Merge branch 'refs/heads/t/para_play_fixes'
[paraslash.git] / web / manual.m4
1 dnl To generate the html version, execute
2 dnl m4 web/manual.m4 | grutatxt --toc
3
4 define(`LOCAL_LINK_NAME', `translit(`$1', `A-Z
5 ', `a-z__')')
6 define(`REMOVE_NEWLINE', `translit(`$1',`
7 ', ` ')')
8
9 define(`REFERENCE', ./``#''`LOCAL_LINK_NAME($1)' (`REMOVE_NEWLINE($2)'))
10 define(`XREFERENCE', `$1' (`REMOVE_NEWLINE($2)'))
11 define(`EMPH', ``_''`REMOVE_NEWLINE($1)'``_'')
12
13 Paraslash user manual
14 =====================
15
16 This document describes how to install, configure and use the paraslash
17 network audio streaming system. Most chapters start with a chapter
18 overview and conclude with an example section. We try to focus on
19 general concepts and on the interaction of the various pieces of the
20 paraslash package. Hence this user manual is not meant as a replacement
21 for the manual pages that describe all command line options of each
22 paraslash executable.
23
24 ------------
25 Introduction
26 ------------
27
28 In this chapter we give an REFERENCE(Overview, overview) of the
29 interactions of the two main programs contained in the paraslash
30 package, followed by REFERENCE(The paraslash executables, brief
31 descriptions) of all executables.
32
33 Overview
34 ~~~~~~~~
35
36 The core functionality of the para suite is provided by two main
37 executables, para_server and para_audiod. The former maintains a
38 database of audio files and streams these files to para_audiod which
39 receives and plays the stream.
40
41 In a typical setting, both para_server and para_audiod act as
42 background daemons whose functionality is controlled by client
43 programs: the para_audioc client controls para_audiod over a local
44 socket while the para_client program connects to para_server over a
45 local or remote networking connection.
46
47 Typically, these two daemons run on different hosts but a local setup
48 is also possible.
49
50 A simplified picture of a typical setup is as follows
51 <<
52 <pre>
53 server_host client_host
54 ~~~~~~~~~~~ ~~~~~~~~~~~
55
56 +-----------+ audio stream +-----------+
57 |para_server| -----------------------------> |para_audiod|
58 +-----------+ +-----------+
59 ^ ^
60 | |
61 | | connect
62 | |
63 | |
64 | +-----------+
65 | |para_audioc|
66 | +-----------+
67 |
68 |
69 | connect +-----------+
70 +-------------------------------------- |para_client|
71 +-----------+
72 </pre>
73 >>
74
75 The paraslash executables
76 ~~~~~~~~~~~~~~~~~~~~~~~~~
77
78 *para_server*
79
80 para_server streams binary audio data (MP3, ...) over local and/or
81 remote networks. It listens on a TCP port and accepts commands such
82 as play, stop, pause, next from authenticated clients. There are
83 many more commands though, see the man page of para_server for a
84 description of all commands.
85
86 It supports three built-in network streaming protocols
87 (senders/receivers): HTTP, DCCP, or UDP. This is explained in more
88 detail in the section on REFERENCE(Networking, networking).
89
90 The built-in audio file selector of paraslash is used to manage your
91 audio files. It maintains statistics on the usage of all available
92 audio files such as last-played time, and the number of times each
93 file was selected.
94
95 Additional information may be added to the database to allow
96 fine-grained selection based on various properties of the audio file,
97 including information found in (ID3) tags. However, old-fashioned
98 playlists are also supported.
99
100 It is also possible to store images (album covers) and lyrics in the
101 database and associate these to the corresponding audio files.
102
103 The section on the REFERENCE(The audio file selector, audio file
104 selector) discusses this topic.
105
106
107 *para_client*
108
109 The client program to connect to para_server. paraslash commands
110 are sent to para_server and the response is dumped to STDOUT. This
111 can be used by any scripting language to produce user interfaces with
112 little programming effort.
113
114 All connections between para_server and para_client are encrypted
115 with a symmetric session key. For each user of paraslash you must
116 create a public/secret RSA key pair for authentication.
117
118 If para_client is started without non-option arguments, an interactive
119 session (shell) is started. Command history and command completion are
120 supported through libreadline.
121
122 *para_audiod*
123
124 The local daemon that collects information from para_server.
125
126 It runs on the client side and connects to para_server. As soon as
127 para_server announces the availability of an audio stream, para_audiod
128 starts an appropriate receiver, any number of filters and a paraslash
129 writer to play the stream.
130
131 Moreover, para_audiod listens on a local socket and sends status
132 information about para_server and para_audiod to local clients on
133 request. Access via this local socket may be restricted by using Unix
134 socket credentials, if available.
135
136
137 *para_audioc*
138
139 The client program which talks to para_audiod. Used to control
140 para_audiod, to receive status info, or to grab the stream at any
141 point of the decoding process. Like para_client, para_audioc supports
142 interactive sessions on systems with libreadline.
143
144 *para_recv*
145
146 A command line HTTP/DCCP/UDP stream grabber. The http mode is
147 compatible with arbitrary HTTP streaming sources (e.g. icecast).
148 In addition to the three network streaming modes, para_recv can also
149 operate in local (afh) mode. In this mode it writes the content of
150 an audio file on the local file system in complete chunks to stdout,
151 optionally 'just in time'. This allows to cut an audio file without
152 first decoding it, and it enables third-party software which is unaware
153 of the particular audio format to send complete frames in real time.
154
155 *para_filter*
156
157 A filter program that reads from STDIN and writes to STDOUT.
158 Like para_recv, this is an atomic building block which can be used to
159 assemble higher-level audio receiving facilities. It combines several
160 different functionalities in one tool: decoders for multiple audio
161 formats and a number of processing filters, among these a normalizer
162 for audio volume.
163
164 *para_afh*
165
166 A small stand-alone program that prints tech info about the given
167 audio file to STDOUT. It can be instructed to print a "chunk table",
168 an array of offsets within the audio file.
169
170 *para_write*
171
172 A modular audio stream writer. It supports a simple file writer
173 output plug-in and optional WAV/raw players for ALSA (Linux) and for
174 coreaudio (Mac OS). para_write can also be used as a stand-alone WAV
175 or raw audio player.
176
177 *para_play*
178
179 A command line audio player.
180
181 *para_gui*
182
183 Curses-based gui that presents status information obtained in a curses
184 window. Appearance can be customized via themes. para_gui provides
185 key-bindings for the most common server commands and new key-bindings
186 can be added easily.
187
188
189 *para_fade*
190
191 An alarm clock and volume-fader for OSS and ALSA.
192
193 -----------
194 Quick start
195 -----------
196
197 This chapter lists the REFERENCE(Requirements, necessary software)
198 that must be installed to compile the paraslash package, describes
199 how to REFERENCE(Installation, compile and install) the paraslash
200 source code and the steps that have to be performed in order to
201 REFERENCE(Quick start, set up) a typical server and client.
202
203 Requirements
204 ~~~~~~~~~~~~
205 For the impatient:
206
207 git clone git://git.tuebingen.mpg.de/osl
208 cd osl && make && sudo make install && sudo ldconfig
209 sudo apt-get install autoconf libssl-dev help2man gengetopt \
210 libmad0-dev libid3tag0-dev libasound2-dev libvorbis-dev \
211 libfaad-dev libspeex-dev libFLAC-dev libsamplerate-dev \
212 libasound2-dev libao-dev libreadline-dev libncurses-dev \
213 libopus-dev
214
215 Detailed description: In any case you'll need
216
217 - XREFERENCE(http://people.tuebingen.mpg.de/maan/osl/, libosl).
218 The _object storage layer_ library is used by para_server. To
219 clone the source code repository, execute
220
221 git clone git://git.tuebingen.mpg.de/osl
222
223 - XREFERENCE(ftp://ftp.gnu.org/pub/gnu/gcc, gcc) or
224 XREFERENCE(http://clang.llvm.org, clang). All gcc versions
225 >= 3.3 are currently supported. Clang version 1.1 or newer
226 should work as well.
227
228 - XREFERENCE(ftp://ftp.gnu.org/pub/gnu/make, gnu make) is
229 also shipped with the disto. On BSD systems the gnu make
230 executable is often called gmake.
231
232 - XREFERENCE(ftp://ftp.gnu.org/pub/gnu/bash, bash). Some
233 scripts which run during compilation require the EMPH(Bourne
234 again shell). It is most likely already installed.
235
236 - XREFERENCE(ftp://ftp.gnu.org/pub/gnu/gengetopt/, gengetopt)
237 is needed to generate the C code for the command line parsers
238 of all paraslash executables.
239
240 - XREFERENCE(ftp://ftp.gnu.org/pub/gnu/help2man, help2man)
241 is used to create the man pages.
242
243 Optional:
244
245 - XREFERENCE(http://www.openssl.org/, openssl) or
246 XREFERENCE(ftp://ftp.gnupg.org/gcrypt/libgcrypt/, libgcrypt).
247 At least one of these two libraries is needed as the backend
248 for cryptographic routines on both the server and the client
249 side. Both openssl and libgcrypt are usually shipped with the
250 distro, but you might have to install the development package
251 (libssl-dev or libgcrypt-dev on debian systems) as well.
252
253 - XREFERENCE(http://www.underbit.com/products/mad/, libmad).
254 To compile in MP3 support for paraslash, the development
255 package must be installed. It is called libmad0-dev on
256 debian-based systems. Note that libmad is not necessary on
257 the server side, i.e. for sending MP3 files.
258
259 - XREFERENCE(http://www.underbit.com/products/mad/,
260 libid3tag). For version-2 ID3 tag support, you'll need
261 the libid3tag development package libid3tag0-dev. Without
262 libid3tag, only version-1 tags are recognized. The mp3 tagger
263 also needs this library for modifying (id3v1 and id3v2) tags.
264
265 - XREFERENCE(http://www.xiph.org/downloads/, ogg vorbis).
266 For ogg vorbis streams you'll need libogg, libvorbis,
267 libvorbisfile. The corresponding Debian packages are called
268 libogg-dev and libvorbis-dev.
269
270 - XREFERENCE(http://www.audiocoding.com/, libfaad). For aac
271 files (m4a) you'll need libfaad (libfaad-dev).
272
273 - XREFERENCE(http://www.speex.org/, speex). In order to stream
274 or decode speex files, libspeex (libspeex-dev) is required.
275
276 - XREFERENCE(http://flac.sourceforge.net/, flac). To stream
277 or decode files encoded with the _Free Lossless Audio Codec_,
278 libFLAC (libFLAC-dev) must be installed.
279
280 - XREFERENCE(http://www.mega-nerd.com/SRC/index.html,
281 libsamplerate). The resample filter will only be compiled if
282 this library is installed. Debian package: libsamplerate-dev.
283
284 - XREFERENCE(ftp://ftp.alsa-project.org/pub/lib/, alsa-lib). On
285 Linux, you'll need to have ALSA's development package
286 libasound2-dev installed.
287
288 - XREFERENCE(http://downloads.xiph.org/releases/ao/,
289 libao). Needed to build the ao writer (ESD, PulseAudio,...).
290 Debian package: libao-dev.
291
292 - XREFERENCE(ftp://ftp.gnu.org/pub/gnu/ncurses, curses). Needed
293 for para_gui. Debian package: libncurses-dev.
294
295 - XREFERENCE(http://cnswww.cns.cwru.edu/php/chet/readline/rltop.html,
296 GNU Readline). If this library (libreadline-dev) is installed,
297 para_client, para_audioc and para_play support interactive
298 sessions.
299
300 Installation
301 ~~~~~~~~~~~~
302 To build the sources from a tarball, execute
303
304 ./configure && make
305
306 To build from git or a gitweb snapshot, run this command instead:
307
308 ./autogen.sh
309
310 There should be no errors but probably some warnings about missing
311 packages which usually implies that not all audio formats will be
312 supported. If headers or libs are installed at unusual locations you
313 might need to tell the configure script where to find them. Try
314
315 ./configure --help
316
317 to see a list of options. If the paraslash package was compiled
318 successfully, execute (optionally)
319
320 make test
321
322 to run the paraslash test suite. If all tests pass, execute as root
323
324 make install
325
326 to install executables under /usr/local/bin and the man pages under
327 /usr/local/man.
328
329 Configuration
330 ~~~~~~~~~~~~~
331
332 *Step 1*: Create a paraslash user
333
334 In order to control para_server at runtime you must create a paraslash
335 user. As authentication is based on the RSA crypto system you'll have
336 to create an RSA key pair. If you already have a user and an RSA key
337 pair, you may skip this step.
338
339 In this section we'll assume a typical setup: You would like to run
340 para_server on some host called server_host as user foo, and you want
341 to connect to para_server from another machine called client_host as
342 user bar.
343
344 As foo@server_host, create ~/.paraslash/server.users by typing the
345 following commands:
346
347 user=bar
348 target=~/.paraslash/server.users
349 key=~/.paraslash/id_rsa.pub.$user
350 perms=AFS_READ,AFS_WRITE,VSS_READ,VSS_WRITE
351 mkdir -p ~/.paraslash
352 echo "user $user $key $perms" >> $target
353
354 Next, change to the "bar" account on client_host and generate the
355 key pair with the commands
356
357 ssh-keygen -q -t rsa -b 2048 -N '' -f $key
358
359 This generates the two files id_rsa and id_rsa.pub in ~/.ssh. Note
360 that para_server won't accept keys shorter than 2048 bits. Moreover,
361 para_client rejects private keys which are world-readable.
362
363 para_server only needs to know the public key of the key pair just
364 created. Copy this public key to server_host:
365
366 src=~/.ssh/id_rsa.pub
367 dest=.paraslash/id_rsa.pub.$LOGNAME
368 scp $src foo@server_host:$dest
369
370 Finally, tell para_client to connect to server_host:
371
372 conf=~/.paraslash/client.conf
373 echo 'hostname server_host' > $conf
374
375
376 *Step 2*: Start para_server
377
378 For this first try, we'll use the info loglevel to make the output
379 of para_server more verbose.
380
381 para_server -l info
382
383 Now you can use para_client to connect to the server and issue
384 commands. Open a new shell as bar@client_host and try
385
386 para_client help
387 para_client si
388
389 to retrieve the list of available commands and some server info.
390 Don't proceed if this doesn't work.
391
392 *Step 3*: Create and populate the database
393
394 An empty database is created with
395
396 para_client init
397
398 This initializes a couple of empty tables under
399 ~/.paraslash/afs_database-0.4. You normally don't need to look at these
400 tables, but it's good to know that you can start from scratch with
401
402 rm -rf ~/.paraslash/afs_database-0.4
403
404 in case something went wrong.
405
406 Next, you need to add some audio files to that database so that
407 para_server knows about them. Choose an absolute path to a directory
408 containing some audio files and add them to the audio file table:
409
410 para_client add /my/mp3/dir
411
412 This might take a while, so it is a good idea to start with a directory
413 containing not too many files. Note that the table only contains data
414 about the audio files found, not the files themselves.
415
416 You may print the list of all known audio files with
417
418 para_client ls
419
420 *Step 4*: Configure para_audiod
421
422 We will have to tell para_audiod that it should receive the audio
423 stream from server_host via http:
424
425 para_audiod -l info -r '.:http -i server_host'
426
427 You should now be able to listen to the audio stream once para_server
428 starts streaming. To activate streaming, execute
429
430 para_client play
431
432 Since no playlist has been specified yet, the "dummy" mode which
433 selects all known audio files is activated automatically. See the
434 section on the REFERENCE(The audio file selector, audio file selector)
435 for how to use playlists and moods to specify which files should be
436 streamed in which order.
437
438 *Troubleshooting*
439
440 If you receive a socket related error on server or audiod startup,
441 make sure you have write permissions to the /var/paraslash directory:
442
443 sudo chown $LOGNAME /var/paraslash
444
445 Alternatively, use the --afs-socket (para_server) or --socket
446 (para_audiod) option to specify a different socket pathname.
447
448 To identify streaming problems try to receive, decode and play the
449 stream manually using para_recv, para_filter and para_write as follows.
450 For simplicity we assume that you're running Linux/ALSA and that only
451 MP3 files have been added to the database.
452
453 para_recv -r 'http -i server_host' > file.mp3
454 # (interrupt with CTRL+C after a few seconds)
455 ls -l file.mp3 # should not be empty
456 para_filter -f mp3dec -f wav < file.mp3 > file.wav
457 ls -l file.wav # should be much bigger than file.mp3
458 para_write -w alsa < file.wav
459
460 Double check what is logged by para_server and use the --loglevel
461 option of para_recv, para_filter and para_write to increase verbosity.
462
463 ---------------
464 User management
465 ---------------
466
467 para_server uses a challenge-response mechanism to authenticate
468 requests from incoming connections, similar to ssh's public key
469 authentication method. Authenticated connections are encrypted using
470 a stream cipher, either RC4 or AES in integer counter mode.
471
472 In this chapter we briefly describe RSA, RC4 and AES, and sketch the
473 REFERENCE(Client-server authentication, authentication handshake)
474 between para_client and para_server. User management is discussed
475 in the section on REFERENCE(The user_list file, the user_list file).
476 These sections are all about communication between the client and the
477 server. Connecting para_audiod is a different matter and is described
478 in a REFERENCE(Connecting para_audiod, separate section).
479
480
481
482 RSA, RC4, AES
483 ~~~~~~~~~~~~~
484
485 RSA is an asymmetric block cipher which is used in many applications,
486 including ssh and gpg. An RSA key consists in fact of two keys,
487 called the public key and the private key. A message can be encrypted
488 with either key and only the counterpart of that key can decrypt
489 the message. While RSA can be used for both signing and encrypting
490 a message, paraslash uses RSA only for the latter purpose. The
491 RSA public key encryption and signatures algorithms are defined in
492 detail in RFC 2437.
493
494 RC4 is a stream cipher, i.e. the input is XORed with a pseudo-random
495 key stream to produce the output. Decryption uses the same function
496 calls as encryption. While RC4 supports variable key lengths,
497 paraslash uses a fixed length of 256 bits, which is considered a
498 strong encryption by today's standards. Since the same key must never
499 be used twice, a different, randomly-generated key is used for every
500 new connection.
501
502 AES, the advanced encryption standard, is a well-known symmetric block
503 cipher, i.e. a transformation operating on fixed-length blocks which
504 is determined by a single key for both encryption and decryption. Any
505 block cipher can be turned into a stream cipher by generating
506 a pseudo-random key stream by encrypting successive values of a
507 counter. The AES_CTR128 stream cipher used in paraslash is obtained
508 in this way from the AES block cipher with a 128 bit block size.
509
510
511 Client-server authentication
512 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
513
514 The authentication handshake between para_client and para_server goes
515 as follows:
516
517 - para_client connects to para_server and sends an
518 authentication request for a user. It does so by connecting
519 to TCP port 2990 of the server host. This port is called the
520 para_server _control port_.
521
522 - para_server accepts the connection and forks a child process
523 which handles the incoming request. The parent process keeps
524 listening on the control port while the child process (also
525 called para_server below) continues as follows.
526
527 - para_server loads the RSA public key of that user, fills a
528 fixed-length buffer with random bytes, encrypts that buffer
529 using the public key and sends the encrypted buffer to the
530 client. The first part of the buffer is the challenge which
531 is used for authentication while the second part is the
532 session key.
533
534 - para_client receives the encrypted buffer and decrypts it
535 with the user's private key, thereby obtaining the challenge
536 buffer and the session key. It sends the SHA1 hash value of
537 the challenge back to para_server and stores the session key
538 for further use.
539
540 - para_server also computes the SHA1 hash of the challenge
541 and compares it against what was sent back by the client.
542
543 - If the two hashes do not match, the authentication has
544 failed and para_server closes the connection.
545
546 - Otherwise the user is considered authenticated and the client
547 is allowed to proceed by sending a command to be executed. From
548 this point on the communication is encrypted using the stream
549 cipher with the session key known to both peers.
550
551 paraslash relies on the quality of the pseudo-random bytes provided
552 by the crypto library (openssl or libgcrypt), on the security of
553 the implementation of the RSA, RC4 and AES crypto routines and on the
554 infeasibility to invert the SHA1 function.
555
556 Neither para_server or para_client create RSA keys on their own. This
557 has to be done once for each user as sketched in REFERENCE(Quick start,
558 Quick start) and discussed in more detail REFERENCE(The user_list
559 file, below).
560
561 The user_list file
562 ~~~~~~~~~~~~~~~~~~
563
564 At startup para_server reads the user list file which contains one
565 line per user. The default location of the user list file may be
566 changed with the --user-list option.
567
568 There should be at least one user in this file. Each user must have
569 an RSA key pair. The public part of the key is needed by para_server
570 while the private key is needed by para_client. Each line of the
571 user list file must be of the form
572
573 user <username> <key> <perms>
574
575 where _username_ is an arbitrary string (usually the user's login
576 name), _key_ is the full path to that user's public RSA key, and
577 _perms_ is a comma-separated list of zero or more of the following
578 permission bits:
579
580 +---------------------------------------------------------+
581 | AFS_READ | read the contents of the databases |
582 +-----------+---------------------------------------------+
583 | AFS_WRITE | change database contents |
584 +-----------+---------------------------------------------+
585 | VSS_READ | obtain information about the current stream |
586 +-----------+---------------------------------------------+
587 | VSS_WRITE | change the current stream |
588 +---------------------------------------------------------+
589
590 The permission bits specify which commands the user is allowed to
591 execute. The output of
592
593 para_client help
594
595 contains in the third column the permissions needed to execute the
596 command.
597
598 It is possible to make para_server reread the user_list file by
599 executing the paraslash "hup" command or by sending SIGHUP to the
600 PID of para_server.
601
602
603 Connecting para_audiod
604 ~~~~~~~~~~~~~~~~~~~~~~
605
606 para_audiod listens on a Unix domain socket. Those sockets are
607 for local communication only, so only local users can connect to
608 para_audiod. The default is to let any user connect but this can be
609 restricted on platforms that support UNIX socket credentials which
610 allow para_audiod to obtain the Unix credentials of the connecting
611 process.
612
613 Use para_audiod's --user-allow option to allow connections only for
614 a limited set of users.
615
616 -----------------------
617 The audio file selector
618 -----------------------
619
620 paraslash comes with a sophisticated audio file selector (AFS),
621 whose main task is to determine which file to stream next, based on
622 information on the audio files stored in a database. It communicates
623 also with para_client whenever an AFS command is executed, for example
624 to answer a database query.
625
626 Besides the traditional playlists, AFS supports audio file selection
627 based on _moods_ which act as a filter that limits the set of all
628 known audio files to those which satisfy certain criteria. It also
629 maintains tables containing images (e.g. album cover art) and lyrics
630 that can be associated with one or more audio files.
631
632 AFS uses XREFERENCE(http://people.tuebingen.mpg.de/maan/osl/, libosl), the
633 object storage layer library, as the backend library for storing
634 information on audio files, playlists, etc. This library offers
635 functionality similar to a relational database, but is much more
636 lightweight than a full database backend.
637
638 In this chapter we sketch the setup of the REFERENCE(The AFS process,
639 AFS process) during server startup and proceed with the description
640 of the REFERENCE(Database layout, layout) of the various database
641 tables. The section on REFERENCE(Playlists and moods, playlists
642 and moods) explains these two audio file selection mechanisms
643 in detail and contains pratical examples. The way REFERENCE(File
644 renames and content changes, file renames and content changes) are
645 detected is discussed briefly before the REFERENCE(Troubleshooting,
646 Troubleshooting) section concludes the chapter.
647
648 The AFS process
649 ~~~~~~~~~~~~~~~
650
651 On startup, para_server forks to create the AFS process which opens
652 the OSL database tables. The server process communicates with the
653 AFS process via pipes and shared memory. Usually, the AFS process
654 awakes only briefly whenever the current audio file changes. The AFS
655 process determines the next audio file, opens it, verifies it has
656 not been changed since it was added to the database and passes the
657 open file descriptor to the server process, along with audio file
658 meta-data such as file name, duration, audio format and so on. The
659 server process then starts to stream the audio file.
660
661 The AFS process also accepts connections from local clients via
662 a well-known socket. However, only child processes of para_server
663 may connect through this socket. All server commands that have the
664 AFS_READ or AFS_WRITE permission bits use this mechanism to query or
665 change the database.
666
667 Database layout
668 ~~~~~~~~~~~~~~~
669
670 *The audio file table*
671
672 This is the most important and usually also the largest table of the
673 AFS database. It contains the information needed to stream each audio
674 file. In particular the following data is stored for each audio file.
675
676 - SHA1 hash value of the audio file contents. This is computed
677 once when the file is added to the database. Whenever AFS
678 selects this audio file for streaming the hash value is
679 recomputed and checked against the value stored in the
680 database to detect content changes.
681
682 - The time when this audio file was last played.
683
684 - The number of times the file has been played so far.
685
686 - The attribute bitmask.
687
688 - The image id which describes the image associated with this
689 audio file.
690
691 - The lyrics id which describes the lyrics associated with
692 this audio file.
693
694 - The audio format id (MP3, OGG, ...).
695
696 - An amplification value that can be used by the amplification
697 filter to pre-amplify the decoded audio stream.
698
699 - The chunk table. It describes the location and the timing
700 of the building blocks of the audio file. This is used by
701 para_server to send chunks of the file at appropriate times.
702
703 - The duration of the audio file.
704
705 - Tag information contained in the audio file (ID3 tags,
706 Vorbis comments, ...).
707
708 - The number of channels
709
710 - The encoding bitrate.
711
712 - The sampling frequency.
713
714 To add or refresh the data contained in the audio file table, the _add_
715 command is used. It takes the full path of either an audio file or a
716 directory. In the latter case, the directory is traversed recursively
717 and all files which are recognized as valid audio files are added to
718 the database.
719
720 *The attribute table*
721
722 The attribute table contains two columns, _name_ and _bitnum_. An
723 attribute is simply a name for a certain bit number in the attribute
724 bitmask of the audio file table.
725
726 Each of the 64 bits of the attribute bitmask can be set for each
727 audio file individually. Hence up to 64 different attributes may be
728 defined. For example, "pop", "rock", "blues", "jazz", "instrumental",
729 "german_lyrics", "speech", whatever. You are free to choose as
730 many attributes as you like and there are no naming restrictions
731 for attributes.
732
733 A new attribute "test" is created by
734
735 para_client addatt test
736 and
737 para_client lsatt
738
739 lists all available attributes. You can set the "test" attribute for
740 an audio file by executing
741
742 para_client setatt test+ /path/to/the/audio/file
743
744 Similarly, the "test" bit can be removed from an audio file with
745
746 para_client setatt test- /path/to/the/audio/file
747
748 Instead of a path you may use a shell wildcard pattern. The attribute
749 is applied to all audio files matching this pattern:
750
751 para_client setatt test+ '/test/directory/*'
752
753 The command
754
755 para_client -- ls -l=v
756
757 gives you a verbose listing of your audio files also showing which
758 attributes are set.
759
760 In case you wonder why the double-dash in the above command is needed:
761 It tells para_client to not interpret the options after the dashes. If
762 you find this annoying, just say
763
764 alias para='para_client --'
765
766 and be happy. In what follows we shall use this alias.
767
768 The "test" attribute can be dropped from the database with
769
770 para rmatt test
771
772 Read the output of
773
774 para help ls
775 para help setatt
776
777 for more information and a complete list of command line options to
778 these commands.
779
780 *Blob tables*
781
782 The image, lyrics, moods and playlists tables are all blob tables.
783 Blob tables consist of three columns each: The identifier which is
784 a positive non-negative number that is auto-incremented, the name
785 (an arbitrary string) and the content (the blob).
786
787 All blob tables support the same set of actions: cat, ls, mv, rm
788 and add. Of course, _add_ is used for adding new blobs to the table
789 while the other actions have the same meaning as the corresponding
790 Unix commands. The paraslash commands to perform these actions are
791 constructed as the concatenation of the table name and the action. For
792 example addimg, catimg, lsimg, mvimg, rmimg are the commands that
793 manipulate or query the image table.
794
795 The add variant of these commands is special as these commands read
796 the blob contents from stdin. To add an image to the image table the
797 command
798
799 para addimg image_name < file.jpg
800
801 can be used.
802
803 Note that the images and lyrics are not interpreted at all, and also
804 the playlist and the mood blobs are only investigated when the mood
805 or playlist is activated with the select command.
806
807 *The score table*
808
809 Unlike all other tables the contents of the score table remain in
810 memory and are never stored on disk. The score table contains two
811 columns: The SHA1 hash value (of an audio file) and its current
812 score.
813
814 However, only those files which are admissible for the current mood
815 or playlist are contained in the score table. The audio file selector
816 always chooses the row with the highest score as the file to stream
817 next. While doing so, it computes the new score and updates the
818 last_played and the num_played fields in the audio file table.
819
820 The score table is recomputed by the select command which loads a
821 mood or playlist. Audio files are chosen for streaming from the rows
822 of the score table on a highest-score-first basis.
823
824
825 Playlists and moods
826 ~~~~~~~~~~~~~~~~~~~
827
828 Playlists and moods offer two different ways of specifying the set of
829 admissible files. A playlist in itself describes a set of admissible
830 files. A mood, in contrast, describes the set of admissible files in
831 terms of attributes and other type of information available in the
832 audio file table. As an example, a mood can define a filename pattern,
833 which is then matched against the names of audio files in the table.
834
835 *Playlists*
836
837 Playlists are accommodated in the playlist table of the afs database,
838 using the aforementioned blob format for tables. A new playlist is
839 created with the addpl command by specifying the full (absolute)
840 paths of all desired audio files, separated by newlines. Example:
841
842 find /my/mp3/dir -name "*.mp3" | para addpl my_playlist
843
844 If _my_playlist_ already exists it is overwritten. To activate the
845 new playlist, execute
846
847 para select p/my_playlist
848
849 The audio file selector will assign scores to each entry of the list,
850 in descending order so that files will be selected in order. If a
851 file could not be opened for streaming, its entry is removed from
852 the score table (but not from the playlist).
853
854 *Moods*
855
856 A mood consists of a unique name and its *mood definition*, which is
857 a set of *mood lines* containing expressions in terms of attributes
858 and other data contained in the database.
859
860 At any time at most one mood can be *active* which means that
861 para_server is going to select only files from that subset of
862 admissible files.
863
864 So in order to create a mood definition one has to write a set of
865 mood lines. Mood lines come in three flavours: Accept lines, deny
866 lines and score lines.
867
868 The general syntax of the three types of mood lines is
869
870
871 accept [with score <score>] [if] [not] <mood_method> [options]
872 deny [with score <score>] [if] [not] <mood_method> [options]
873 score <score> [if] [not] <mood_method> [options]
874
875
876 Here <score> is either an integer or the string "random" which assigns
877 a random score to all matching files. The score value changes the
878 order in which admissible files are going to be selected, but is of
879 minor importance for this introduction.
880
881 So we concentrate on the first two forms, i.e. accept and deny
882 lines. As usual, everything in square brackets is optional, i.e.
883 accept/deny lines take the following form when ignoring scores:
884
885 accept [if] [not] <mood_method> [options]
886
887 and analogously for the deny case. The "if" keyword is only syntactic
888 sugar and has no function. The "not" keyword just inverts the result,
889 so the essence of a mood line is the mood method part and the options
890 following thereafter.
891
892 A *mood method* is realized as a function which takes an audio file
893 and computes a number from the data contained in the database.
894 If this number is non-negative, we say the file *matches* the mood
895 method. The file matches the full mood line if it either
896
897 - matches the mood method and the "not" keyword is not given,
898 or
899 - does not match the mood method, but the "not" keyword is given.
900
901 The set of admissible files for the whole mood is now defined as those
902 files which match at least one accept mood line, but no deny mood line.
903 More formally, an audio file F is admissible if and only if
904
905 (F ~ AL1 or F ~ AL2...) and not (F ~ DL1 or F ~ DN2 ...)
906
907 where AL1, AL2... are the accept lines, DL1, DL2... are the deny
908 lines and "~" means "matches".
909
910 The cases where no mood lines of accept/deny type are defined need
911 special treatment:
912
913 - Neither accept nor deny lines: This treats all files as
914 admissible (in fact, that is the definition of the dummy mood
915 which is activated automatically if no moods are available).
916
917 - Only accept lines: A file is admissible iff it matches at
918 least one accept line:
919
920 F ~ AL1 or F ~ AL2 or ...
921
922 - Only deny lines: A file is admissible iff it matches no
923 deny line:
924
925 not (F ~ DL1 or F ~ DN2 ...)
926
927
928
929 *List of mood_methods*
930
931 no_attributes_set
932
933 Takes no arguments and matches an audio file if and only if no
934 attributes are set.
935
936 is_set <attribute_name>
937
938 Takes the name of an attribute and matches iff that attribute is set.
939
940 path_matches <pattern>
941
942 Takes a filename pattern and matches iff the path of the audio file
943 matches the pattern.
944
945 artist_matches <pattern>
946 album_matches <pattern>
947 title_matches <pattern>
948 comment_matches <pattern>
949
950 Takes an extended regular expression and matches iff the text of the
951 corresponding tag of the audio file matches the pattern. If the tag
952 is not set, the empty string is matched against the pattern.
953
954 year ~ <num>
955 bitrate ~ <num>
956 frequency ~ <num>
957 channels ~ <num>
958 num_played ~ <num>
959
960 Takes a comparator ~ of the set {<, =, <=, >, >=, !=} and a number
961 <num>. Matches an audio file iff the condition <val> ~ <num> is
962 satisfied where val is the corresponding value of the audio file
963 (value of the year tag, bitrate in kbit/s, frequency in Hz, channel
964 count, play count).
965
966 The year tag is special as its value is undefined if the audio file
967 has no year tag or the content of the year tag is not a number. Such
968 audio files never match. Another difference is the special treatment
969 if the year tag is a two-digit number. In this case either 1900 or
970 2000 is added to the tag value, depending on whether the number is
971 greater than 2000 plus the current year.
972
973
974 *Mood usage*
975
976 To create a new mood called "my_mood", write its definition into
977 some temporary file, say "tmpfile", and add it to the mood table
978 by executing
979
980 para addmood my_mood < tmpfile
981
982 If the mood definition is really short, you may just pipe it to the
983 client instead of using temporary files. Like this:
984
985 echo "$MOOD_DEFINITION" | para addmood my_mood
986
987 There is no need to keep the temporary file since you can always use
988 the catmood command to get it back:
989
990 para catmood my_mood
991
992 A mood can be activated by executing
993
994 para select m/my_mood
995
996 Once active, the list of admissible files is shown by the ls command
997 if the "-a" switch is given:
998
999 para ls -a
1000
1001
1002 *Example mood definition*
1003
1004 Suppose you have defined attributes "punk" and "rock" and want to define
1005 a mood containing only Punk-Rock songs. That is, an audio file should be
1006 admissible if and only if both attributes are set. Since
1007
1008 punk and rock
1009
1010 is obviously the same as
1011
1012 not (not punk or not rock)
1013
1014 (de Morgan's rule), a mood definition that selects only Punk-Rock
1015 songs is
1016
1017 deny if not is_set punk
1018 deny if not is_set rock
1019
1020
1021
1022 File renames and content changes
1023 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1024
1025 Since the audio file selector knows the SHA1 of each audio file that
1026 has been added to the afs database, it recognizes if the content of
1027 a file has changed, e.g. because an ID3 tag was added or modified.
1028 Also, if a file has been renamed or moved to a different location,
1029 afs will detect that an entry with the same hash value already exists
1030 in the audio file table.
1031
1032 In both cases it is enough to just re-add the new file. In the
1033 first case (file content changed), the audio table is updated, while
1034 metadata such as the num_played and last_played fields, as well as
1035 the attributes, remain unchanged. In the other case, when the file
1036 is moved or renamed, only the path information is updated, all other
1037 data remains as before.
1038
1039 It is possible to change the behaviour of the add command by using the
1040 "-l" (lazy add) or the "-f" (force add) option.
1041
1042 Troubleshooting
1043 ~~~~~~~~~~~~~~~
1044
1045 Use the debug loglevel (-l debug) to show debugging info. All paraslash
1046 executables have a brief online help which is displayed when -h is
1047 given. The --detailed-help option prints the full help text.
1048
1049 If para_server crashed or was killed by SIGKILL (signal 9), it
1050 may refuse to start again because of "dirty osl tables". In this
1051 case you'll have to run the oslfsck program of libosl to fix your
1052 database:
1053
1054 oslfsck -fd ~/.paraslash/afs_database-0.4
1055
1056 However, make sure para_server isn't running before executing oslfsck.
1057
1058 If you don't mind to recreate your database you can start
1059 from scratch by removing the entire database directory, i.e.
1060
1061 rm -rf ~/.paraslash/afs_database-0.4
1062
1063 Be aware that this removes all attribute definitions, all playlists
1064 and all mood definitions and requires to re-initialize the tables.
1065
1066 Although oslfsck fixes inconsistencies in database tables it doesn't
1067 care about the table contents. To check for invalid table contents, use
1068
1069 para_client check
1070
1071 This prints out references to missing audio files as well as invalid
1072 playlists and mood definitions.
1073
1074 Similarly, para_audiod refuses to start if its socket file exists, since
1075 this indicates that another instance of para_audiod is running. After
1076 a crash a stale socket file might remain and you must run
1077
1078 para_audiod --force
1079
1080 once to fix it up.
1081
1082 ---------------------------------------
1083 Audio formats and audio format handlers
1084 ---------------------------------------
1085
1086 Audio formats
1087 ~~~~~~~~~~~~~
1088
1089 The following audio formats are supported by paraslash:
1090
1091 *MP3*
1092
1093 Mp3, MPEG-1 Audio Layer 3, is a common audio format for audio storage,
1094 designed as part of its MPEG-1 standard. An MP3 file is made up of
1095 multiple MP3 frames, which consist of a header and a data block. The
1096 size of an MP3 frame depends on the bit rate and on the number
1097 of channels. For a typical CD-audio file (sample rate of 44.1 kHz
1098 stereo), encoded with a bit rate of 128 kbit, an MP3 frame is about
1099 400 bytes large.
1100
1101 *OGG/Vorbis*
1102
1103 OGG is a standardized audio container format, while Vorbis is an
1104 open source codec for lossy audio compression. Since Vorbis is most
1105 commonly made available via the OGG container format, it is often
1106 referred to as OGG/Vorbis. The OGG container format divides data into
1107 chunks called OGG pages. A typical OGG page is about 4KB large. The
1108 Vorbis codec creates variable-bitrate (VBR) data, where the bitrate
1109 may vary considerably.
1110
1111 *OGG/Speex*
1112
1113 Speex is an open-source speech codec that is based on CELP (Code
1114 Excited Linear Prediction) coding. It is designed for voice
1115 over IP applications, has modest complexity and a small memory
1116 footprint. Wideband and narrowband (telephone quality) speech are
1117 supported. As for Vorbis audio, Speex bit-streams are often stored
1118 in OGG files. As of 2012 this codec is considered obsolete since the
1119 Oppus codec, described below, surpasses its performance in all areas.
1120
1121 *OGG/Opus*
1122
1123 Opus is a lossy audio compression format standardized through RFC
1124 6716 in 2012. It combines the speech-oriented SILK codec and the
1125 low-latency CELT (Constrained Energy Lapped Transform) codec. Like
1126 OGG/Vorbis and OGG/Speex, Opus data is usually encapsulated in OGG
1127 containers. All known software patents which cover Opus are licensed
1128 under royalty-free terms.
1129
1130 *AAC*
1131
1132 Advanced Audio Coding (AAC) is a standardized, lossy compression
1133 and encoding scheme for digital audio which is the default audio
1134 format for Apple's iPhone, iPod, iTunes. Usually MPEG-4 is used as
1135 the container format and audio files encoded with AAC have the .m4a
1136 extension. A typical AAC frame is about 700 bytes large.
1137
1138 *WMA*
1139
1140 Windows Media Audio (WMA) is an audio data compression technology
1141 developed by Microsoft. A WMA file is usually encapsulated in the
1142 Advanced Systems Format (ASF) container format, which also specifies
1143 how meta data about the file is to be encoded. The bit stream of WMA
1144 is composed of superframes, each containing one or more frames of
1145 2048 samples. For 16 bit stereo a WMA superframe is about 8K large.
1146
1147 *FLAC*
1148
1149 The Free Lossless Audio Codec (FLAC) compresses audio without quality
1150 loss. It gives better compression ratios than a general purpose
1151 compressor like zip or bzip2 because FLAC is designed specifically
1152 for audio. A FLAC-encoded file consists of frames of varying size, up
1153 to 16K. Each frame starts with a header that contains all information
1154 necessary to decode the frame.
1155
1156 Meta data
1157 ~~~~~~~~~
1158
1159 Unfortunately, each audio format has its own conventions how meta
1160 data is added as tags to the audio file.
1161
1162 For MP3 files, ID3, version 1 and 2 are widely used. ID3 version 1
1163 is rather simple but also very limited as it supports only artist,
1164 title, album, year and comment tags. Each of these can only be at most
1165 32 characters long. ID3, version 2 is much more flexible but requires
1166 a separate library being installed for paraslash to support it.
1167
1168 Ogg vorbis, ogg speex and flac files contain meta data as Vorbis
1169 comments, which are typically implemented as strings of the form
1170 "[TAG]=[VALUE]". Unlike ID3 version 1 tags, one may use whichever
1171 tags are appropriate for the content.
1172
1173 AAC files usually use the MPEG-4 container format for storing meta
1174 data while WMA files wrap meta data as special objects within the
1175 ASF container format.
1176
1177 paraslash only tracks the most common tags that are supported by
1178 all tag variants: artist, title, year, album, comment. When a file
1179 is added to the AFS database, the meta data of the file is extracted
1180 and stored in the audio file table.
1181
1182 Chunks and chunk tables
1183 ~~~~~~~~~~~~~~~~~~~~~~~
1184
1185 paraslash uses the word "chunk" as common term for the building blocks
1186 of an audio file. For MP3 files, a chunk is the same as an MP3 frame,
1187 while for OGG files a chunk is an OGG page, etc. Therefore the chunk
1188 size varies considerably between audio formats, from a few hundred
1189 bytes (MP3) up to 16K (FLAC).
1190
1191 The chunk table contains the offsets within the audio file that
1192 correspond to the chunk boundaries of the file. Like the meta data,
1193 the chunk table is computed and stored in the database whenever an
1194 audio file is added.
1195
1196 The paraslash senders (see below) always send complete chunks. The
1197 granularity for seeking is therefore determined by the chunk size.
1198
1199 Audio format handlers
1200 ~~~~~~~~~~~~~~~~~~~~~
1201
1202 For each audio format paraslash contains an audio format handler whose
1203 first task is to tell whether a given file is a valid audio file of
1204 this type. If so, the audio file handler extracts some technical data
1205 (duration, sampling rate, number of channels etc.), computes the
1206 chunk table and reads the meta data.
1207
1208 The audio format handler code is linked into para_server and executed
1209 via the _add_ command. The same code is also available as a stand-alone
1210 tool, para_afh, which prints the technical data, the chunk table
1211 and the meta data of a file. Moreover, all audio format handlers are
1212 combined in the afh receiver which is part of para_recv and para_play.
1213
1214 ----------
1215 Networking
1216 ----------
1217
1218 Paraslash uses different network connections for control and data.
1219 para_client communicates with para_server over a dedicated TCP control
1220 connection. To transport audio data, separate data connections are
1221 used. For these data connections, a variety of transports (UDP, DCCP,
1222 HTTP) can be chosen.
1223
1224 The chapter starts with the REFERENCE(The paraslash control
1225 service, control service), followed by a section on the various
1226 REFERENCE(Streaming protocols, streaming protocols) in which the data
1227 connections are described. The way audio file headers are embedded into
1228 the stream is discussed REFERENCE(Streams with headers and headerless
1229 streams, briefly) before the REFERENCE(Networking examples, example
1230 section) which illustrates typical commands for real-life scenarios.
1231
1232 Both IPv4 and IPv6 are supported.
1233
1234 The paraslash control service
1235 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1236
1237 para_server is controlled at runtime via the paraslash control
1238 connection. This connection is used for server commands (play, stop,
1239 ...) as well as for afs commands (ls, select, ...).
1240
1241 The server listens on a TCP port and accepts connections from clients
1242 that connect the open port. Each connection causes the server to fork
1243 off a client process which inherits the connection and deals with that
1244 client only. In this classical accept/fork approach the server process
1245 is unaffected if the child dies or goes crazy for whatever reason. In
1246 fact, the child process can not change address space of server process.
1247
1248 The section on REFERENCE(Client-server authentication, client-server
1249 authentication) above described the early connection establishment
1250 from the crypto point of view. Here it is described what happens
1251 after the connection (including crypto setup) has been established.
1252 There are four processes involved during command dispatch as sketched
1253 in the following diagram.
1254
1255 <<
1256 <pre>
1257 server_host client_host
1258 ~~~~~~~~~~~ ~~~~~~~~~~~
1259
1260 +-----------+ connect +-----------+
1261 |para_server|<------------------------------ |para_client|
1262 +-----------+ +-----------+
1263 | ^
1264 | fork +---+ |
1265 +----------> |AFS| |
1266 | +---+ |
1267 | ^ |
1268 | | |
1269 | | connect (cookie) |
1270 | | |
1271 | | |
1272 | fork +-----+ inherited connection |
1273 +---------->|child|<--------------------------+
1274 +-----+
1275 </pre>
1276 >>
1277
1278 Note that the child process is not a child of the afs process,
1279 so communication of these two processes has to happen via local
1280 sockets. In order to avoid abuse of the local socket by unrelated
1281 processes, a magic cookie is created once at server startup time just
1282 before the server process forks off the AFS process. This cookie is
1283 known to the server, AFS and the child, but not to unrelated processes.
1284
1285 There are two different kinds of commands: First there are commands
1286 that cause the server to respond with some answer such as the list
1287 of all audio files. All but the addblob commands (addimg, addlyr,
1288 addpl, addmood) are of this kind. The addblob commands add contents
1289 to the database, so they need to transfer data the other way round,
1290 from the client to the server.
1291
1292 There is no knowledge about the server commands built into para_client,
1293 so it does not know about addblob commands. Instead, the server sends
1294 a special "awaiting data" packet for these commands. If the client
1295 receives this packet, it sends STDIN to the server, otherwise it
1296 dumps data from the server to STDOUT.
1297
1298 Streaming protocols
1299 ~~~~~~~~~~~~~~~~~~~
1300
1301 A network (audio) stream usually consists of one streaming source,
1302 the _sender_, and one or more _receivers_ which read data over the
1303 network from the streaming source.
1304
1305 Senders are thus part of para_server while receivers are part of
1306 para_audiod. Moreover, there is the stand-alone tool para_recv which
1307 can be used to manually download a stream, either from para_server
1308 or from a web-based audio streaming service.
1309
1310 The following three streaming protocols are supported by paraslash:
1311
1312 - HTTP. Recommended for public streams that can be played by
1313 any player like mpg123, xmms, itunes, winamp, etc. The HTTP
1314 sender is supported on all operating systems and all platforms.
1315
1316 - DCCP. Recommended for LAN streaming. DCCP is currently
1317 available only for Linux.
1318
1319 - UDP. Recommended for multicast LAN streaming.
1320
1321 See the Appendix on REFERENCE(Network protocols, network protocols)
1322 for brief descriptions of the various protocols relevant for network
1323 audio streaming with paraslash.
1324
1325 It is possible to activate more than one sender simultaneously.
1326 Senders can be controlled at run time and via config file and command
1327 line options.
1328
1329 Note that audio connections are _not_ encrypted. Transport or Internet
1330 layer encryption should be used if encrypted data connections are
1331 needed.
1332
1333 Since DCCP and TCP are both connection-oriented protocols, connection
1334 establishment/teardown and access control are very similar between
1335 these two streaming protocols. UDP is the most lightweight option,
1336 since in contrast to TCP/DCCP it is connectionless. It is also the
1337 only protocol supporting IP multicast.
1338
1339 The HTTP and the DCCP sender listen on a (TCP/DCCP) port waiting for
1340 clients to connect and establish a connection via some protocol-defined
1341 handshake mechanism. Both senders maintain two linked lists each:
1342 The list of all clients which are currently connected, and the list
1343 of access control entries which determines who is allowed to connect.
1344 IP-based access control may be configured through config file and
1345 command line options and via the "allow" and "deny" sender subcommands.
1346
1347 Upon receiving a GET request from the client, the HTTP sender sends
1348 back a status line and a message. The body of this message is the
1349 audio stream. This is common practice and is supported by many popular
1350 clients which can thus be used to play a stream offered by para_server.
1351 For DCCP things are a bit simpler: No messages are exchanged between
1352 the receiver and sender. The client simply connects and the sender
1353 starts to stream.
1354
1355 DCCP is an experimental protocol which offers a number of new features
1356 not available for TCP. Both ends can negotiate these features using
1357 a built-in negotiation mechanism. In contrast to TCP/HTTP, DCCP is
1358 datagram-based (no retransmissions) and thus should not be used over
1359 lossy media (e.g. WiFi networks). One useful feature offered by DCCP
1360 is access to a variety of different congestion-control mechanisms
1361 called CCIDs. Two different CCIDs are available per default on Linux:
1362
1363
1364 - _CCID 2_. A Congestion Control mechanism similar to that
1365 of TCP. The sender maintains a congestion window and halves
1366 this window in response to congestion.
1367
1368
1369 - _CCID-3_. Designed to be fair when competing for bandwidth.
1370 It has lower variation of throughput over time compared with
1371 TCP, which makes it suitable for streaming media.
1372
1373 Unlike the HTTP and DCCP senders, the UDP sender maintains only a
1374 single list, the _target list_. This list describes the set of clients
1375 to which the stream is sent. There is no list for access control and
1376 no "allow" and "deny" commands for the UDP sender. Instead, the "add"
1377 and "delete" commands can be used to modify the target list.
1378
1379 Since both UDP and DCCP offer an unreliable datagram-based transport,
1380 additional measures are necessary to guard against disruptions over
1381 networks that are lossy or which may be subject to interference (as
1382 is for instance the case with WiFi). Paraslash uses FEC (Forward
1383 Error Correction) to guard against packet losses and reordering. The
1384 stream is FEC-encoded before it is sent through the UDP socket and
1385 must be decoded accordingly on the receiver side.
1386
1387 The packet size and the amount of redundancy introduced by FEC can
1388 be configured via the FEC parameters which are dictated by server
1389 and may also be configured through the "sender" command. The FEC
1390 parameters are encoded in the header of each network packet, so no
1391 configuration is necessary on the receiver side. See the section on
1392 REFERENCE(Forward error correction, FEC) below.
1393
1394 Streams with headers and headerless streams
1395 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1396
1397 For OGG/Vorbis, OGG/Speex and wma streams, some of the information
1398 needed to decode the stream is only contained in the audio file
1399 header of the container format but not in each data chunk. Clients
1400 must be able to obtain this information in case streaming starts in
1401 the middle of the file or if para_audiod is started while para_server
1402 is already sending a stream.
1403
1404 This is accomplished in different ways, depending on the streaming
1405 protocol. For connection-oriented streams (HTTP, DCCP) the audio file
1406 header is sent prior to audio file data. This technique however does
1407 not work for the connectionless UDP transport. Hence the audio file
1408 header is periodically being embedded into the UDP audio data stream.
1409 By default, the header is resent after five seconds. The receiver has
1410 to wait until the next header arrives before it can start decoding
1411 the stream.
1412
1413 Examples
1414 ~~~~~~~~
1415
1416 The "si" (server info) command lists some information about the
1417 currently running server process.
1418
1419 -> Show PIDs, number of connected clients, uptime, and more:
1420
1421 para_client si
1422
1423 The sender command of para_server prints information about senders,
1424 like the various access control lists, and it allows to (de-)activate
1425 senders and to change the access permissions at runtime.
1426
1427 -> List all senders
1428
1429 para_client sender
1430
1431 -> Obtain general help for the sender command:
1432
1433 para_client help sender
1434
1435 -> Get help for a specific sender (contains further examples):
1436
1437 s=http # or dccp or udp
1438 para_client sender $s help
1439
1440 -> Show status of the http sender
1441
1442 para_client sender http status
1443
1444 By default para_server activates both the HTTP and th DCCP sender on
1445 startup. This can be changed via command line options or para_server's
1446 config file.
1447
1448 -> List config file options for senders:
1449
1450 para_server -h
1451
1452 All senders share the "on" and "off" commands, so senders may be
1453 activated and deactivated independently of each other.
1454
1455 -> Switch off the http sender:
1456
1457 para_client sender http off
1458
1459 -> Receive a DCCP stream using CCID2 and write the output into a file:
1460
1461 host=foo.org; ccid=2; filename=bar
1462 para_recv --receiver "dccp --host $host --ccid $ccid" > $filename
1463
1464 Note the quotes around the arguments for the dccp receiver. Each
1465 receiver has its own set of command line options and its own command
1466 line parser, so arguments for the dccp receiver must be protected
1467 from being interpreted by para_recv.
1468
1469 -> Start UDP multicast, using the default multicast address:
1470
1471 para_client sender udp add 224.0.1.38
1472
1473 -> Receive FEC-encoded multicast stream and write the output into a file:
1474
1475 filename=foo
1476 para_recv -r udp > $filename
1477
1478 -> Add an UDP unicast for a client to the target list of the UDP sender:
1479
1480 t=client.foo.org
1481 para_client sender udp add $t
1482
1483 -> Receive this (FEC-encoded) unicast stream:
1484
1485 filename=foo
1486 para_recv -r 'udp -i 0.0.0.0' > $filename
1487
1488 -> Create a minimal config for para_audiod for HTTP streams:
1489
1490 c=$HOME/.paraslash/audiod.conf.min; s=server.foo.com
1491 echo receiver \".:http -i $s\" > $c
1492 para_audiod --config $c
1493
1494 -------
1495 Filters
1496 -------
1497
1498 A paraslash filter is a module which transforms an input stream into
1499 an output stream. Filters are included in the para_audiod executable
1500 and in the stand-alone tool para_filter which usually contains the
1501 same modules.
1502
1503 While para_filter reads its input stream from STDIN and writes
1504 the output to STDOUT, the filter modules of para_audiod are always
1505 connected to a receiver which produces the input stream and a writer
1506 which absorbs the output stream.
1507
1508 Some filters depend on a specific library and are not compiled in
1509 if this library was not found at compile time. To see the list of
1510 supported filters, run para_filter and para_audiod with the --help
1511 option. The output looks similar to the following:
1512
1513 Available filters:
1514 compress wav amp fecdec wmadec prebuffer oggdec aacdec mp3dec
1515
1516 Out of these filter modules, a chain of filters can be constructed,
1517 much in the way Unix pipes can be chained, and analogous to the use
1518 of modules in gstreamer: The output of the first filter becomes the
1519 input of the second filter. There is no limitation on the number of
1520 filters and the same filter may occur more than once.
1521
1522 Like receivers, each filter has its own command line options which
1523 must be quoted to protect them from the command line options of
1524 the driving application (para_audiod or para_filter). Example:
1525
1526 para_filter -f 'mp3dec --ignore-crc' -f 'compress --damp 1'
1527
1528 For para_audiod, each audio format has its own set of filters. The
1529 name of the audio format for which the filter should be applied can
1530 be used as the prefix for the filter option. Example:
1531
1532 para_audiod -f 'mp3:prebuffer --duration 300'
1533
1534 The "mp3" prefix above is actually interpreted as a POSIX extended
1535 regular expression. Therefore
1536
1537 para_audiod -f '.:prebuffer --duration 300'
1538
1539 activates the prebuffer filter for all supported audio formats (because
1540 "." matches all audio formats) while
1541
1542 para_audiod -f 'wma|ogg:prebuffer --duration 300'
1543
1544 activates it only for wma and ogg streams.
1545
1546 Decoders
1547 ~~~~~~~~
1548
1549 For each supported audio format there is a corresponding filter
1550 which decodes audio data in this format to 16 bit PCM data which
1551 can be directly sent to the sound device or any other software that
1552 operates on undecoded PCM data (visualizers, equalizers etc.). Such
1553 filters are called _decoders_ in general, and xxxdec is the name of
1554 the paraslash decoder for the audio format xxx. For example, the mp3
1555 decoder is called mp3dec.
1556
1557 Note that the output of the decoder is about 10 times larger than
1558 its input. This means that filters that operate on the decoded audio
1559 stream have to deal with much more data than filters that transform
1560 the audio stream before it is fed to the decoder.
1561
1562 Paraslash relies on external libraries for most decoders, so these
1563 libraries must be installed for the decoder to be included in the
1564 executables. For example, the mp3dec filter depends on the mad library.
1565
1566 Forward error correction
1567 ~~~~~~~~~~~~~~~~~~~~~~~~
1568
1569 As already mentioned REFERENCE(Streaming protocols, earlier),
1570 paraslash uses forward error correction (FEC) for the unreliable UDP
1571 and DCCP transports. FEC is a technique which was invented already
1572 in 1960 by Reed and Solomon and which is widely used for the parity
1573 calculations of storage devices (RAID arrays). It is based on the
1574 algebraic concept of finite fields, today called Galois fields, in
1575 honour of the mathematician Galois (1811-1832). The FEC implementation
1576 of paraslash is based on code by Luigi Rizzo.
1577
1578 Although the details require a sound knowledge of the underlying
1579 mathematics, the basic idea is not hard to understand: For positive
1580 integers k and n with k < n it is possible to compute for any k given
1581 data bytes d_1, ..., d_k the corresponding r := n -k parity bytes p_1,
1582 ..., p_r such that all data bytes can be reconstructed from *any*
1583 k bytes of the set
1584
1585 {d_1, ..., d_k, p_1, ..., p_r}.
1586
1587 FEC-encoding for unreliable network transports boils down to slicing
1588 the audio stream into groups of k suitably sized pieces called _slices_
1589 and computing the r corresponding parity slices. This step is performed
1590 in para_server which then sends both the data and the parity slices
1591 over the unreliable network connection. If the client was able
1592 to receive at least k of the n = k + r slices, it can reconstruct
1593 (FEC-decode) the original audio stream.
1594
1595 From these observations it is clear that there are three different
1596 FEC parameters: The slice size, the number of data slices k, and the
1597 total number of slices n. It is crucial to choose the slice size
1598 such that no fragmentation of network packets takes place because
1599 FEC only guards against losses and reordering but fails if slices are
1600 received partially.
1601
1602 FEC decoding in paralash is performed through the fecdec filter which
1603 usually is the first filter (there can be other filters before fecdec
1604 if these do not alter the audio stream).
1605
1606
1607 Volume adjustment (amp and compress)
1608 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1609
1610 The amp and the compress filter both adjust the volume of the audio
1611 stream. These filters operate on uncompressed audio samples. Hence
1612 they are usually placed directly after the decoding filter. Each
1613 sample is multiplied with a scaling factor (>= 1) which makes amp
1614 and compress quite expensive in terms of computing power.
1615
1616 *amp*
1617
1618 The amp filter amplifies the audio stream by a fixed scaling factor
1619 that must be known in advance. For para_audiod this factor is derived
1620 from the amplification field of the audio file's entry in the audio
1621 file table while para_filter uses the value given at the command line.
1622
1623 The optimal scaling factor F for an audio file is the largest real
1624 number F >= 1 such that after multiplication with F all samples still
1625 fit into the sample interval [-32768, 32767]. One can use para_filter
1626 in combination with the sox utility to compute F:
1627
1628 para_filter -f mp3dec -f wav < file.mp3 | sox -t wav - -e stat -v
1629
1630 The amplification value V which is stored in the audio file table,
1631 however, is an integer between 0 and 255 which is connected to F
1632 through the formula
1633
1634 V = (F - 1) * 64.
1635
1636 To store V in the audio file table, the command
1637
1638 para_client -- touch -a=V file.mp3
1639
1640 is used. The reader is encouraged to write a script that performs
1641 these computations :)
1642
1643 *compress*
1644
1645 Unlike the amplification filter, the compress filter adjusts the volume
1646 of the audio stream dynamically without prior knowledge about the peak
1647 value. It maintains the maximal volume of the last n samples of the
1648 audio stream and computes a suitable amplification factor based on that
1649 value and the various configuration options. It tries to chose this
1650 factor such that the adjusted volume meets the desired target level.
1651
1652 Note that it makes sense to combine amp and compress.
1653
1654 Misc filters (wav and prebuffer)
1655 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1656
1657 These filters are rather simple and do not modify the audio stream at
1658 all. The wav filter is only useful with para_filter and in connection
1659 with a decoder. It asks the decoder for the number of channels and the
1660 sample rate of the stream and adds a Microsoft wave header containing
1661 this information at the beginning. This allows to write wav files
1662 rather than raw PCM files (which do not contain any information about
1663 the number of channels and the sample rate).
1664
1665 The prebuffer filter simply delays the output until the given time has
1666 passed (starting from the time the first byte was available in its
1667 input queue) or until the given amount of data has accumulated. It
1668 is mainly useful for para_audiod if the standard parameters result
1669 in buffer underruns.
1670
1671 Both filters require almost no additional computing time, even when
1672 operating on uncompressed audio streams, since data buffers are simply
1673 "pushed down" rather than copied.
1674
1675 Examples
1676 ~~~~~~~~
1677
1678 -> Decode an mp3 file to wav format:
1679
1680 para_filter -f mp3dec -f wav < file.mp3 > file.wav
1681
1682 -> Amplify a raw audio file by a factor of 1.5:
1683
1684 para_filter -f amp --amp 32 < foo.raw > bar.raw
1685
1686 ------
1687 Output
1688 ------
1689
1690 Once an audio stream has been received and decoded to PCM format,
1691 it can be sent to a sound device for playback. This part is performed
1692 by paraslash _writers_ which are described in this chapter.
1693
1694 Writers
1695 ~~~~~~~
1696
1697 A paraslash writer acts as a data sink that consumes but does not
1698 produce audio data. Paraslash writers operate on the client side and
1699 are contained in para_audiod and in the stand-alone tool para_write.
1700
1701 The para_write program reads uncompressed audio data from STDIN. If
1702 this data starts with a wav header, sample rate, sample format and
1703 channel count are read from the header. Otherwise CD audio (44.1KHz
1704 16 bit little endian, stereo) is assumed but this can be overridden
1705 by command line options. para_audiod, on the other hand, obtains
1706 the sample rate and the number of channels from the decoder.
1707
1708 Like receivers and filters, each writer has an individual set of
1709 command line options, and for para_audiod writers can be configured
1710 per audio format separately. It is possible to activate more than
1711 one writer for the same stream simultaneously.
1712
1713 OS-dependent APIs
1714 ~~~~~~~~~~~~~~~~~
1715
1716 Unfortunately, the various flavours of Unix on which paraslash
1717 runs on have different APIs for opening a sound device and starting
1718 playback. Hence for each such API there is a paraslash writer that
1719 can play the audio stream via this API.
1720
1721 *ALSA*. The _Advanced Linux Sound Architecture_ is only available on
1722 Linux systems. Although there are several mid-layer APIs in use by
1723 the various Linux distributions (ESD, Jack, PulseAudio), paraslash
1724 currently supports only the low-level ALSA API which is not supposed
1725 to be change. ALSA is very feature-rich, in particular it supports
1726 software mixing via its DMIX plugin. ALSA is the default writer on
1727 Linux systems.
1728
1729 *OSS*. The _Open Sound System_ is the only API on *BSD Unixes and
1730 is also available on Linux systems, usually provided by ALSA as an
1731 emulation for backwards compatibility. This API is rather simple but
1732 also limited. For example only one application can open the device
1733 at any time. The OSS writer is activated by default on BSD Systems.
1734
1735 *OSX*. Mac OS X has yet another API called CoreAudio. The OSX writer
1736 for this API is only compiled in on such systems and is of course
1737 the default there.
1738
1739 *FILE*. The file writer allows to capture the audio stream and
1740 write the PCM data to a file on the file system rather than playing
1741 it through a sound device. It is supported on all platforms and is
1742 always compiled in.
1743
1744 *AO*. _Libao_ is a cross-platform audio library which supports a wide
1745 variety of platforms including PulseAudio (gnome), ESD (Enlightened
1746 Sound Daemon), AIX, Solaris and IRIX. The ao writer plays audio
1747 through an output plugin of libao.
1748
1749 Examples
1750 ~~~~~~~~
1751
1752 -> Use the OSS writer to play a wav file:
1753
1754 para_write --writer oss < file.wav
1755
1756 -> Enable ALSA software mixing for mp3 streams
1757
1758 para_audiod --writer 'mp3:alsa -d plug:swmix'
1759
1760
1761 ---
1762 Gui
1763 ---
1764
1765 para_gui executes an arbitrary command which is supposed to print
1766 status information to STDOUT. It then displays this information in
1767 a curses window. By default the command
1768
1769 para_audioc -- stat -p
1770
1771 is executed, but this can be customized via the --stat-cmd option. In
1772 particular it possible to use
1773
1774 para_client -- stat -p
1775
1776 to make para_gui work on systems on which para_audiod is not running.
1777
1778 Key bindings
1779 ~~~~~~~~~~~~
1780
1781 It is possible to bind keys to arbitrary commands via custom
1782 key-bindings. Besides the internal keys which can not be changed (help,
1783 quit, loglevel, version...), the following flavours of key-bindings
1784 are supported:
1785
1786 - external: Shutdown curses before launching the given command.
1787 Useful for starting other ncurses programs from within
1788 para_gui, e.g. aumix or dialog scripts. Or, use the mbox
1789 output format to write a mailbox containing one mail for each
1790 (admissible) file the audio file selector knows about. Then
1791 start mutt from within para_gui to browse your collection!
1792
1793 - display: Launch the command and display its stdout in
1794 para_gui's bottom window.
1795
1796 - para: Like display, but start "para_client <specified
1797 command>" instead of "<specified command>".
1798
1799 The general form of a key binding is
1800
1801 key_map k:m:c
1802
1803 which maps key k to command c using mode m. Mode may be x, d or p
1804 for external, display and paraslash commands, respectively.
1805
1806 Themes
1807 ~~~~~~
1808
1809 Currently there are only two themes for para_gui. It is easy, however,
1810 to add more themes. To create a new theme one has to define the
1811 position, color and geometry for for each status item that should be
1812 shown by this theme. See gui_theme.c for examples.
1813
1814 The "." and "," keys are used to switch between themes.
1815
1816 Examples
1817 ~~~~~~~~
1818
1819 -> Show server info:
1820
1821 key_map "i:p:si"
1822
1823 -> Jump to the middle of the current audio file by pressing F5:
1824
1825 key_map "<F5>:p:jmp 50"
1826
1827 -> vi-like bindings for jumping around:
1828
1829 key_map "l:p:ff 10"
1830 key_map "h:p:ff 10-"
1831 key_map "w:p:ff 60"
1832 key_map "b:p:ff 60-"
1833
1834 -> Print the current date and time:
1835
1836 key_map "D:d:date"
1837
1838 -> Call other curses programs:
1839
1840 key_map "U:x:aumix"
1841 key_map "!:x:/bin/bash"
1842 key_map "^E:x:/bin/sh -c 'vi ~/.paraslash/gui.conf'"
1843
1844 -----------
1845 Development
1846 -----------
1847
1848 Tools
1849 ~~~~~
1850
1851 In order to compile the sources from the git repository (rather than
1852 from tar balls) and for contributing non-trivial changes to the
1853 paraslash project, some additional tools should be installed on a
1854 developer machine.
1855
1856 http://git.or.cz/ (git). As described in more detail REFERENCE(Git
1857 branches, below), the git source code management tool is used for
1858 paraslash development. It is necessary for cloning the git repository
1859 and for getting updates.
1860
1861 ftp://ftp.gnu.org/pub/gnu/m4/ (m4). Some input files for gengetopt
1862 are generated from templates by the m4 macro processor.
1863
1864 ftp://ftp.gnu.org/pub/gnu/autoconf/ (autoconf) GNU autoconf creates
1865 the configure file which is shipped in the tarballs but has to be
1866 generated when compiling from git.
1867
1868 http://www.triptico.com/software/grutatxt.html (grutatxt). The
1869 HTML version of this manual and some of the paraslash web pages are
1870 generated by the grutatxt plain text to HTML converter. If changes
1871 are made to these text files the grutatxt package must be installed
1872 to regenerate the HTML files.
1873
1874 http://www.stack.nl/~dimitri/doxygen/ (doxygen). The documentation
1875 of paraslash's C sources uses the doxygen documentation system. The
1876 conventions for documenting the source code is described in the
1877 REFERENCE(Doxygen, Doxygen section).
1878
1879 ftp://ftp.gnu.org/pub/gnu/global (global). This is used to generate
1880 browsable HTML from the C sources. It is needed by doxygen.
1881
1882 Git branches
1883 ~~~~~~~~~~~~
1884
1885 Paraslash has been developed using the git source code management
1886 tool since 2006. Development is organized roughly in the same spirit
1887 as the git development itself, as described below.
1888
1889 The following text passage is based on "A note from the maintainer",
1890 written by Junio C Hamano, the maintainer of git.
1891
1892 There are four branches in the paraslash repository that track the
1893 source tree: "master", "maint", "next", and "pu".
1894
1895 The "master" branch is meant to contain what is well tested and
1896 ready to be used in a production setting. There could occasionally be
1897 minor breakages or brown paper bag bugs but they are not expected to
1898 be anything major, and more importantly quickly and easily fixable.
1899 Every now and then, a "feature release" is cut from the tip of this
1900 branch, named with three dotted decimal digits, like 0.4.2.
1901
1902 Whenever changes are about to be included that will eventually lead to
1903 a new major release (e.g. 0.5.0), a "maint" branch is forked off from
1904 "master" at that point. Obvious, safe and urgent fixes after the major
1905 release are applied to this branch and maintenance releases are cut
1906 from it. New features never go to this branch. This branch is also
1907 merged into "master" to propagate the fixes forward.
1908
1909 A trivial and safe enhancement goes directly on top of "master".
1910 New development does not usually happen on "master", however.
1911 Instead, a separate topic branch is forked from the tip of "master",
1912 and it first is tested in isolation; Usually there are a handful such
1913 topic branches that are running ahead of "master". The tip of these
1914 branches is not published in the public repository to keep the number
1915 of branches that downstream developers need to worry about low.
1916
1917 The quality of topic branches varies widely. Some of them start out as
1918 "good idea but obviously is broken in some areas" and then with some
1919 more work become "more or less done and can now be tested by wider
1920 audience". Luckily, most of them start out in the latter, better shape.
1921
1922 The "next" branch is to merge and test topic branches in the latter
1923 category. In general, this branch always contains the tip of "master".
1924 It might not be quite rock-solid production ready, but is expected to
1925 work more or less without major breakage. The maintainer usually uses
1926 the "next" version of paraslash for his own pleasure, so it cannot
1927 be _that_ broken. The "next" branch is where new and exciting things
1928 take place.
1929
1930 The two branches "master" and "maint" are never rewound, and "next"
1931 usually will not be either (this automatically means the topics that
1932 have been merged into "next" are usually not rebased, and you can find
1933 the tip of topic branches you are interested in from the output of
1934 "git log next"). You should be able to safely build on top of them.
1935
1936 However, at times "next" will be rebuilt from the tip of "master" to
1937 get rid of merge commits that will never be in "master". The commit
1938 that replaces "next" will usually have the identical tree, but it
1939 will have different ancestry from the tip of "master".
1940
1941 The "pu" (proposed updates) branch bundles the remainder of the
1942 topic branches. The "pu" branch, and topic branches that are only in
1943 "pu", are subject to rebasing in general. By the above definition
1944 of how "next" works, you can tell that this branch will contain quite
1945 experimental and obviously broken stuff.
1946
1947 When a topic that was in "pu" proves to be in testable shape, it
1948 graduates to "next". This is done with
1949
1950 git checkout next
1951 git merge that-topic-branch
1952
1953 Sometimes, an idea that looked promising turns out to be not so good
1954 and the topic can be dropped from "pu" in such a case.
1955
1956 A topic that is in "next" is expected to be polished to perfection
1957 before it is merged to "master". Similar to the above, this is
1958 done with
1959
1960 git checkout master
1961 git merge that-topic-branch
1962 git branch -d that-topic-branch
1963
1964 Note that being in "next" is not a guarantee to appear in the next
1965 release (being in "master" is such a guarantee, unless it is later
1966 found seriously broken and reverted), nor even in any future release.
1967
1968 Coding Style
1969 ~~~~~~~~~~~~
1970
1971 The preferred coding style for paraslash coincides more or less
1972 with the style of the Linux kernel. So rather than repeating what is
1973 written XREFERENCE(http://www.kernel.org/doc/Documentation/CodingStyle,
1974 there), here are the most important points.
1975
1976 - Burn the GNU coding standards.
1977 - Never use spaces for indentation.
1978 - Tabs are 8 characters, and thus indentations are also 8 characters.
1979 - Don't put multiple assignments on a single line.
1980 - Avoid tricky expressions.
1981 - Don't leave whitespace at the end of lines.
1982 - The limit on the length of lines is 80 columns.
1983 - Use K&R style for placing braces and spaces:
1984
1985 if (x is true) {
1986 we do y
1987 }
1988
1989 - Use a space after (most) keywords.
1990 - Do not add spaces around (inside) parenthesized expressions.
1991 - Use one space around (on each side of) most binary and ternary operators.
1992 - Do not use cute names like ThisVariableIsATemporaryCounter, call it tmp.
1993 - Mixed-case names are frowned upon.
1994 - Descriptive names for global variables are a must.
1995 - Avoid typedefs.
1996 - Functions should be short and sweet, and do just one thing.
1997 - The number of local variables shouldn't exceed 10.
1998 - Gotos are fine if they improve readability and reduce nesting.
1999 - Don't use C99-style "// ..." comments.
2000 - Names of macros defining constants and labels in enums are capitalized.
2001 - Enums are preferred when defining several related constants.
2002 - Always use the paraslash wrappers for allocating memory.
2003 - If the name of a function is an action or an imperative.
2004 command, the function should return an error-code integer
2005 (<0 means error, >=0 means success). If the name is a
2006 predicate, the function should return a "succeeded" boolean.
2007
2008
2009 Doxygen
2010 ~~~~~~~
2011
2012 Doxygen is a documentation system for various programming
2013 languages. The API reference on the paraslash web page is generated
2014 by doxygen.
2015
2016 It is more illustrative to look at the source code for examples than
2017 to describe the conventions in this manual, so we only describe which
2018 parts of the code need doxygen comments, but leave out details on
2019 documentation conventions.
2020
2021 As a rule, only the public part of the C source is documented with
2022 Doxygen. This includes structures, defines and enumerations in header
2023 files as well as public (non-static) C functions. These should be
2024 documented completely. For example, each parameter and the return
2025 value of a public function should get a descriptive doxygen comment.
2026
2027 No doxygen comments are necessary for static functions and for
2028 structures and enumerations in C files (which are used only within
2029 this file). This does not mean, however, that those entities need
2030 no documentation at all. Instead, common sense should be applied to
2031 document what is not obvious from reading the code.
2032
2033 --------
2034 Appendix
2035 --------
2036
2037 Network protocols
2038 ~~~~~~~~~~~~~~~~~
2039
2040 *IP*. The _Internet Protocol_ is the primary networking protocol
2041 used for the Internet. All protocols described below use IP as the
2042 underlying layer. Both the prevalent IPv4 and the next-generation
2043 IPv6 variant are being deployed actively worldwide.
2044
2045 *Connection-oriented and connectionless protocols*. Connectionless
2046 protocols differ from connection-oriented ones in that state
2047 associated with the sending/receiving endpoints is treated
2048 implicitly. Connectionless protocols maintain no internal knowledge
2049 about the state of the connection. Hence they are not capable of
2050 reacting to state changes, such as sudden loss or congestion on the
2051 connection medium. Connection-oriented protocols, in contrast, make
2052 this knowledge explicit. The connection is established only after
2053 a bidirectional handshake which requires both endpoints to agree
2054 on the state of the connection, and may also involve negotiating
2055 specific parameters for the particular connection. Maintaining an
2056 up-to-date internal state of the connection also in general means
2057 that the sending endpoints perform congestion control, adapting to
2058 qualitative changes of the connection medium.
2059
2060 *Reliability*. In IP networking, packets can be lost, duplicated,
2061 or delivered out of order, and different network protocols handle
2062 these problems in different ways. We call a transport-layer protocol
2063 _reliable_, if it turns the unreliable IP delivery into an ordered,
2064 duplicate- and loss-free delivery of packets. Sequence numbers
2065 are used to discard duplicates and re-arrange packets delivered
2066 out-of-order. Retransmission is used to guarantee loss-free
2067 delivery. Unreliable protocols, in contrast, do not guarantee ordering
2068 or data integrity.
2069
2070 *Classification*. With these definitions the protocols which are used
2071 by paraslash for steaming audio data may be classified as follows.
2072
2073 - HTTP/TCP: connection-oriented, reliable,
2074 - UDP: connectionless, unreliable,
2075 - DCCP: connection-oriented, unreliable.
2076
2077 Below we give a short descriptions of these protocols.
2078
2079 *TCP*. The _Transmission Control Protocol_ provides reliable,
2080 ordered delivery of a stream and a classic window-based congestion
2081 control. In contrast to UDP and DCCP (see below), TCP does not have
2082 record-oriented or datagram-based syntax, i.e. it provides a stream
2083 which is unaware and independent of any record (packet) boundaries.
2084 TCP is used extensively by many application layers. Besides HTTP (the
2085 Hypertext Transfer Protocol), also FTP (the File Transfer protocol),
2086 SMTP (Simple Mail Transfer Protocol), SSH (Secure Shell) all sit on
2087 top of TCP.
2088
2089 *UDP*. The _User Datagram Protocol_ is the simplest transport-layer
2090 protocol, built as a thin layer directly on top of IP. For this reason,
2091 it offers the same best-effort service as IP itself, i.e. there is no
2092 detection of duplicate or reordered packets. Being a connectionless
2093 protocol, only minimal internal state about the connection is
2094 maintained, which means that there is no protection against packet
2095 loss or network congestion. Error checking and correction (if at all)
2096 are performed in the application.
2097
2098 *DCCP*. The _Datagram Congestion Control Protocol_ combines the
2099 connection-oriented state maintenance known from TCP with the
2100 unreliable, datagram-based transport of UDP. This means that it
2101 is capable of reacting to changes in the connection by performing
2102 congestion control, offering multiple alternative approaches. But it
2103 is bound to datagram boundaries (the maximum packet size supported
2104 by a medium), and like UDP it lacks retransmission to protect
2105 against loss. Due to the use of sequence numbers, it is however
2106 able to react to loss (interpreted as a congestion indication) and
2107 to ignore out-of-order and duplicate packets. Unlike TCP it allows
2108 to negotiate specific, binding features for a connection, such as
2109 the choice of congestion control: classic, window-based congestion
2110 control known from TCP is available as CCID-2, rate-based, "smooth"
2111 congestion control is offered as CCID-3.
2112
2113 *HTTP*. _The Hypertext Transfer Protocol_ is an application layer
2114 protocol on top of TCP. It is spoken by web servers and is most often
2115 used for web services. However, as can be seen by the many Internet
2116 radio stations and YouTube/Flash videos, http is by far not limited to
2117 the delivery of web pages only. Being a simple request/response based
2118 protocol, the semantics of the protocol also allow the delivery of
2119 multimedia content, such as audio over http.
2120
2121 *Multicast*. IP multicast is not really a protocol but a technique
2122 for one-to-many communication over an IP network. The challenge is to
2123 deliver information to a group of destinations simultaneously using
2124 the most efficient strategy to send the messages over each link of
2125 the network only once. This has benefits for streaming multimedia:
2126 the standard one-to-one unicast offered by TCP/DCCP means that
2127 n clients listening to the same stream also consume n-times the
2128 resources, whereas multicast requires to send the stream just once,
2129 irrespective of the number of receivers. Since it would be costly to
2130 maintain state for each listening receiver, multicast often implies
2131 connectionless transport, which is the reason that it is currently
2132 only available via UDP.
2133
2134 Abstract socket namespace
2135 ~~~~~~~~~~~~~~~~~~~~~~~~~
2136 UNIX domain sockets are a traditional way to communicate between
2137 processes on the same machine. They are always reliable (see above)
2138 and don't reorder datagrams. Unlike TCP and UDP, UNIX domain sockets
2139 support passing open file descriptors or process credentials to
2140 other processes.
2141
2142 The usual way to set up a UNIX domain socket (as obtained from
2143 socket(2)) for listening is to first bind the socket to a file system
2144 pathname and then call listen(2), then accept(2). Such sockets are
2145 called _pathname sockets_ because bind(2) creates a special socket
2146 file at the specified path. Pathname sockets allow unrelated processes
2147 to communicate with the listening process by binding to the same path
2148 and calling connect(2).
2149
2150 There are two problems with pathname sockets:
2151
2152 * The listing process must be able to (safely) create the
2153 socket special in a directory which is also accessible to
2154 the connecting process.
2155
2156 * After an unclean shutdown of the listening process, a stale
2157 socket special may reside on the file system.
2158
2159 The abstract socket namespace is a non-portable Linux feature which
2160 avoids these problems. Abstract sockets are still bound to a name,
2161 but the name has no connection with file system pathnames.
2162
2163 License
2164 ~~~~~~~
2165
2166 Paraslash is licensed under the GPL, version 2. Most of the code
2167 base has been written from scratch, and those parts are GPL V2
2168 throughout. Notable exceptions are FEC and the WMA decoder. See the
2169 corresponding source files for licencing details for these parts. Some
2170 code sniplets of several other third party software packages have
2171 been incorporated into the paraslash sources, for example log message
2172 coloring was taken from the git sources. These third party software
2173 packages are all published under the GPL or some other license
2174 compatible to the GPL.
2175
2176 Acknowledgements
2177 ~~~~~~~~~~~~~~~~
2178
2179 Many thanks to Gerrit Renker who read an early draft of this manual
2180 and contributed significant improvements.
2181
2182 ----------
2183 References
2184 ----------
2185
2186 Articles
2187 ~~~~~~~~
2188 - Reed, Irving S.; Solomon, Gustave (1960),
2189 XREFERENCE(http://kom.aau.dk/~heb/kurser/NOTER/KOFA01.PDF,
2190 Polynomial Codes over Certain Finite Fields), Journal of the
2191 Society for Industrial and Applied Mathematics (SIAM) 8 (2):
2192 300-304, doi:10.1137/0108018)
2193
2194 RFCs
2195 ~~~~
2196
2197 - XREFERENCE(http://www.ietf.org/rfc/rfc768.txt, RFC 768) (1980):
2198 User Datagram Protocol
2199 - XREFERENCE(http://www.ietf.org/rfc/rfc791.txt, RFC 791) (1981):
2200 Internet Protocol
2201 - XREFERENCE(http://www.ietf.org/rfc/rfc2437.txt, RFC 2437) (1998):
2202 RSA Cryptography Specifications
2203 - XREFERENCE(http://www.ietf.org/rfc/rfc4340.txt, RFC 4340)
2204 (2006): Datagram Congestion Control Protocol (DCCP)
2205 - XREFERENCE(http://www.ietf.org/rfc/rfc4341.txt, RFC 4341) (2006):
2206 Congestion Control ID 2: TCP-like Congestion Control
2207 - XREFERENCE(http://www.ietf.org/rfc/rfc4342.txt, RFC 4342) (2006):
2208 Congestion Control ID 3: TCP-Friendly Rate Control (TFRC)
2209 - XREFERENCE(http://www.ietf.org/rfc/rfc6716.txt, RFC 6716) (2012):
2210 Definition of the Opus Audio Codec
2211
2212 Application web pages
2213 ~~~~~~~~~~~~~~~~~~~~~
2214
2215 - XREFERENCE(http://people.tuebingen.mpg.de/maan/paraslash/, paraslash)
2216 - XREFERENCE(http://paraslash.systemlinux.org/, paraslash (alternative page))
2217 - XREFERENCE(http://xmms2.org/wiki/Main_Page, xmms)
2218 - XREFERENCE(http://www.mpg123.de/, mpg123)
2219 - XREFERENCE(http://gstreamer.freedesktop.org/, gstreamer)
2220 - XREFERENCE(http://www.icecast.org/, icecast)
2221 - XREFERENCE(http://beesbuzz.biz/code/audiocompress.php, Audio Compress)
2222
2223 External documentation
2224 ~~~~~~~~~~~~~~~~~~~~~~
2225
2226 - XREFERENCE(http://kernel.org/pub/linux/kernel/people/hpa/raid6.pdf,
2227 H. Peter Anvin: The mathematics of Raid6)
2228 - XREFERENCE(http://info.iet.unipi.it/~luigi/fec_ccr.ps.gz,
2229 Luigi Rizzo: Effective Erasure Codes for reliable Computer
2230 Communication Protocols)
2231
2232 Code
2233 ~~~~
2234 - XREFERENCE(http://info.iet.unipi.it/~luigi/vdm.tar.gz,
2235 Original FEC implementation by Luigi Rizzo)
2236