mediarecorder(1) mediarecorder(1) NAME mediarecorder - record movies, still images, or audio files SYNOPSIS mediarecorder [-media image | movie | audio ] [-source video | screen ] [-device devicename] [-jack jackname] [-outputfile filename] [-fixedfile ] [-allowmenu menustring] [-execstring shellcmdstring] [-noclipbin ] [-screensel optionsstring] [-delay n] [-duration n] [-start ] [-exitdone ] [-movieoptions optionstring] [-imageoptions optionstring] [-audiooptions optionstring] DESCRIPTION mediarecorder is a graphical tool to record digital media files from the camera, microphone, external video and audio hardware, and workstation screen. To supplement the discussion below, you may wish to consult the online help, accessible from the Help menu within the tool. mediarecorder records three different types of files: movie files (such as QuickTime), still image files (such as TIFF, JFIF or GIF), and audio files (such as AIFF/AIFC). You can record movie or still image files from the system camera, an external video source connected to the video input jacks, or from the workstation screen. You can record audio files from the system microphone, an external analog audio source connected to the Line In jacks of your system, or from a digital audio input source. To choose the type of file and signal source to record, click one of the three icon buttons in the lower left corner of the mediarecorder display. Each button presents a pop-up menu listing the signal sources from which you can record. Each menu entry will also have a submenu listing the available audio/video jacks or screen selection options. To customize the parameters for recording, such as image size, frame rate, audio rate, or compression, choose one of the options from the Tasks menu. Each media type has its own submenu, listing several presets appropriate for various uses. Choose the task which most closely suits your needs. To view the settings corresponding to a particular task, choose Show Task Settings from the Tasks menu. To finely control recording parameters, choose Custom Settings from the appropriate Tasks menu. To begin recording, click the button with the circular dot on the right side of the mediarecorder display, or choose Record from the Edit menu. To stop recording, click the button with the square on the right side of the mediarecorder display. To review a clip you have just recorded, click its icon, then click the button with the triangle on the right side of the mediarecorder display. CLIP BIN/SAVE ALL CLIPS The clip bin is an icon view inside the mediarecorder display. It shows icons for image, audio, and movie files contained in the current directory. The current directory is denoted by the path name above the icon display. mediarecorder will create all new media files in the current directory. By default, mediarecorder uses the current working directory for the clip bin. To choose a different directory, you can edit the pathname above the icon display, or drag a desktop icon for the desired directory into the icon pocket next to the path name. When mediarecorder records a file, it gives the file a name, comprised of the type of file (image, audio, or movie), a number, and an extension for the file type (such as .mov for QuickTime movies or .gif for GIF files). To rename the file, click on the icon name in the clip bin, type a new name for the file, and press Enter. By default, mediarecorder will create a new file with a new name every time you record. This is useful if you want to record several different pieces of material. To record the same piece of material several times, but keep only the best clip, uncheck the Save All Clips option on the Options menu. When Save All Clips is unchecked, mediarecorder will create all new clips with the same name, replacing the previous recording. SETTING A RECORD DELAY mediarecorder has a time delay feature, similar to the self-timer on some photographic cameras. This is especially useful if you want to capture open menus in another application. To set up a record delay, choose one of the options from the Set Record Delay entry on the Edit menu, or use the -delay option from the command line. TEMPORARY FILES/SCRATCH DISKS mediarecorder makes use of temporary files when recording movies. By default, the temporary files are placed in the current working directory, except if the environment variable TMPDIR is set. If TMPDIR is set, the temporary files are placed in the path named by TMPDIR. By default, mediarecorder uses only one temporary directory. It is possible to specify more than one temporary directory for movie recording. If you specify temporary directories on different disk volumes, mediarecorder can record to multiple disks at once, increasing the amount of video which can be written to disk. To do this, choose the Set Scratch Disks command from the Options menu. Type or drop icons for each temporary directory, then click the Add button. To remove a directory from the list, select it and click Remove. DIRECT MOVIE RECORDING When mediarecorder records movies made with software codecs such as QuickTime Video, Cinepak, or Indeo, it records the raw data to a temporary file, then performs software compression after the video capture is complete. When mediarecorder records QuickTime movies with uncompressed video, or QuickTime movies from sources which support hardware JPEG compression, it can record directly to the output movie file. This is faster and uses less disk space during capture when compared to the non-direct approach. By default, mediarecorder will perform direct recording whenever it can. However, when mediarecorder records direct to the output file, it (by definition) uses only one disk volume-that containing the output movie file. If you want to record using multiple disk volumes, you may wish to turn off direct recording. To turn off direct recording, uncheck the "Record directly to movie file when possible" checkbox in the Custom Settings dialog, under the Tasks/Movie menu. HIGH SCHEDULING PRIORITY FOR VIDEO CAPTURE When you record movies from live video using mediarecorder, it will attempt to run its internal video recording threads at a very high priority. It will also attempt to pin down physical memory buffers for video capture. Both of these actions help achieve optimal video captures. However, in order to take advantage of these features, you must run mediarecorder as the root user. If you do not run mediarecorder as the root user, it will record video just fine, but at possibly reduced efficiency compared to what it would achieve while running as the root user. HANDLING DROPPED VIDEO FRAMES Depending on the disk bandwidth available, and the frame rate you have chosen for capture, mediarecorder may not be able to write all incoming video to disk. This occurs mostly with uncompressed video, or software compression. Hardware-based JPEG compression, such as that supported by the O2, Octane/IMPACT Compression, or the Cosmo Compress option board, typically will not have problems with dropped video. mediarecorder deals with dropped video frames by extending the duration of video frames immediately preceding them. The resulting movie plays properly, with some loss of motion in the places where frames were dropped. For many applications, it is important to record movies which contain no dropped frames at all. To do this, check the "Abort recording if video frames are dropped" checkbox in the Custom Settings dialog, under the Tasks/Movie menu. When this box is checked, mediarecorder will automatically abort any capture if it drops video frames. All movie files created when this option is checked are thus guaranteed to contain every video frame. By default, the "Abort recording if video frames are dropped" option is disabled. However, the "Video Production (JPEG)", "Video Production (Cross-Platform)" and "Video Production, Uncompressed" tasks in the Tasks/Movie menu will enable this option. ACHIEVING HIGH QUALITY VIDEO CAPTURES To achieve the best video captures without dropping frames, it helps to observe the following guidelines: You are strongly advised to run mediarecorder as the root user. See the section entitled "High Scheduling Priority for Video Capture" elsewhere in this manual page. Consider devoting one or more entire disk drives solely as scratch disks for video recording. Furthermore, you should delete all files from each drive before each capture. To direct mediarecorder to use particular disks for scratch purposes, see the sections entitled "Temporary Files/Scratch Disks" and "Direct Movie Recording" elsewhere in this manual page. The video signal you use should be stable and free of distortion. Examples include professional or prosumer-quality VTRs, laserdisc players, or any time-base corrected video source. A weak or distorted video signal (such as a weak broadcast signal, or a poor quality consumer VHS tape) can cause the system video hardware to lose a frame. Do not use or invoke other applications concurrently while recording video. Do not move windows on the workstation display while recording video. Check for the presence of any cron(1m) jobs which may activate while you are recording video, and temporarily disable them. Finally, you may want to ensure that rtnetd(1m) is enabled on your system. This helps prevent network traffic from interfering with video recording. SCREEN RECORDING AND MONITOR REFRESH RATES When you record movies from the screen using mediarecorder, the list of available image rates for movie recording is typically dependent on the refresh rate of the workstation monitor. To change the monitor refresh rate, use the xscreen(1) (graphical, for O2 systems) or setmon(1) (command line, all systems) utilities. SCREEN RECORDING AND THE X ADVISOR When mediarecorder records movie files from the screen, it can obtain information from the X server about what is actually changing on the screen. This is called the "X advisor". With the X advisor, mediarecorder can know when nothing has changed on screen, so it can avoid writing unnecessary screen data to disk. This can improve the frame rate, and help screen recording files to occupy less disk space. This is because the screen advisor helps mediarecorder to capture the least amount of data necessary to accurately reproduce the screen activity in the movie file. The X advisor is most useful for scenarios involving screen controls such as buttons and menus, because very often, little of the screen changes. There are situations in which the X advisor is less helpful. These scenarios typically involve captures of 3D animations, where everything in the window is updated on a frequent basis. In those situations, when users want a full rate capture, it is best to turn off the X advisor. This is because it uses CPU resources and can prevent the capture from actually achieving full rate. By default, the "Screen Recording (Windows, Buttons, Menus)" task uses the screen advisor, and the "Screen Recording (Graphics, Rendering)" tasks does not use the X advisor. To enable or disable the X advisor, set or clear the "Use X advisor to optimize screen recording" checkbox in the Custom Settings dialog, under the Tasks/Movie menu. When you record from the screen using software, the X advisor is always used as an intrinsic part of the recording process. Therefore, the checkbox is disabled in the Custom Settings dialog. COMMAND LINE OPTIONS The following is a description of the options which may be specified on the command line. -media Specify the initial media type for the tool at startup. Valid choices are image, movie, or audio. Only one option may be specified. -source Specify the initial signal source for the tool at startup, when recording movies, or still images. Valid choices are video or screen. Only one option may be specified. This option is ignored if -media does not also appear on the command line. -device devicename Specify the initial hardware device name for the tool at startup. If the device does not exist on the current system, or is not specified, the default device will be used. Only one device name may be specified. This option is ignored if -media does not also appear on the command line. The valid device names for movie capture from live video are: vino Indy built-in (VINO) video impact IMPACT Video or Octane Video boards impactcomp IMPACT and Octane Compression, used alone or in conjunction with IMPACT/Octane Video option boards evo Octane Personal Video option board ev1 Galileo, Indy Video, Indy Video 601, Indigo2 Video, Indigo2 Video for IMPACT cosmo Cosmo JPEG compression board, used in conjunction with Galileo, Indy Video, Indy Video 601, Indigo2 Video, or Indigo2 Video for IMPACT option boards mvp O2 built-in video hardware The valid device names for movie capture from the screen are: xlib Screen capture via the X Window System. This option uses the system's built-in graphics hardware, and is available regardless of system type or installed video hardware. evo Screen capture via the Octane Personal Video option board cosmo Screen capture via the Cosmo JPEG compression board, used in conjunction with Galileo, Indy Video, Indy Video 601, Indigo2 Video, or Indigo2 Video for IMPACT mvp Screen capture via the O2 built-in video hardware The valid device names for still image capture from live video are: vino Indy built-in (VINO) video impact IMPACT Video or Octane Video option boards evo Octane Personal Video option board ev1 Galileo, Indy Video, Indy Video 601, Indigo2 Video, Indigo2 Video for IMPACT option boards mvp O2 built-in video hardware -jack jackname Specify the initial jack name for the tool at startup, when recording video or audio files. If the named jack does not exist, or is not specified, the system default jack will be used. Only one jack name may be specified. This option is ignored if -media does not also appear on the command line, or if the source is something other than video or audio. The valid jack names for MVP video ("mvp") are: camera composite yc The valid jack names for Impact Video and Octane Video ("impact") are: node_dig_video_in_1 node_dig_video_in_2 node_dig_video_in_dl The valid jack names for Impact Compression and Octane Compression ("impactcomp") are: node_analog_video_in.Composite node_analog_video_in.SVideo Note that if both Impact Video and Impact Compression are installed, or both Octane Video and Octane Compression are installed, you may access both jack lists at once, from either the "impact" or "impactcomp" devices. The valid jack names for the Octane Personal Video option board ("evo") are: camera Digital yc composite The valid jack names for Indy Built-in Video ("vino") are: indycam yc composite The valid jack names for Galileo Video ("ev1") (with digital breakout box) are: yc1 yc2 yc3 composite1 composite2 composite3 yryby1 yryby2 parallel1 serial1 The valid jack names for Galileo Video ("ev1") (with IndyCam in place of the digital breakout box) are: indycam yc1 yc2 yc3 composite1 composite2 composite3 yryby1 yryby2 The valid jack names for Indigo2 Video, and Indigo2 Video for IMPACT ("ev1") are: indycam composite1 yc1 composite2 The valid jack names for Indy Video ("ev1") are: composite1 yc1 composite2 The valid jack names for Indy Video 601 ("ev1") are: composite1 yc1 composite2 parallel1 serial1 The valid jack names for the Cosmo Compression board ("cosmo") are the same as those for the corresponding Galileo family product installed on the system-Galileo Video, Indigo2 Video, Indigo2 Video for IMPACT, Indy Video, or Indy Video 601. The valid jack names for audio recording are the labels associated with each of the audio input interfaces on the system. To learn what the labels are for a given system, inspect the output of audiopanel(1) when invoked with the -print option. Case is important for audio jack names, and names containing spaces must be enclosed in quotes. -outputfile filename Specify name of output file. The application will choose a default name if this option is not specified. -fixedfile Fix the name of the output file, so it will not automatically increment, and cannot be altered by the user. -allowmenu menustring This option allows you to restrict the media types and signal sources which appear in the menus. This option must be followed by a string containing groups of media types and sources, separated by commas. No whitespace is allowed. Media types and signal sources not explicitly specified on this list will not appear in the menus. The valid media types are image, movie, or audio. The valid signal sources are video or screen. No signal source may be specified if the media type is audio. The groups are separated by commas. Multiple source names are separated with a plus sign. This example will allow everything to appear on the menus: -allowmenu movie=video+screen,image=video+screen,audio This example allowing only the audio options on the menus. The user will be able to record only audio files: -allowmenu audio This example allows only image capture from screen and movie capture from video on the menus: -allowmenu movie=video,image=screen This example allows only movie capture from screen and movie capture from live video on the menus: -allowmenu movie=video+screen This example allows only movie capture from the screen on the menus: -allowmenu movie=screen -execstring shellcmdstring This is a string to execute after any file is written to disk. The sequence %s within the string will be substituted with name of the captured file. -noclipbin This option is obsolete. It is maintained for compatibility with previous versions of the application. -screensel optionstring This option controls the type and placement of the onscreen selection rectangle used when performing screen captures. This option is followed by a mandatory string containing any of the following suboptions, separated by commas. No whitespace is allowed: method= [ adjustable | fullscreenntsc | fullscreenpal | fullscreenntscsif | fullscreenpalsif | ntscfull | palfull ] Type of screen selection. adjustable provides a user-resizable frame. fullscreenntsc and fullscreenpal record the entire workstation screen, resized to fit either an NTSC or a PAL sized movie file, respectively. fullscreenntscsif and fullscreenpalsif record the entire workstation screen, resized to fit an NTSC or PAL SIF (half-sized) movie file, respectively. ntscfull and palfull record an NTSC or PAL sized subrectangle of the workstation display to a NTSC or PAL sized movie. Not all screen selection options are available on all systems. In particular, the ntscfull, and palfull options are available on O2, Indy or Indigo2 with the Cosmo Compress option, and Octane with the Octane Personal Video option. The fullscreenntsc and fullscreenpal options are available on Indy and Indigo2 with the Cosmo Compress option, or Octane with the Octane Personal Video option. The fullscreenntscsif and fullscreenpalsif options are available only on Octane with the Octane Personal Video option. None of the screen selection options except adjustable are recognized if the recording device is xlib. This is true regardless of the workstation model or hardware options present. Please note that when recording the entire workstation screen using one of the video hardware options, typically only the top 960 lines of the workstation display are captured. This is due to limitations imposed by the video hardware option, not by the mediarecorder application. x= x y= y Origin of upper left corner of screen selection, in root window pixel units. The origin is not allowed if any of the fullscreen selection methods are specified. w= w h= h Width and height of screen selection, in pixels. The dimensions may be specified only if the screen selection method is adjustable. -delay n This option sets the the time between when the record button is pressed, and when recording actually begins to n seconds. This is most useful for screen captures, as it gives the user time to remove the mediarecorder application window from the display, or activate menus, before recording starts. The number of seconds is rounded up to the nearest value in the following list: 3, 5, 10, 30, 60, or 300 seconds. These values correspond to those found in the menus displayed by the graphical interface. -start Begin recording immediately upon starting, without waiting for user interaction. -duration n Causes recording to stop n seconds after it starts. -exitdone This option will cause mediarecorder to automatically exit after recording is finished. The application will exit regardless of whether it successfully records a media file, an error occurs while recording, or recording is aborted by the user. -movieoptions optionstring Set the parameters for movie files. This option is followed by a mandatory string containing any of the following suboptions, separated by commas. No whitespace is allowed: format= [ sgi | qt ] File format for movie file. qt specifies Apple Computer's QuickTime file format, and is the recommended file format. sgi specifies the older Silicon Graphics movie file format. The SGI format is supported primarily for backwards compatibility with applications running on earlier releases of IRIX. framesize= [ full | half | quarter ] Size of video frame to record. This is ignored when recording movies from the screen. imgrate= n This specifies not the actual frame rate, but an index into the list of frame rates which are valid at the time of capture. The first element of the list is zero, and is always the nominal frame rate (29.97 for NTSC, 25 for PAL). Successively higher numbers typically set integer dividends of the native frame rate (one-half nominal rate for n = 1, one-third nominal rate for n = 2, and so on). The nominal frame rate for most screen sources is the monitor refresh rate, typically 60, 72, or 76 Hz. If the index exceeds the end of the list of available frame rates, the last entry (lowest rate available) is used. compression= codecname codecname is a string naming a compression codec. The allowable codecs depend upon the movie format, and are found in the system include file <dmedia/dm_image.h>. spatialq= n Spatial quality factor for the codec. This is a decimal number between 0 and 1. Not all codecs support this feature, and it is ignored if not supported. temporalq= n Temporal quality factor for the codec. This is a decimal number between 0 and 1. Not all codecs support this feature, and it is ignored if not supported. keyframe= n Key frame distance for the image track in the output movie file. The movie file will be written with a key frame every n frames. Not all codecs support this feature, and it is ignored if not supported. colorspace= [ xbgr | xrgb | yuv422 ] Color space for image track of output movie file. xbgr specifies SGI-compatible 8-bits per component RGB. xrgb specifies QuickTime-compatible 8-bits per component RGB. yuv422 specifies YCbCr (YUV) color space with 4:2:2 color component sampling. rgb and yuv are retained for compatibility with earlier versions of the application, but their use is deprecated. rgb is a synonym for xbgr, and yuv is a synonym for yuv422. interlacing= [ even | odd | none ] Specifies field interlacing for output movie file. none specifies a movie recorded as frames, with no video field information. even specifies field encoding for PAL video, and odd specifies field encoding for NTSC video. Typically, you should only specify even or odd if you are recording JPEG, Motion JPEG A, or uncompressed video. audioenable= [ true | false ] True if audio track should be recorded in movie file, False if no audio track should be recording. audrate= n Sampling rate for movie audio track. The valid rates are found in the system include file <dmedia/audio.h>. audchannels= n Number of channels in movie audio track. This may be 1, 2, or 4. audwidth= n Width of samples for movie audio track. This may be 8, 16, or 24. droppedframeserror= [ true | false ] Determines if recording should abort if any frames are dropped while recording. screenadvisor= [ true | false ] Determines if the X screen advisor should be used when recording from the screen. Note that this option applies only when the device used for screen recording is something other than "xlib". When the device is "xlib" (the default), the screen advisor is always used and this option is ignored. For more information, see the section entitled "Screen Recording and the X Advisor" elsewhere in this manual page. The default is false. -imageoptions optionstring Sets the parameters for image files. This option is followed by a mandatory string containing any of the following suboptions, separated by commas. No whitespace is allowed: format= [ sgi | tiff | jfif | gif ] File format for output image file. framesize= [ full | half | quarter ] Size of video frame to record. This is ignored when recording a still image frame from the screen. -audiooptions optionstring Sets the parameters for audio files. This option is followed by a mandatory string containing any of the following suboptions, separated by commas. No whitespace is allowed: format= [ aiff | aifc | wave ] File format for output audio file. audrate= n Sampling rate for audio file. The valid rates are found in the system include file <dmedia/audio.h>. audchannels= n Number of channels to record into audio file. This may be 1, 2, or 4. audwidth= n Width of samples for audio file. This may be 8, 16, or 24. EXAMPLES To use mediarecorder to record a movie from live video, into a file called TEMPMOVIE which your application will later access, do this: mediarecorder -media movie -source video -allowmenu movie=video -fixedfile -outputfile TEMPMOVIE This will direct mediarecorder to record movies from live video and will prevent changing from that mode. It will place the resultant movie file in TEMPMOVIE. The user cannot change this name while mediarecorder is running. To use mediarecorder to record either still images or screen snapshots into a file called TEMPIMAGE which your application will later access, do this: mediarecorder -media image -source screen -allowmenu image=video+screen -fixedFile -outputfile TEMPIMAGE This will force screen snapshot mode and will allow the user to choose between screen snapshot mode and video still frame mode. It will place the resultant image file in TEMPIMAGE. The user cannot change this name while mediarecorder is running. To use mediarecorder to record QuickTime format movies from the IndyCam using Cinepak compression do this: mediarecorder -media movie -source video -jack indycam -movieoptions format=qt,compression="Apple Compact Video" This will direct mediarecorder to record movies from video, using the IndyCam if it is connected to the system. The resulting movie is saved in QuickTime movie format using Cinepak video compression. The user may change both the recording mode (movie, image, audio) and the source (video, screen) using mediarecorder's graphical interface. The name for the resulting file will be generated by mediarecorder. To use mediarecorder to record a movie file from the screen, then to run the program "myprog" whenever it writes a file, do this: mediarecorder -media movie -source screen -execstring "myprog %s" This will direct mediarecorder to capture a movie from the screen, using the default settings. Every time the user records a file, mediarecorder will execute the program "myprog", passing it the name of the file the user just captured. To automatically record a 10-second movie file from the screen using the X-based software facility, and have the application automatically exit when it is finished, do this: mediarecorder -media movie -source screen -delay 3 -duration 10 -start -exitdone -screensel method=adjustable,x=10,y=10,w=259,h=341 -movieoptions format=qt,compression="Apple Animation",colorspace=xrgb,interlacing=none,spatialq=0.8,temporalq=0.8,keyframe=24 To automatically record a 10-second NTSC video clip from the Octane or IMPACT compression board, to a movie file using Motion JPEG A, and have the application automatically exit when it is finished, do this: mediarecorder -media movie -source video -device impactcomp -delay 3 -duration 10 -start -exitdone -movieoptions format=qt,framesize=full,compression="Motion JPEGA",colorspace=yuv422,interlacing=odd,spatialq=0.8,temporalq=0.8 To automatically record a still image from the workstation display, to a SGI format image file, and have the application automatically exit when done, do this: mediarecorder -media image -source screen -delay 3 -start -exitdone -imageoptions format=sgi -screensel method=adjustable,x=15,y=15,w=243,h=591 To automatically record a 5 second AIFF audio clip from the default audio input, and have the application exit automatically when done, try this: mediarecorder -media audio -delay 3 -duration 5 -start -exitdone -audiooptions format=aiff,audrate=48000,audchannels=2,audwidth=16 NOTES Commonly, users wish to snap pictures or movies from a shell script, and embed the results in a web page. You can do this with mediarecorder, as it provides command line arguments to record files without human interaction. However, if you do not want an onscreen graphical interface, the vidtomem(1) or the dmrecord(1) utilities may be more suitable for your needs. mediarecorder does not create MPEG files. To create an MPEG file, record a QuickTime movie (uncompressed video would be ideal if you have the disk space) and use either the mediaconvert(1) graphical tool or the dmconvert(1) utility to convert the QuickTime file into an MPEG movie. When recording hardware JPEG-compressed movie files on an O2 system, from Octane Compression /IMPACT Compression, or from the Cosmo Compress board, choose SGI file format if you want a movie which is backwards- compatible with IRIX 5.3 and 6.2 systems. When mediarecorder records hardware JPEG files into QuickTime movies, it writes each frame as a pair of video fields. The resulting file will not play on IRIX 6.2 and earlier platforms. To record motion JPEG files which are compatible with QuickTime on other platforms, choose "QuickTime JPEG A" compression in the Movie Settings Dialog, or specify "Motion JPEGA" as the compression choice in the movieoptions options string from the command line. Keep in mind that movies created using Motion JPEG A compression will not be playable with releases of IRIX earlier than IRIX 6.5, except for IRIX 6.3 with the appropriate patch set. Though you may select video jacks from the menu built in to the application, in order to adjust video parameters such as signal timing (NTSC or PAL), hue, saturation, or video synchronization source, you must use vcp(1), the Video Control Panel. Likewise, to choose audio inputs for video and screen recording, or to adjust audio input levels, you must use audiopanel(1), the Audio Control Panel. mediarecorder requires overlay planes to run properly. Currently, all graphics hardware with the exception of Onyx4 UltimateVision support overlay planes by default. On Onyx4 UltimateVision, you must modify your XF86Config(4) file under the devices section, and change: Option "OpenGLOverlay" "off" to: Option "OpenGLOverlay" "on" Once overlays have been enabled, you must restart the X server for the changes to take effect. mediarecorder requires a root window with an OpenGL capable visual in order to capture from the screen. If the root window is not OpenGL capable, mediarecorder will generate an MVR GLX error when a capture is attempted. The workaround is to manually specify the visual using the -visid param in your Xservers file. Refer to the Xsgi(1) manpage for more information. FILES $TMPDIR/.vkme* Directories created by mediarecorder to hold temporary data during movie capture. They are removed immediately after mediarecorder writes the output movie file. SEE ALSO moviemaker(1), mediaplayer(1), audiopanel(1), vcp(1), soundtrack(1), cdplayer(1), vidtomem(1), mediaconvert(1), dmconvert(1), dmrecord(1), dmplay(1), xscreen(1), setmon(1) Page 15