecopy (format=6 chans=1 rate=22050 bits=16 B/s=0 sample-0)
VDec: vo config request - 240 x 180 (preferred colorspace: Planar YV12)
VDec: using Planar YV12 as output csp (no 0)
Movie-Aspect is undefined - no prescaling applied.
videocodec: libavcodec (240x180 fourcc=31564c46 [FLV1])
VIDEO CODEC ID: 22
AUDIO CODEC ID: 10007, TAG: 0
Writing header...
[NULL # 0x67d110]codec not compatible with flv
Floating point exception
c) Is there a way for mencoder to decode and encode flvs correctly?
So the only way I've found so far to join flvs, is to use ffmpeg to go back and forth between flv and avi, and use mencoder to join the avis:
ffmpeg -i input.flv -vcodec rawvideo -acodec pcm_s16le -r 15 file.avi
mencoder -o output.avi -oac copy -ovc copy -noskip file.avi file.avi
ffmpeg -i output.avi output.flv
d) There must be a better way to achieve this... Which one?
e) Because of the problem of the framerate, though, only flvs with constant framerate (like the one I recorded through facebook) will be converted correctly to avis, but this won't work for the flvs I seem to be recording (like this one or this one). Is there a way to do this for these flvs too?
Any help would be very appreciated.
I thought it would be a nice learning exercise to rewrite it in Ruby.
It was.
Six months later and three gems later, here's the released product.
I'll still be working a bit on it, but it works.
You'll encounter a very subtle problem here because most video and audio formats (especially in ordinary containers) use "global headers," meaning at the start of the file they have a single header which specifies compression information (like width, height, etc) for the whole file. Concatting two streams will clearly fail, as it will now have two headers instead of one and the muxer may not like this. Converting to AVI probably is resolving the issue in your case because mencoder has code to concat AVIs--that code properly handles the header issue.
After posting my question on mencoder's mailing list, trying other things, I resorted to write my own tool! I started from flvtool and after some digging in the code and writing about 40 lines of code, it works, with no loss in quality (since there is no transcoding).
I'll release it asap, in the meantime anyone interested can contact me.
dont know if this will actually work but try using this command :
cat yourVideos/*.flv >> big.flv
this will probably damage meta information so after executing that command use "flvtool" (ruby script you can find it with google) to fix it.
Related
I'm trying to encode 6 arbitrary mono audio streams into a single AAC 5.1 track in an mp4 container (here with test streams):
ffmpeg -f lavfi -i testsrc=duration=10:size=100x100:rate=30 -f lavfi -i aevalsrc="-2+random(0)" -filter_complex "[1:a][1:a][1:a][1:a][1:a][1:a]join=inputs=6:channel_layout=5.1:map=0.0-FL|1.0-FR|2.0-FC|3.0-LFE|4.0-BL|5.0-BR[a]" -map '0:v' -map "[a]" -c:a aac -channel_layout 5.1 -t 10 testlfe.mp4
5 of the channels replicate the input audio just fine (modulo encoding). However, the LFE channel is lowpassed. Extracting with:
ffmpeg -i testlfe.mp4 -filter_complex "channelsplit=channel_layout=5.1:channels=LFE[LFE]" -map '[LFE]' testlfe.wav
I get a lowpassed rumble, instead of the original full white noise
(from ffmpeg -i testlfe.wav -lavfi showspectrumpic=s=640x320 testlfe.png)
Is there a way to prevent the lowpass from happening?
I couldn't find any references whether that's inherent to the AAC 5.1 encoding, something that ffmpeg does, or inherent to the decoding process. (I did decode my same test files using something that uses Microsoft MediaFoundation and the LFE channel was still lowpassed).
Turns out, the AAC codec inherently limits the LFE bandwidth, so there's no way around it.
(thanks to kesh in the comments) Wikipedia's Advanced Audio Encoding article claims the upper limit is 120Hz which matches my spectrogram, but doesn't cite a source. The actual ISO/IEC 13818-7:2006(en) Standard costs a bunch of money to read as usual, but in the free glossary there is an entry:
low frequency enhancement ( LFE ) channel:
limited bandwidth channel for low frequency audio effects in a multichannel system
Encode with
ffmpeg -i 6channels.wav -filter "channelmap=0|1|2|3|4|5:6.0(front)" -c:a libfdk_aac -ac 6 -profile:a aac_he -vbr 1 -cutoff 18000 -movflags +faststart 6channels-vbr1-fdk.m4a
It can also be done with regular aac codec.
I'm using latest static build of ffmpeg windows.
My input file (.mkv) is:
[video] - 1080, V_MPEG4/ISO/AVC, 14.6 Mbps, ID#0
[audio] - DTS 5.1, 1510 Kbps, ID#1
[subtitles] - S_TEXT/ASS Lossless English, ID#14
My problem is this: I convert the audio, so that my target player, a XB1 console (media support faq), is able to play audio/video. However sometimes its rather difficult to hear or parts may be in foreign language, so I want to force the english subtitles into the mix at the same time I convert the audio.
Currently for the audio, I use the following command
ffmpeg -i input.mkv -codec copy -acodec ac3 output.mkv
Can I somehow tie in the forced subtitles (onto the video) in order to save an extra process of taking the output.mkv and trying to force subtitles on?
Edit: I've tried using the following command to extract subtitles to be able to edit them
ffmpeg -i Movie.mkv -map 0:s:14 subs.srt
However i get the error: Stream map '0:s:14' matches no streams
Edit2: attempted to extract subtitles and succeeded with
ffmpeg -i input.mkv -map 0:14 -c copy subtitles.ass
but still looking to force the subtitles, nonetheless!
Also - a little bonus to this question - can I somehow extract the .ass file and edit it to only produce subtitles for foreign parts - so english audio doesn't have subtitles during the movie but foreign audio does have subtitles?
Cheers
Edit3:
When I try to use both of the commands at once (my earlier mentioned audio converter & one from the ffmpeg wiki)
ffmpeg -i input.mkv -codec copy -acodec ac3 -vf "ass=subs.ass" output.mkv
I get the following error from ffmpeg,
Filtergraph 'ass=subs.ass' was defined for video output stream 0:0 but codec copy was selected.
Filtering and streamcopy cannot be used together.
Since your media player does not support subtitles, the text has to be burnt onto the video image. For that, use
ffmpeg -i input.mkv -vf "ass=subs.ass" -c:v libx264 -crf 20 -c:a ac3 output.mkv
This will re-encode the video, since text is being added. The CRF value controls the video quality. Lower values produce better quality but larger files. 18 to 28 is a decent range to try.
I'd like to get the best .3g2 quality for my video encoding. I've read about this format on wiki, but as a newbie, I'm not sure to get the best sound/video bitrate.
Format sizes are 128x96, 176x144, 352x288, 704x576, and max 1408x1152 pixels.
I couldn't find the highest video bitrate.
I've tried this :
ffmpeg -i C:\projet.avi -s 1408x1152 -vf "split [a], pad=iw*2:ih [b], [a] alphaextract, [b] overlay=w" -vb 1220000000 -ac 1 -ar 8000 C:\projet.alpha.3g2
it says
"[h263 # 0000000000666bc0] bitrate tolerance 4000000 too small for bitrate 12200000000, overriding
[libopencore_amrnb # 000000000066aee0] bitrate not supported: use one of 4.75k, 5.15k, 5.90k, 6.70k, 7.40k, 7.95k, 10.20k, 12.20k, using 12.20k"
I don't understand what it means ?
Others : I've tried to mention stereo sound, 44000Hz, but it seems not supported.
Any help/Suggestions ? Thank you !
P.S: I need alpha channel and .3g2 for an Augmented Reality software on iOS and Android.
LAST EDIT : I'm still with
ffmpeg -i c:/test.avi -s 1408x1152 -vf "split [a], pad=iw*2:ih [b], [a] alphaextract, [b] overlay=w" -vb 4000000 -ac 1 -ar 8000 C:\test.alpha.3g2
It's working well except error messages, and bad quality sound. Any more ideas ?
For video, your supplied bitrate (1200M!) is too high. Use -q:v 1 instead of -b:v for best quality.
For audio, by default, ffmpeg tells the encoder to use a bitrate of 128k unless specified otherwise. Which is not valid for the default encoder used here. So the encoder has overridden the generic value with 12.20k.
If your application supports it, use
ffmpeg -i C:\projet.avi -filter_complex "[0]alphaextract[a];[0][a]hstack" -c:v libx264 -crf 20 -c:a aac C:\projet.alpha.3g2
So I'm trying to stream on YouTube using a raspberry pi. The idea is for one raspberry pi to be used to stream the connected webcam and for another to display the stream, sort of like a surveillance camera. Both raspberry pi's are currently using Raspbian.
So is it possible for me to stream directly to YouTube on a Raspberry Pi.
You can use any Pi supported RTMP/Flash encoder to publish a YouTube live event. One example is ffmpeg which can be compiled on Raspbian.
Create your YouTube live event using the guide. You can find the various encoder settings here.
When everything is ready you can start streaming. For a 640x480#25 700k video stream the command will be something like:
ffmpeg -f v4l2 -framerate 25 -video_size 640x480 -i /dev/video0 -c:v libx264 -b:v 700k -maxrate 700k -bufsize 700k -an -f flv rtmp://<youtube_rtmp_server/<youtube_live_stream_id>
"So is it possible for me to stream directly to YouTube on a Raspberry
Pi?"
Yes. But you're going to need to do a bit of configuring and get different hardware depending on your project needs.
For my project, a day and night doorway "security camera" that streams live to youtube, I chose a Raspberry Pi Zero W running raspbian (headless) and a camera module with auto IR switching capabilities and IR lights.
I have edited the raspbian image so all of the configurations of the wifi and camera module interfaces, code, and dependencies I need are pre-installed, so I can just flash an sd card, slap it in a pi+camera+powersupply setup and it does its thing.
So, for this answer to be helpful at all, you're going to need to do plenty of research on FFMPEG, know what it is, learn what it does, and get it installed on your board... You should be able to run a few tests getting FFMPEG to just spit out maybe a 10-second long video from your camera. I wouldn't bother reading any more of my ramblings if you have not got that far yet, because things are about to get specific.
So, your board is online, you can see it on the network, it's got internet, it's got ffmpeg, it's ready to go.
Here is the ffmpeg "stream command" I use to start the live stream:
raspivid -o - -t 0 -vf -hf -fps 60 -b 12000000 -rot 180 | ffmpeg -re -ar 44100 -ac 2 -acodec pcm_s16le -f s16le -ac 2 -i /dev/zero -i - -vcodec copy -acodec aac -ab 384k -g 17 -strict experimental -f flv rtmp://a.rtmp.youtube.com/live2/SESSION_ID
I arrived at this "stream command" above by tweaking each parameter you see, one by one, and in different combinations and I eventually got a really crisp 1080p stream with no buffering issues at all except for the occasional bit of wifi lag that comes around on my setup. You are going to need to do a ton of research into what every parameter does to get things just right and trust me it's going to be a pain figuring out what does what in the beginning. I would lurk all around StackOverflow and other resources and just plug around and see what you can get to come out of your setup when it comes to these FFMPEG commands.
To test if this "stream command" or any other you find works for you, just change SESSION_ID at the end to your stream key and run it in the console.
After you get an output you are happy with, figure out on your own how you want to trigger your camera to start streaming, if you want it to start recording as soon as the board is ready to start sending data, you accomplish this by putting your "stream command" in /etc/rc.local and it will run that command as soon as it can.
For my project, I use 18650 cells charged by solar panels as the power source so I have to be conscious about the power I use so I wrote some NodeJS program monitor just that.
Alright, that's enough talking into the wind for now. Hopefully, any of this helped someone out there, cheers.
Audio working! This worked for me from a raspberry pi 4 with an rbp v1.3 camera and cheap usb audio interface. Also gets the default audio which you can set in the alsamixer:
raspivid -o - -t 0 -vf -hf -fps 30 -b 6000000 | ffmpeg -f alsa -ac 1 -ar 44100 -i default -acodec pcm_s16le -f s16le -f h264 -i - -vcodec copy -acodec aac -ab 128k -g 60 -strict -2 -f flv rtmp://<destination/streamkey>
I am writing a small mp3 conversion tool. We upload a mp3 file & would like to convert it to a 96kbps file & a 320 kbps file. I have written the conversion script & it runs. But these files do not play.
Am i missing something?
the code i've written is:
/usr/local/bin/ffmpeg -i test.mp3 -vn -ar 441000 -ac 2 -ab 96k -f mp2 music/96/test.mp3 2>&1
Thanks!
There are two major issues here:
The audio rate you're setting (-ar 441000) is incorrect. You want 44.1 kHz (-ar 44100).
You're forcing the use of MPEG2 audio (-f mp2), which is not what you want, and is probably not supported by the player you're using either. Leave that option out entirely; the .mp3 extension on the output file will be used as a hint anyway.