ffserver is a streaming server for both audio and video. It
supports several live feeds, streaming from files and time shifting on
live feeds (you can seek to positions in the past on each live feed,
provided you specify a big enough feed storage in
ffserver receives prerecorded files or FFM streams from some
ffmpeg instance as input, then streams them over
ffserver instance will listen on some port as specified
in the configuration file. You can launch one or more instances of
ffmpeg and send one or more FFM streams to the port where
ffserver is expecting to receive them. Alternately, you can make
ffserver launch such
ffmpeg instances at startup.
Input streams are called feeds, and each one is specified by a
<Feed> section in the configuration file.
For each feed you can have different output streams in various
formats, each one specified by a
<Stream> section in the
ffserver supports an HTTP interface which exposes the current status of the server.
Simply point your browser to the address of the special status stream specified in the configuration file.
For example if you have:
<Stream status.html> Format status # Only allow local people to get the status ACL allow localhost ACL allow 192.168.0.0 192.168.255.255 </Stream>
then the server will post a page with the status information when the special stream ‘status.html’ is requested.
When properly configured and running, you can capture video and audio in real time from a suitable capture card, and stream it out over the Internet to either Windows Media Player or RealAudio player (with some restrictions).
It can also stream from files, though that is currently broken. Very often, a web server can be used to serve up the files just as well.
It can stream prerecorded video from .ffm files, though it is somewhat tricky to make it work correctly.
First, build the kit. It *really* helps to have installed LAME first. Then when
you run the ffserver ./configure, make sure that you have the
--enable-libmp3lame flag turned on.
LAME is important as it allows for streaming audio to Windows Media Player. Don’t ask why the other audio types do not work.
As a simple test, just run the following two command lines where INPUTFILE is some file which you can decode with ffmpeg:
ffserver -f doc/ffserver.conf & ffmpeg -i INPUTFILE http://localhost:8090/feed1.ffm
At this point you should be able to go to your Windows machine and fire up Windows Media Player (WMP). Go to Open URL and enter
You should (after a short delay) see video and hear audio.
WARNING: trying to stream test1.mpg doesn’t work with WMP as it tries to transfer the entire file before starting to play. The same is true of AVI files.
You should edit the ffserver.conf file to suit your needs (in terms of frame rates etc). Then install ffserver and ffmpeg, write a script to start them up, and off you go.
Maybe you didn’t install LAME, or got your ./configure statement wrong. Check the ffmpeg output to see if a line referring to MP3 is present. If not, then your configuration was incorrect. If it is, then maybe your wiring is not set up correctly. Maybe the sound card is not getting data from the right input source. Maybe you have a really awful audio interface (like I do) that only captures in stereo and also requires that one channel be flipped. If you are one of these people, then export ’AUDIO_FLIP_LEFT=1’ before starting ffmpeg.
Yes, they do.
Yes, it does. Who knows why?
Yes, it does. Any thoughts on this would be gratefully received. These differences extend to embedding WMP into a web page. [There are two object IDs that you can use: The old one, which does not play well, and the new one, which does (both tested on the same system). However, I suspect that the new one is not available unless you have installed WMP 7].
You can replay video from .ffm files that was recorded earlier. However, there are a number of caveats, including the fact that the ffserver parameters must match the original parameters used to record the file. If they do not, then ffserver deletes the file before recording into it. (Now that I write this, it seems broken).
You can fiddle with many of the codec choices and encoding parameters, and there are a bunch more parameters that you cannot control. Post a message to the mailing list if there are some ’must have’ parameters. Look in ffserver.conf for a list of the currently available controls.
It will automatically generate the ASX or RAM files that are often used in browsers. These files are actually redirections to the underlying ASF or RM file. The reason for this is that the browser often fetches the entire file before starting up the external viewer. The redirection files are very small and can be transferred quickly. [The stream itself is often ’infinite’ and thus the browser tries to download it and never finishes.]
* When you connect to a live stream, most players (WMP, RA, etc) want to buffer a certain number of seconds of material so that they can display the signal continuously. However, ffserver (by default) starts sending data in realtime. This means that there is a pause of a few seconds while the buffering is being done by the player. The good news is that this can be cured by adding a ’?buffer=5’ to the end of the URL. This means that the stream should start 5 seconds in the past – and so the first 5 seconds of the stream are sent as fast as the network will allow. It will then slow down to real time. This noticeably improves the startup experience.
You can also add a ’Preroll 15’ statement into the ffserver.conf that will add the 15 second prebuffering on all requests that do not otherwise specify a time. In addition, ffserver will skip frames until a key_frame is found. This further reduces the startup delay by not transferring data that will be discarded.
* You may want to adjust the MaxBandwidth in the ffserver.conf to limit the amount of bandwidth consumed by live streams.
It turns out that (on my machine at least) the number of frames successfully grabbed is marginally less than the number that ought to be grabbed. This means that the timestamp in the encoded data stream gets behind realtime. This means that if you say ’Preroll 10’, then when the stream gets 10 or more seconds behind, there is no Preroll left.
Fixing this requires a change in the internals of how timestamps are handled.
Yes (subject to the limitation outlined above). Also note that whenever you start ffserver, it deletes the ffm file (if any parameters have changed), thus wiping out what you had recorded before.
The format of the
?date=xxxxxx is fairly flexible. You should use one
of the following formats (the ’T’ is literal):
* YYYY-MM-DDTHH:MM:SS (localtime) * YYYY-MM-DDTHH:MM:SSZ (UTC)
You can omit the YYYY-MM-DD, and then it refers to the current day. However note that ‘?date=16:00:00’ refers to 16:00 on the current day – this may be in the future and so is unlikely to be useful.
You use this by adding the ?date= to the end of the URL for the stream. For example: ‘http://localhost:8080/test.asf?date=2002-07-26T23:05:00’.
FFM and FFM2 are formats used by ffserver. They allow storing a wide variety of video and audio streams and encoding options, and can store a moving time segment of an infinite movie or a whole movie.
FFM is version specific, and there is limited compatibility of FFM files generated by one version of ffmpeg/ffserver and another version of ffmpeg/ffserver. It may work but it is not guaranteed to work.
FFM2 is extensible while maintaining compatibility and should work between differing versions of tools. FFM2 is the default.
All the numerical options, if not specified otherwise, accept a string representing a number as input, which may be followed by one of the SI unit prefixes, for example: ’K’, ’M’, or ’G’.
If ’i’ is appended to the SI unit prefix, the complete prefix will be interpreted as a unit prefix for binary multiplies, which are based on powers of 1024 instead of powers of 1000. Appending ’B’ to the SI unit prefix multiplies the value by 8. This allows using, for example: ’KB’, ’MiB’, ’G’ and ’B’ as number suffixes.
Options which do not take arguments are boolean options, and set the corresponding value to true. They can be set to false by prefixing the option name with "no". For example using "-nofoo" will set the boolean option with name "foo" to false.
Some options are applied per-stream, e.g. bitrate or codec. Stream specifiers are used to precisely specify which stream(s) a given option belongs to.
A stream specifier is a string generally appended to the option name and
separated from it by a colon. E.g.
-codec:a:1 ac3 contains the
a:1 stream specifier, which matches the second audio stream. Therefore, it
would select the ac3 codec for the second audio stream.
A stream specifier can match several streams, so that the option is applied to all
of them. E.g. the stream specifier in
-b:a 128k matches all audio
An empty stream specifier matches all streams. For example,
-codec: copy would copy all the streams without reencoding.
Possible forms of stream specifiers are:
Matches the stream with this index. E.g.
-threads:1 4 would set the
thread count for the second stream to 4.
stream_type is one of following: ’v’ for video, ’a’ for audio, ’s’ for subtitle, ’d’ for data, and ’t’ for attachments. If stream_index is given, then it matches stream number stream_index of this type. Otherwise, it matches all streams of this type.
If stream_index is given, then it matches the stream with number stream_index in the program with the id program_id. Otherwise, it matches all streams in the program.
Matches the stream by a format-specific ID.
These options are shared amongst the ff* tools.
Show help. An optional parameter may be specified to print help about a specific item.
Possible values of arg are:
Print detailed information about the decoder named decoder_name. Use the ‘-decoders’ option to get a list of all decoders.
Print detailed information about the encoder named encoder_name. Use the ‘-encoders’ option to get a list of all encoders.
Print detailed information about the demuxer named demuxer_name. Use the ‘-formats’ option to get a list of all demuxers and muxers.
Print detailed information about the muxer named muxer_name. Use the ‘-formats’ option to get a list of all muxers and demuxers.
Print detailed information about the filter name filter_name. Use the ‘-filters’ option to get a list of all filters.
Show available formats.
Show all codecs known to libavcodec.
Note that the term ’codec’ is used throughout this documentation as a shortcut for what is more correctly called a media bitstream format.
Show available decoders.
Show all available encoders.
Show available bitstream filters.
Show available protocols.
Show available libavfilter filters.
Show available pixel formats.
Show available sample formats.
Show channel names and standard channel layouts.
Set the logging level used by the library. Adding "repeat+" indicates that repeated log output should not be compressed to the first line and the "Last message repeated n times" line will be omitted. "repeat" can also be used alone. If "repeat" is used alone, and with no prior loglevel set, the default loglevel will be used. If multiple loglevel parameters are given, using ’repeat’ will not change the loglevel. loglevel is a number or a string containing one of the following values:
Show nothing at all; be silent.
Only show fatal errors which could lead the process to crash, such as and assert failure. This is not currently used for anything.
Only show fatal errors. These are errors after which the process absolutely cannot continue after.
Show all errors, including ones which can be recovered from.
Show all warnings and errors. Any message related to possibly incorrect or unexpected events will be shown.
Show informative messages during processing. This is in addition to warnings and errors. This is the default value.
info, except more verbose.
Show everything, including debugging information.
By default the program logs to stderr, if coloring is supported by the
terminal, colors are used to mark errors and warnings. Log coloring
can be disabled setting the environment variable
NO_COLOR, or can be forced setting
the environment variable
The use of the environment variable
NO_COLOR is deprecated and
will be dropped in a following FFmpeg version.
Dump full command line and console output to a file named
program-YYYYMMDD-HHMMSS.log in the current
This file can be useful for bug reports.
It also implies
Setting the environment variable
FFREPORT to any value has the
same effect. If the value is a ’:’-separated key=value sequence, these
options will affect the report; options values must be escaped if they
contain special characters or the options delimiter ’:’ (see the
“Quoting and escaping” section in the ffmpeg-utils manual). The
following option is recognized:
set the file name to use for the report;
%p is expanded to the name
of the program,
%t is expanded to a timestamp,
%% is expanded
to a plain
Errors in parsing the environment variable are not fatal, and will not appear in the report.
Allows setting and clearing cpu flags. This option is intended for testing. Do not use it unless you know what you’re doing.
ffmpeg -cpuflags -sse+mmx ... ffmpeg -cpuflags mmx ... ffmpeg -cpuflags 0 ...
Possible flags for this option are:
Set OpenCL environment options. This option is only available when
FFmpeg has been compiled with
options must be a list of key=value option pairs separated by ’:’. See the “OpenCL Options” section in the ffmpeg-utils manual for the list of supported options.
These options are provided directly by the libavformat, libavdevice and libavcodec libraries. To see the list of available AVOptions, use the ‘-help’ option. They are separated into two categories:
These options can be set for any container, codec or device. Generic options are listed under AVFormatContext options for containers/devices and under AVCodecContext options for codecs.
These options are specific to the given container, device or codec. Private options are listed under their corresponding containers/devices/codecs.
For example to write an ID3v2.3 header instead of a default ID3v2.4 to an MP3 file, use the ‘id3v2_version’ private option of the MP3 muxer:
ffmpeg -i input.flac -id3v2_version 3 out.mp3
All codec AVOptions are obviously per-stream, so the chapter on stream specifiers applies to them
Note ‘-nooption’ syntax cannot be used for boolean AVOptions, use ‘-option 0’/‘-option 1’.
Note2 old undocumented way of specifying per-stream AVOptions by prepending v/a/s to the options name is now obsolete and will be removed soon.
Use ‘configfile’ instead of ‘/etc/ffserver.conf’.
Enable no-launch mode. This option disables all the Launch directives within the various <Stream> sections. Since ffserver will not launch any ffmpeg instances, you will have to launch them manually.
Enable debug mode. This option increases log verbosity, directs log messages to stdout.
ffserver-all, the ‘doc/ffserver.conf’ example, ffmpeg, ffplay, ffprobe, ffmpeg-utils, ffmpeg-scaler, ffmpeg-resampler, ffmpeg-codecs, ffmpeg-bitstream-filters, ffmpeg-formats, ffmpeg-devices, ffmpeg-protocols, ffmpeg-filters
The FFmpeg developers.
For details about the authorship, see the Git history of the project
(git://source.ffmpeg.org/ffmpeg), e.g. by typing the command
git log in the FFmpeg source directory, or browsing the
online repository at http://source.ffmpeg.org.
Maintainers for the specific components are listed in the file ‘MAINTAINERS’ in the source code tree.