avcapture(1) avcapture(1)
avcapture - capture and synchronization tool for uncompressed video and
audio
avcapture [ -a # ] [ -b # ] [ -c # ] [ -d # ] [ -f %s ]
[ -h ] [ -i # ] [ -o # ] [ -p # ] [ -t %s ]
[ -v vlan_cmd_file ] [ -w # ] [ -A ] [ -B ]
[ -D ] [ -F ] [ -R ] [ -T ] [ -V ] [ filename ]
-a <audio format>
specifies the file format of the audio data
1 = no audio (default)
2 = QuickTime (currently not implemented)
3 = AIFC
-b <buffer size>
allows you to specify the maximum number of buffer elements to be
allocated. By default, 30 buffers are used, but the value can be
adjusted for I/O load balancing. The maximum number of buffers which
can be allocated is limited by the amount of available physical
memory in the system.
-c <colorspace>
specifies the image colorspace (see /usr/include/dmedia/vl.h or the
Digital Media Programming Guide for definitions). The default is
VL_FORMAT_DIGITAL_COMPONENT (i.e. CCIR601 YCbCr).
1 = CCIR601 YCbCr (default) [YUV(A) with headroom]
2 = SMPTE YUV [YUV(A) full range]
3 = RP175 [RGB(A) with headroom]
4 = RGB [RGB(A) full range]
-d <video device number>
specifies the video device to be used (as reported by vlinfo). This
need not be specified unless you have more than one video device in
your system; it defaults to 0, i.e. the first device found.
-f <capture format>
specifies the capture format of the video data:
frames = uncompressed frames
fields = uncompressed fields
-h prints a usage message giving a summary of available flags.
Page 1
avcapture(1) avcapture(1)
-i <number of frames>
specifies the number of video images to capture. By default 60
fields are transferred.
-p <VL packing type>
specifies the pixel packing; see /usr/include/dmedia/vl.h for
definitions. The default is VL_PACKING_R422_8; (This is the most
efficient selection, since it only takes 16 bits per pixel; all 444
packing selections take 24 bits per pixel; all 4444 packing
selections take 32 bits per pixel).
1 = R242_8 (default)
2 = 242_8
3 = 444_8
4 = R444_8
5 = 4444_8
6 = R4444_8
-t </dev/ttyc#>
specify the tty device port for serial RS422 V-LAN control
-v <V-LAN command file>
instructs the program to parse the specified file for V-LAN commands
and to issue them immediately.
-w <scatter gather>
specifies the number of images to simultaneously write to disk using
writev(2)
-A abort if any image is dropped (when a VL sequence lost event is
generated).
-B calculates the video disk I/O performance and prints the results
after the data is captured to disk.
-D enables loquacious output for debugging purposes.
-F swap field dominance (works only in fields mode).
-R enables Rice encoding (DIVO only).
Page 2
avcapture(1) avcapture(1)
-T trigger with VITC.
-V enables verbose mode and prints the values of parameters used to
setup the hardware for audio and video transfers.
<filename>
specifies the name of the file to store the video data. If the -a3
option is chosen audio data will be stored in a separate audio file,
filename.aifc. If the files already exist they will be truncated.
The utility avcapture uses the Digital Media libraries to record a
sequence of video images input from a Silicon Graphics video device.
Audio data can also be captured in a separate AIFC file with the -a
option. If the audio and video hardware are locked to a common external
reference signal the audio and video data are guaranteed to be
synchronized.
avcapture is best suited to store large volumes of audio and video data
in realtime. This is possible if a striped logical volume disk array of
sufficient size and configuration is used with the XFS file system. The
logical volume, xlv_mgr and xlv_assemble, and file system, mkfs_xfs,
administration tools to configure your disk array. Another set of
production utilities, diskalign and diskperf are also available and make
an otherwise daunting task much simpler to perform.
Files created with this utility will be written in the QuickTime file
format. See /usr/share/src/dmedia/movie for details on how to create and
edit QuickTime files using the SGI movie library.
The size of a video image is determined by the combination of the capture
mode, fields or frames, colorspace, and packing. The image size and
video format are then used to compute the required data transfer rate.
The video input timing and format must be set beforehand with vcp. If
audio input is selected with the -a option the default audio device is
used. The audio control parameters can be adjusted with apanel.
avcapture can be run in two ways, burst mode or stream mode. In stream
mode it is assumed that the disk array onto which the data will be
recorded has been performance tested for the specified data format. The
video data is recorded directly to a file through a small fifo in memory
as defined by the -b buffer size option. If the disk device is not setup
for real-time capture the -b option can be used to record in batch mode
fashion using as many buffers as can be stored in memory with frame
accurate control in small intervals over the desired duration of time.
The -v option specifies a text file containing V-LAN edit commands.
Robust error checking is not performed, therefore, unpredictable behavior
should be expected if invalid V-LAN commands are issued or if the
commands are issued in the wrong order. Permissions and ownership of the
Page 3
avcapture(1) avcapture(1)
command file must be properly set. If V-LAN control is chosen and the CO
(Coincidence On) and SC (Set Coincidence) commands are issued the GPI
input of the video device must be connected to the V-LAN transmitter to
accurately initiate the data transfer. This option should be used in
conjunction with the -t flag. A special cable must be constructed for
the trigger input. See Appendix A of the DIVO XIO Board Owner's Guide
for details.
avplayback(1), diskalign(1), diskperf(1), memtovid(1), orion(1),
vidtomen(1)
avcapture runs in "compatibility mode" when using the Octane Digital
Video board with Irix 6.4. This is due to the evolution of the VL API.
The DIVO board supports the new API on Irix 6.4, but Octane Digital Video
will not support the new API until Irix 6.5. Consequently, there is a
requirement to map the colorspace and packing specified in the new API
model back to the format and packing model of the old API; unfortunately,
not all combinations are supported. Refer to the avcapture source in
/usr/share/src/dmedia/tools.
PPPPaaaaggggeeee 4444 [ Back ]
|
