1AXFER-TRANSFER(1) General Commands Manual AXFER-TRANSFER(1)
2
6 axfer-transfer - transferrer of audio data frame for sound devices and
7 nodes.
8
11 axfer transfer direction [ common-options ] [ backend-options ] [
12 filepath ]
13 axfer transfer direction [ common-options ] [ backend-options ] -I |
15 --separate-channels filepath ...
16 direction = capture | playback
18 common-options = ( read OPTIONS section )
20 backend-options = ( read OPTIONS section )
22 filepaths = ( read OPTIONS section )
24
27 The transfer subcommand of axfer performs transmission of audio data
28 frames for devices available in supported backends. This program is es‐
29 sentially designed to use alsa-lib APIs (libasound backend) to handle
30 sound devices supported by Linux sound subsystem (ALSA).
31
34 Direction
35 capture
36 Operates for capture transmission.
37 playback
40 Operates for playback transmission.
41 Filepath
44 Filepath is handled as a path relative to current working directory of
45 run time if it's not full path from root directory.
46 The standard input or output is used if filepath is not specified or
48 given as '-' .
49 For playback transmission, container format of given filepath is de‐
51 tected automatically and metadata is used for parameters of sample for‐
52 mat, channels, rate, duration. If nothing detected, content of given
53 file path is handled as raw data. In this case, the parameters should
54 be indicated as options.
55 Multiple filepaths are allowed with -I | --separate-channels option. In
57 this case, standard input and output is not available. The same
58 filepath is not allowed except for paths listed below:
59 - /dev/null
60 - /dev/zero
61 - /dev/full
62 - /dev/random
63 - /dev/urandom
64 Common options
67 -h, --help
68 Print help messages and finish run time.
69 -q, --quiet
72 Quiet mode. Suppress messages (not sound :))
73 -v, --verbose
76 Verbose mode. Runtime dumps supplemental information according
77 to the number of this option given in command line.
78 -d, --duration=#
81 Interrupt after # seconds. A value of zero means infinity. The
82 default is zero, so if this option is omitted then the transmis‐
83 sion process will run until it is killed. Either -d or -s option
84 is available exclusively.
85 -s, --samples=#
88 Interrupt after transmission of # number of data frames. A value
89 of zero means infinity. The default is zero, so if this options
90 is omitted then the transmission process will run until it is
91 killed. Either -d or -s option is available exclusively.
92 -f, --format=FORMAT
95 Indicate format of audio sample. This is required for capture
96 transmission, or playback transmission with files including raw
97 audio data.
98 Available sample format is listed below:
100 - [S8|U8|S16|U16|S32|U32][_LE|_BE]
101 - [S24|U24][_LE|_BE]
102 - FLOAT[_LE|_BE]
103 - FLOAT64[_LE|_BE]
104 - IEC958_SUBFRAME[_LE|_BE]
105 - MU_LAW
106 - A_LAW
107 - [S20|U20][_LE|_BE]
108 - [S24|U24][_3LE|_3BE]
109 - [S20|U20][_3LE|_3BE]
110 - [S18|U18][_3LE|_3BE]
111 - DSD_U8
112 - DSD_[U16|U32][_LE|_BE]
113 If endian-ness is omitted, host endian-ness is used.
115 Some special formats are available:
117 - cd (16 bit little endian, 44100, stereo) [= -f S16_LE -c 2 -r
118 44100]
119 - cdr (16 bit big endian, 44100, stereo) [= -f S16_BE -c 2 -f
120 44100]
121 - dat (16 bit little endian, 48000, stereo) [= -f S16_LE -c 2
122 -r 48000]
123 If omitted, U8 is used as a default. Actual available formats
125 are restricted by each transmission backend.
126 Unavailable sample format is listed below. These format has size
128 of data frame unaligned to byte unit.
129 - IMA_ADPCM
131 - MPEG
132 - GSM
133 - SPECIAL
134 - G723_24
135 - G723_24_1B
136 - G723_40
137 - G723_40_1B
138 -c, --channels=#
141 Indicate the number of audio data samples per frame. This is re‐
142 quired for capture transmission, or playback transmission with
143 files including raw audio data. The value should be between 1 to
144 256 . If omitted, 1 is used as a default.
145 -r, --rate=#
148 Indicate the number of audio data frame per second. This is re‐
149 quired for capture transmission, or playback transmission with
150 files including raw audio data. If the value is less than 1000 ,
151 it's interpreted by kHz unit. The value should be between 2000
152 and 192000 . If omitted, 8000 is used as a default.
153 -t, --file-type=TYPE
156 Indicate the type of file. This is required for capture trans‐
157 mission. Available types are listed below:
158 - wav: Microsoft/IBM RIFF/Wave format
159 - au, sparc: Sparc AU format
160 - voc: Creative Tech. voice format
161 - raw: raw data
162 When nothing is indicated, for capture transmission, the type is
164 decided according to suffix of filepath , and raw type is used
165 for fallback.
166 -I, --separate-channels
169 Indicate this option when several files are going to be handled.
170 For capture transmission, if one filepath is given as filepath ,
171 a list of filepaths is generated in a formula '<filepath>-<se‐
172 quential number>[.suffix]'. The suffix is omitted when raw for‐
173 mat of container is used.
174 --dump-hw-params
177 Dump hardware parameters and finish run time if backend supports
178 it.
179 --xfer-backend=BACKEND
182 Select backend of transmission from a list below. The default is
183 libasound.
184 - libasound
185 - libffado (optional if compiled)
186 Backend options for libasound
189 -D, --device=NODE
190 This option is used to select PCM node in libasound configura‐
192 tion space. Available nodes are listed by pcm operation of list
193 subcommand.
194 -N, --nonblock
197 With this option, PCM substream is opened in non-blocking mode.
199 When audio data frame is not available in buffer of the PCM sub‐
200 stream, I/O operation immediately returns without blocking
201 process. This option implicitly uses --waiter-type option as
202 well to prevent heavy consumption of CPU time.
203 -M, --mmap
206 With this option, audio data frame is processed directly in buf‐
208 fer of PCM substream if selected node supports this operation.
209 Without the option, temporary buffers are used to copy audio
210 data frame for buffer of PCM substream. This option implicitly
211 uses --waiter-type option as well to prevent heavy consumption
212 of CPU time.
213 -F, --period-size=#
216 This option configures given value to period_size hardware pa‐
218 rameter of PCM substream. The parameter indicates the number of
219 audio data frame per period in buffer of the PCM substream. Ac‐
220 tual number is decided as a result of interaction between each
221 implementation of PCM plugin chained from the selected PCM node,
222 and in-kernel driver or PCM I/O plugins.
223 Ideally, the same amount of audio data frame as the value should
225 be handled in one I/O operation. Actually, it is not, depending
226 on implementation of the PCM plugins, in-kernel driver, PCM I/O
227 plugins and scheduling model. For 'hw' PCM plugin in 'irq'
228 scheduling model, the value is used to decide intervals of hard‐
229 ware interrupt, thus the same amount of audio data frame as the
230 value is expected to be available for one I/O operation.
231 --period-time=#
234 This option configures given value to period_time hardware pa‐
236 rameter of PCM substream. This option is similar to --pe‐
237 riod-size option, however its unit is micro-second.
238 -B, --buffer-size=#
241 This option configures given value to buffer_size hardware pa‐
243 rameter of PCM substream. The parameter indicates the number of
244 audio data frame in buffer of PCM substream. Actual number is
245 decided as a result of interaction between each implementation
246 of PCM plugin chained from the selected PCM node, and in-kernel
247 driver or PCM I/O plugins.
248 Ideally, this is multiples of the number of audio data frame per
250 period, thus the size of period. Actually, it is not, depending
251 on implementation of the PCM plugins, in-kernel driver and PCM
252 I/O plugins.
253 --buffer-time=#
256 This option configures given value to buffer_time hardware pa‐
258 rameter of PCM substream. This option is similar to --buf‐
259 fer-size option, however its unit is micro-second.
260 --waiter-type=TYPE
263 This option indicates the type of waiter for event notification.
265 At present, four types are available; default , select , poll
266 and epoll . With default type, 'snd_pcm_wait()' is used. With
267 select type, 'select(2)' system call is used. With poll type,
268 'poll(2)' system call is used. With epoll type, Linux-specific
269 'epoll(7)' system call is used.
270 This option should correspond to one of --nonblock or --mmap op‐
272 tions, or timer value of --sched-model option. Neither this op‐
273 tion nor --test-nowait is available at the same time.
274 --sched-model=MODEL
277 This option selects scheduling model for process of this pro‐
279 gram. One of irq or timer is available. In detail, please read
280 'SCHEDULING MODEL' section.
281 When nothing specified, irq model is used.
283 -A, --avail-min=#
286 This option configures given value to avail-min software parame‐
288 ter of PCM substream. In blocking mode, the value is used as
289 threshold of the number of available audio data frames in buffer
290 of PCM substream to wake up process blocked by I/O operation. In
291 non-blocking mode, any I/O operation returns -EAGAIN until the
292 available number of audio data frame reaches the threshold.
293 This option has an effect in cases neither --mmap nor timer
295 value of --sched-model option is used.
296 -R, --start-delay=#
299 This option configures given value to start_threshold software
301 parameter of PCM substream. The value is used as threshold to
302 start PCM substream automatically. At present, this option has
303 an effect in cases neither --mmap nor timer value of
304 --sched-model option is used.
305 For playback transmission, when the number of accumulated audio
307 data frame in buffer of PCM substream to which this program
308 writes out reaches the threshold, the PCM substream starts auto‐
309 matically without an explicit call of snd_pcm_start() to the PCM
310 substream.
311 For capture transmission, this option is useless. The number of
313 accumulated audio data frame is not increased without an ex‐
314 plicit call of snd_pcm_start() to the PCM substream.
315 This option has an effect in cases neither --mmap nor timer
317 value of --sched-model option is used.
318 -T, --stop-delay=#
321 This option configures given value to stop_threshold software
323 parameter of PCM substream. The value is used as threshold to
324 stop PCM substream automatically. At present, this option has an
325 effect in cases neither --mmap nor timer value of --sched-model
326 option is used.
327 For capture transmission, when the number of accumulated audio
329 data frame in buffer of PCM substream to which a driver or
330 alsa-lib PCM plugins write reaches the threshold, the PCM sub‐
331 stream stops automatically without an explicit call of
332 snd_pcm_stop() to the PCM substream. This is a case that this
333 program leaves the audio data frames without reading for a
334 while.
335 For playback transmission, when the number available audio data
337 frame in buffer of PCM substream from which a driver or alsa-lib
338 PCM plugins read reaches the threshold, the PCM substream stops
339 automatically without an explicit call of snd_pcm_stop() to the
340 PCM substream. This is a case that this program leaves the audio
341 data frames without writing for a while.
342 This option has an effect in cases neither --mmap nor timer
344 value of --sched-model option is used.
345 --disable-resample
348 This option has an effect for 'plug' plugin in alsa-lib to sup‐
350 press conversion of sampling rate for audio data frame.
351 --disable-channels
354 This option has an effect for 'plug' plugin in alsa-lib to sup‐
356 press conversion of channels for audio data frame.
357 --disable-format
360 This option has an effect for 'plug' plugin in alsa-lib to sup‐
362 press conversion of sample format for audio data frame.
363 --disable-softvol
366 This option has an effect for 'softvol' plugin in alsa-lib to
368 suppress conversion of samples for audio data frame via addi‐
369 tional control element.
370 --fatal-errors
373 This option suppresses recovery operation from XRUN state of
375 running PCM substream, then process of this program is going to
376 finish as usual.
377 --test-nowait
380 This option disables any waiter for I/O event notification. I/O
382 operations are iterated till any of audio data frame is avail‐
383 able. The option brings heavy load in consumption of CPU time.
384 Backend options for libffado
387 This backend is automatically available when configure script detects
388 ffado_streaming_init() symbol in libffado shared object.
389 -p, --port=#
392 This option uses given value to decide which 1394 OHCI con‐
394 troller is used to communicate. When Linux system has two 1394
395 OHCI controllers, 0 or 1 are available. Neither this option nor
396 -g is available at the same time. If nothing specified, libffado
397 performs to communicate to units on IEEE 1394 bus managed by all
398 of 1394 OHCI controller available in Linux system.
399 -n, --node=#
402 This option uses given value to decide which unit is used to
404 communicate. This option requires -p option to indicate which
405 1394 OHCI controller is used to communicate to the specified
406 unit.
407 -g, --guid=HEXADECIMAL
410 This option uses given value to decide a target unit to communi‐
412 cate. The value should be prefixed with '0x' and consists of
413 hexadecimal literal letters (0-9, a-f, A-F). Neither this option
414 nor -p is available at the same time. If nothing specified,
415 libffado performs to communicate to units on IEEE 1394 bus man‐
416 aged by all of 1394 OHCI controller available in Linux system.
417 --frames-per-period=#
420 This option uses given value to decide the number of audio data
422 frame in one read/write operation. The operation is blocked till
423 the number of available audio data frame exceeds the given
424 value. As a default, 512 audio data frames is used.
425 --periods-per-buffer=#
428 This option uses given value to decide the size of intermediate
430 buffer between this program and libffado. As a default, 2 peri‐
431 ods per buffer is used.
432 --slave
435 This option allows this program to run slave mode. In this mode,
437 libffado adds unit directory into configuration ROM of 1394 OHCI
438 controller where Linux system runs. The unit directory can be
439 found by the other node on the same bus. Linux system running
440 on the node can transfer isochronous packet with audio data
441 frame to the unit. This program can receive the packet and de‐
442 multiplex the audio data frame.
443 --snoop
446 This option allows this program to run snoop mode. In this mode,
448 libffado listens isochronous channels to which device transfers
449 isochronous packet. When isochronous communication starts by any
450 unit on the same bus, the packets can be handled by this pro‐
451 gram.
452 --sched-priority=#
455 This option executes pthread_setschedparam() in a call of
457 ffado_streaming_init() to configure scheduling policy and given
458 value as its priority for threads related to isochronous commu‐
459 nication. The given value should be within RLIMIT_RTPRIO param‐
460 eter of process. Please read getrlimit(2) for details.
461
464 During transmission, SIGINT and SIGTERM will close handled files and
465 PCM substream to be going to finish run time.
466 SIGTSTP will suspend PCM substream and SIGCONT will resume it. No XRUNs
468 are expected. With libffado backend, the suspend/resume is not sup‐
469 ported and runtime is aborted immediately.
470 The other signals perform default behaviours.
472
475 $ axfer transfer playback -d 1 something
476 The above will transfer audio data frame in 'something' file for play‐
478 back during 1 second. The sample format is detected automatically as a
479 result to parse 'something' as long as it's compliant to one of Micro‐
480 soft/IBM RIFF/Wave, Sparc AU, Creative Tech. voice formats. If nothing
481 detected, -r , -c and -f should be given, or -f should be given with
482 special format.
483 $ axfer transfer playback -r 22050 -c 1 -f S16_LE -t raw something
486 The above will transfer audio data frame in 'something' file including
488 no information of sample format, as sample format of 22050 Hz, monau‐
489 ral, signed 16 bit little endian PCM for playback. The transmission
490 continues till catching SIGINT from keyboard or SIGTERM by kill(1) .
491 $ axfer transfer capture -d 10 -f cd something.wav
494 The above will transfer audio data frame to 'something.wav' file as
496 sample format of 44.1 kHz, 2 channels, signed 16 bit little endian PCM,
497 during 10 seconds. The file format is Microsoft/IBM RIFF/Wave according
498 to suffix of the given filepath .
499 $ axfer transfer capture -s 1024 -r 48000 -c 2 -f S32_BE -I -t au channels
502 The above will transfer audio data frame as sample format of 48.0 kHz,
504 2 channels, signed 32 bit big endian PCM for 1,024 number of data
505 frames to files named 'channels-1.au' and 'channels-2.au'.
506
509 In a design of ALSA PCM core, runtime of PCM substream supports two
510 modes; period-wakeup and no-period-wakeup. These two modes are for
511 different scheduling models.
512 IRQ-based scheduling model
515 As a default, period-wakeup mode is used. In this mode, in-kernel driv‐
516 ers should operate hardware to generate periodical notification for
517 transmission of audio data frame. The interval of notification is
518 equivalent to the same amount of audio data frame as one period of buf‐
519 fer, against actual time.
520 In a handler assigned to the notification, a helper function of ALSA
522 PCM core is called to update a position to head of hardware transmis‐
523 sion, then compare it with a position to head of application operation
524 to judge overrun/underrun (XRUN) and to wake up blocked processes.
525 For this purpose, hardware IRQ of controller for serial audio bus such
527 as Inter-IC sound is typically used. In this case, the controller gen‐
528 erates the IRQ according to transmission on the serial audio bus. In
529 the handler assigned to the IRQ, direct media access (DMA) transmission
530 is requested between dedicated host memory and device memory.
531 If target hardware doesn't support this kind of mechanism, the periodi‐
533 cal notification should be emulated by any timer; e.g. hrtimer, kernel
534 timer. External PCM plugins generated by PCM plugin SDK in alsa-lib
535 should also emulate the above behaviour.
536 In this mode, PCM applications are programmed according to typical way
538 of I/O operations. They execute blocking system calls to read/write au‐
539 dio data frame in buffer of PCM substream, or blocking system calls to
540 wait until any audio data frame is available. In axfer , this is called
541 IRQ-based scheduling model and a default behaviour. Users can explic‐
542 itly configure this mode by usage of --sched-model option with irq
543 value.
544 Timer-based scheduling model
547 The no-period-wakeup mode is an optional mode of runtime of PCM sub‐
548 stream. The mode assumes a specific feature of hardware and assist of
549 in-kernel driver and PCM applications. In this mode, in-kernel drivers
550 don't operate hardware to generate periodical notification for trans‐
551 mission of audio data frame. The hardware should automatically con‐
552 tinue transmission of audio data frame without periodical operation of
553 the drivers; e.g. according to auto-triggered DMA transmission, a chain
554 of registered descriptors.
555 In this mode, nothing wakes up blocked processes, therefore PCM appli‐
557 cations should be programmed without any blocking operation. For this
558 reason, this mode is enabled when the PCM applications explicitly con‐
559 figure hardware parameter to runtime of PCM substream, to prevent dis‐
560 order of existing applications. Additionally, nothing maintains timing
561 for transmission of audio data frame, therefore the PCM applications
562 should voluntarily handle any timer to queue audio data frame in buffer
563 of the PCM substream for lapse of time. Furthermore, instead of driver,
564 the PCM application should call a helper function of ALSA PCM core to
565 update a position to head of hardware transmission and to check XRUN.
566 In axfer , this is called timer-based scheduling model and available as
568 long as hardware/driver assists no-period-wakeup runtime. Users should
569 explicitly set this mode by usage of --sched-model option with timer
570 value.
571 In the scheduling model, PCM applications need to care of available
573 space on PCM buffer by lapse of time, typically by yielding CPU and
574 wait for rescheduling. For the yielding, timeout is calculated for
575 preferable amount of PCM frames to process. This is convenient to a
576 kind of applications, like sound servers. when an I/O thread of the
577 server wait for the timeout, the other threads can process audio data
578 frames for server clients. Furthermore, with usage of rewinding/for‐
579 warding, applications can achieve low latency between transmission po‐
580 sition and handling position even if they uses large size of PCM buf‐
581 fers.
582 Advantages and issues
585 Ideally, timer-based scheduling model has some advantages than
586 IRQ-based scheduling model. At first, no interrupt context runs for PCM
587 substream. The PCM substream is handled in any process context only. No
588 need to care of race conditions between IRQ and process contexts. This
589 reduces some concerns for some developers of drivers and applications.
590 Secondary, CPU time is not used for handlers on the interrupt context.
591 The CPU time can be dedicated for the other tasks. This is good in a
592 point of Time Sharing System. Thirdly, hardware is not configured to
593 generate interrupts. This is good in a point of reduction of overall
594 power consumption possibly.
595 In either scheduling model, the hardware should allow drivers to read
597 the number of audio data frame transferred between the dedicated memory
598 and the device memory for audio serial bus. However, in timer-based
599 scheduling model, fine granularity and accuracy of the value is impor‐
600 tant. Actually hardware performs transmission between dedicated memory
601 and device memory for a small batch of audio data frames or bytes. In a
602 view of PCM applications, the granularity in current transmission is
603 required to decide correct timeout for each I/O operation. As of Linux
604 kernel v4.21, ALSA PCM interface between kernel/userspace has no fea‐
605 ture to report it.
606
609 The transfer subcommand of axfer is designed to keep compatibility to
610 aplay(1). However some options below are not compatible due to several
611 technical reasons.
612 -I, --separate-channels
615 This option is supported just for files to store audio data
616 frames corresponding to each channel. In aplay(1) implementa‐
617 tion, this option has an additional effect to use PCM buffer
618 aligned to non-interleaved order if a target device supports. As
619 of 2018, PCM buffer of non-interleaved order is hardly used by
620 sound devices.
621 -A, --avail-min=#
624 This option indicates threshold to wake up blocked process in a
625 unit of audio data frame. Against aplay(1) implementation, this
626 option has no effect with --mmap option as well as timer of
627 --sched-model option.
628 -R, --start-delay=#
631 This option indicates threshold to start prepared PCM substream
632 in a unit of audio data frame. Against aplay(1) implementation,
633 this option has no effect with --mmap option as well as timer of
634 --sched-model option.
635 -T, --stop-delay=#
638 This option indicates threshold to stop running PCM substream in
639 a unit of audio data frame. Against aplay(1) implementation,
640 this option has no effect with --mmap option as well as timer of
641 --sched-model option.
642 --max-file-time=#
645 This option is unsupported. In aplay(1) implementation, the op‐
646 tion has an effect for capture transmission to save files up to
647 the same number of data frames as the given value by second
648 unit, or the maximum number of data frames supported by used
649 file format. When reaching to the limitation, used file is
650 closed, then new file is opened and audio data frames are writ‐
651 ten. However, this option requires extra handling of files and
652 shall increase complexity of main loop of axfer.
653 --use-strftime=FORMAT
656 This option is unsupported. In aplay(1) implementation, the op‐
657 tion has an effect for capture transmission to generate file
658 paths according to given format in which some extra formats are
659 available as well as formats supported by strftime(3). However,
660 this option requires extra string processing for file paths and
661 it's bothersome if written in C language.
662 --process-id-file=FILEPATH
665 This option is unsupported. In aplay(1) implementation, the op‐
666 tion has an effect to create a file for given value and write
667 out process ID to it. This file allows users to get process ID
668 and send any POSIX signal to aplay process. However, this idea
669 has some troubles for file locking when multiple aplay processes
670 run with the same file.
671 -V, --vumeter=TYPE
674 This option is not supported at present. In aplay(1) implementa‐
675 tion, this option has an effect to occupy stdout with some ter‐
676 minal control characters and display vumeter for monaural and
677 stereo channels. However, some problems lay; this feature is
678 just for audio data frames with PCM format, this feature brings
679 disorder of terminal after aborting, stdout is not available for
680 pipeline.
681 -i, --interactive
684 This option is not supported at present. In aplay(1) implementa‐
685 tion, this option has an effect to occupy stdin for key input
686 and suspend/resume PCM substream according to pushed enter key.
687 However, this feature requires an additional input handling in
688 main loop and leave bothersome operation to maintain PCM sub‐
689 stream.
690 -m, --chmap=CH1,CH2,...
693 ALSA PCM core and control core doesn't support this feature,
694 therefore remapping should be done in userspace. This brings
695 overhead to align audio data frames, especially for mmap opera‐
696 tion. Furthermore, as of alsa-lib v1.1.8, some plugins don't
697 support this feature expectedly, thus this option is a lack of
698 transparent operation. At present, this option is not supported
699 yet not to confuse users.
700 SIGTSTP, SIGCONT
703 This performs suspend/resume of PCM substream. In aplay(1) im‐
704 plementation, these operations bring XRUN state to the sub‐
705 stream, and suspend/resume is done in interactive mode in the
706 above. Some developers use the signal for recovery test from
707 XRUN. At present, no alternative is supported for the test.
708 SIGUSR1
711 This is not supported. In aplay(1) implementation, this signal
712 is assigned to a handler to close a current file to store audio
713 data frame and open a new file to continue processing. However,
714 as well as --max-file-time option, this option should increase
715 complexity of main loop of axfer.
716
719 Modular structure
720 This program consists of three modules; xfer , mapper and container .
721 Each module has an abstraction layer to enable actual implementation.
722 -------- ---------- -------------
724 device <-> | xfer | <-> | mapper | <-> | container | <-> file
725 -------- ---------- -------------
726 libasound single wav
727 libffado multiple au
728 voc
729 raw
730 The xfer module performs actual transmission to devices and nodes. The
732 module can have several transmission backends. As a default backend,
733 libasound backend is used to perform transmission via alsa-lib APIs.
734 The module allows each backend to parse own command line options.
735 The container module performs to read/write audio data frame via de‐
737 scriptor for file/stream of multimedia container or raw data. The mod‐
738 ule automatically detect type of multimedia container and parse parame‐
739 ters in its metadata of data header. At present, three types of multi‐
740 media containers are supported; Microsoft/IBM RIFF/Wave ( wav ), Sparc
741 AU ( au ) and Creative Technology voice ( voc ). Additionally, a spe‐
742 cial container is prepared for raw audio data ( raw ).
743 The mapper module handles buffer layout and alignment for transmission
745 of audio data frame. The module has two implementations; single and
746 multiple . The single backend uses one container to construct the buf‐
747 fer. The multiple backend uses several containers to construct it.
748 Care of copying audio data frame
751 Between the xfer module and mapper module, a pointer to buffer includ‐
752 ing audio data frames is passed. This buffer has two shapes for inter‐
753 leaved and non-interleaved order. For the former, the pointer points to
754 one buffer. For the latter, the pointer points to an array in which
755 each element points to one buffer. Between the mapper module and con‐
756 tainer module, a pointer to one buffer is passed because supported me‐
757 dia containers including raw type store audio data frames in inter‐
758 leaved order.
759 In passing audio data frame between the modules, axfer is programmed to
761 avoid copying between a buffer to another buffer as much as possible.
762 For example, in some scenarios below, no copying occurs between mod‐
763 ules.
764 - xfer(mmap/interleaved), mapper(single), container(any)
766 - xfer(mmap/non-interleaved), mapper(multiple), containers(any)
767 Unit test
770 For each of the mapper and container module, unit test is available. To
771 run the tests, execute below command:
772 $ make test
774 Each test iterates writing to file and reading to the file for many
776 times and it takes long time to finish. Please take care of the execu‐
777 tion time if running on any CI environment.
778
781 axfer(1), axfer-list(1), alsamixer(1), amixer(1)
782
785 Takashi Sakamoto <o-takashi@sakamocchi.jp>
786alsa-utils 28 November 2018 AXFER-TRANSFER(1)