------------------------- MIXVIEW MANUAL --------------------------------- Introduction Usage: mixview [-D display -s inskip -d duration] [soundfile] Mixview is an interactive, graphics-based soundfile editing program. It will eventually incorporate all cmix sound-processing routines. It is built on the X Window System library and the lxt toolkit, designed by Roger Davis. It is invoked as indicated above. If any options are specified, a sfile name must also be supplied. If it is invoked with no options & no filename, it will create an untitled default file (in memory only). Mixview is mapped to the screen as two windows-a waveform display window and a control panel window. The display window shows the soundfile at the current vertical and horizontal resolution, and above this is a menu bar with seven pulldown menus. An additional menu is popped up using the right mouse button in the display window. The control panel window is divided into two subpanels, the first with scrolling/paging controls and displays, and the second with display resolution controls and additional function buttons. After the windows are both located on the screen, the display window may be resized at will, using whatever resizing technique your window manager uses. You must also then use the resize button to update the display (see below). Mixview expects to find an environmental variable called SFDIR, which is set to the directory in which you keep your soundfiles. SFDIR can be the top of a tree containing several soundfile subdirs if so desired. If no SFDIR variable is found, the current working dir is searched for soundfiles. IMPORTANT NOTE: Mixview requires three mouse buttons to function. If you are working with a machine and server that provide no button 3 action, contact me at the following email address, and I will probably arrange a workaround in mixview. I have tested mixview under both twm and uwm window managers. The version of twm in X11R3 is flawed and will not work with mixview. X11R4 is fine. doug@woof.columbia.edu ______________________________________________________________________________ *****************************Sound Files************************************* Mixview is designed to read 16-bit short integer and 32-bit floating point soundfiles with either the BSD soundfile header (as used and generated by csound and cmix) or, if you are working on a NeXT, the NeXT soundfile header. On NeXT workstations and SPARC workstations with audio capability, 8-bit mulaw encoded (CODEC) soundfiles may also be read in, where they are CONVERTED AUTOMATICALLY to 16-bit files. These may also be played back on these machines. ______________________________________________________________________________ ****************************** COMMANDS ************************************** A) Paging/ Scrolling Operations 1) Key Commands (when cursor is in display window): f - forward 1/4 screen b - back 1/4 screen F - forward 1 screen (actually 9/10) B - backward 1 screen (actually 9/10) NOTE: F and B may be prefaced with a number, i.e., "12F" will go forward 12 pages. 2) Panel items for scrolling: Scrollbar Drag or click on grey bar to move to any location in the file. Percent of your way through the file is indicated. Paging Buttons (to either side of scrollbar) When clicked, will scroll 1/4 screen forward or back, as above. Goto Button Type in a time (erase current numbers first by backspacing over them) then click button to set goto time to center of screen. If time is left blank, the current insert point will be set to the screen center. If no insert point is set, the display will go to the beginning of the file. Set Screen Times (Item in canvas menu -- see below) If the "Set Screen Times" item is selected from the canvas menu, a dialog window will prompt the user for beginning and ending times for the display screen. The horizontal resolution will be adjusted accordingly. The times may be adjusted slightly by the program due to internal calculations. B) Display Control 1) Horizontal Resolution (grainsize): "grain" bar drag or click to set to desired grainsize. This scale is logarithmic from an arbitrary value of "0" (1/32 sample per pixel) to "1024" (8192 samples per pixel). x in & x out these buttons halve and double, respectively, the current grainsize. "zoom" used in conjunction with the region-selection operations (see below), this allows any selected region of the screen to be expanded to fill the screen. Also sets grainsize after using "+" and "-" (see above). If hit again it will unzoom to original, as long as no other grainsize changed have been made. 2) Vertical Resolution (amplitude scale): Y in & Y out These buttons halve and double, respectively, the amplitude scale. Useful for examining low-level sections of a soundfile. To reset the screen to the default scale, click on the "reset" button (see below). Also see "Setting the Vertical Scale". 3) Resizing Screen: Resize Button At any time when working with mixview, including when it is first mapped to the screen, you may resize the entire frame to any size convenient for you. The control and information panels will remain a constant height -- but the viewscreen will resize to as large or as small as you wish. To redisplay the file with a larger screen, press the "resize" button. If resizing to a smaller screen, this SHOULD happen automatically, but it depends on the window manager that you use. Mixview works best under twm. 4) Resetting Screen: Reset Button The "reset" button serves to redisplay the screen at the default vertical scaling and without insert points or edit regions displayed. It is also useful for simply refreshing the display if it is disturbed in any way. 5) Changing the Soundfile Window Cursor The default for the soundfile window is a vertical hairline, indicating the current insert time, once the mouse is clicked in the window. If the CURSOR MODE item is selected from the small menu in the window (right mouse button), a pullright menu will allow the choice of vertical and horizontal crosshairs. When this is done, the amplitude of the waveform at the insert point will also be displayed. 6) Changing the Display Mode The default display mode is "Line", which means the the waveform is drawn as a continuous line, oscilloscope-style. If the "Display Mode" item is selected from the small menu in the window (right mouse button), a pullright menu will allow the choice of a shadow-style display, in which the space between 0.0 and the amplitude values is filled in with the foreground color. This improves the visibility of certain types of waveforms at certain resolutions. 7) Setting the Grainsize When the "Set grainsize" item is selected from the canvas menu, a dialog window will prompt the user to enter a specific grainsize. The value may be adjusted slightly by the program due to internal calculations. The default is the current value. 8) Setting the Vertical Scale When the "Set vertical scale" item is selected from the canvas menu, a dialog window will prompt the user to enter a specific amplitude that will represent the maximum + and - values on the screen. The default is +- current_peak_amp. C) Editing Operations: Setting Edit Points 1) Insert Point The insert point may be set by clicking the left mouse button at any point on the viewscreen. The insert time will be displayed just to the right of the vertical hair that is drawn. Using the "goto" button will also set the insert point at the goto time (see above). 2) Edit Regions The edit region may be set by clicking the middle mouse button anywhere to the LEFT of the insert point. A reverse-video box will highlight the selected region of the file, and the end time of the region will be displayed to the right of the box edge. This region may be dragged to any size by holding down the middle button and moving the pointer (minimum size is one grain at current resolution). To set the end point off-screen, set the insert point, then scroll to the desired endpoint and click the middle button. If a region has been selected, the "goto" button may be used to reset the endpoint of the region to any time after the insert point. To select the entire file, click the right mouse button in the soundfile window, and chose the "select all" option under "edit special" in the small menu that will appear. D) Editing Operations: Adjusting Edit Points 1) Insert Point (coarse adjst.) Click or drag the pointer to any new location using the left mouse button. To move by large amounts, use the goto button. 2) Insert Point (fine adjst.) Use keyboard commands "h" (back 1 grain) and "l" (forward 1 grain). These are modeled after vi commands. NOTE: As with all screen keybd. commands, these will only work when the pointer is inside the bounds of the soundfile window. h and l may be prefaced with a number, i.e., "12h" will go backward 12 grains. 3) Edit Region Endpt. (coarse) Drag the middle button as described above. 4) Edit Region Endpt. (fine) Use keyboard commands "+" (increase by 1 grain) and "-" (decrease by 1 grain). The key is not needed to produce the plus command. Also see above note. 5) Edit Region (Total Area) The "l" and "h" commands will shift the entire region in the same manner as it does the insert point (assuming the region has already been set on the screen). 6) Unset Region If the or key is pressed when a box is displayed, the region will be "unset" and the insert point will be set to the old region's start time. If the bar is pressed instead, the insert point will be set to the first grain AFTER the end of the old region. 7) Edit Special -> Recall Edit If this command is selected from the small menu in the soundfile window (popped up by a right mouse button press), the previously selected region in which an edit operation took place will be re- selected. This does not include zoom regions or play regions. 8) Edit Special -> Select All See the notes preceding this section. E) Editing Commands: An overview The following three sections deal with the commands for editing and processing the soundfile. Most of these commands will display a pop-up dialog window which will prompt the user for certain editing or processing parameters (see section N). 1) The Data-Curve Entry Window Many of the commands for writing buffers into the file have an option for fixed or curve amplitude values. If the "fixed" option is chosen (it is the default), the buffer will be written out to the file multiplied by the amplitude value set in the dialog window. If the "curve" option is chosen, upon confirming the edit, a second box will be displayed in which the user may draw a curve which will be mapped over the duration of the edit region, and will determine the amplitude envelope for the buffer. Once the curve-entry window is displayed, the user may enter the curve as a series of connected segments or as a freely drawn line, depending on the mouse button used. The RESET button will set the curve to all zeros. To change the maximum and minimum values for the curve, select "set range" from the small menu displayed when the right mouse button is depressed. NOTE: This window will soon be replaced with the new generic data-entry window (see the Pitch Track command, below). 2) The Multiple Buffer Option These same buffer-writing commands have an option to write multiple copies of the buffer to the file. These will be written one after another, and are first stored in the cut buffer so that a curve (see above) may be placed over the entire set of copies. This set of copies is left in the cut buffer and may be used again, until another read operation is done. F) Editing Commands: EDIT Menu Many of the commands in mixview have keyboard equivalents, indicated in []. The cursor MUST be in the display window (canvas) for these to be received by the program. 1) Buffer-> allows you to chose between an internal cut buffer (that you have already created) or an external one (selected region in another window). See Transferring Data Between Windows. 2) Copy [^c] copies the selected region into the internal cut buffer. 3) Remove [^r] copies the selected region into the buffer and then erases (sets to zeros) the region. 4) Splice out [^x] copies the region into the buffer and splices out the original region, i.e., shortens the file by the length of the region and moves the remainder to the left. 5) Erase [^e] erases region WITHOUT storing it in buffer. This cannot be undone. 6) Zap [^z] splices out the region WITHOUT storing it. This cannot be undone. 7) Mix [^m] adds the contents of the cut buffer, sample by sample, to the file beginning at the insertion point. 8) Replace [^R] replaces (destructively writes) the contents of the cut buffer to the file beginning at the insertion point. 9) Splice in [^v] inserts the contents of the cut buffer at the insertion point, shifting the contents to the right. Opposite of splice out. 10) Multiply [^M] multiplies the contents of the cut buffer, sample by sample, to the file beginning at the insertion point. This effectively ring-modulates the file by the buffer. 11) Crossfade [^C] combines the contents of the cut buffer with the file, beginning at the insertion point, according to the curve drawn in the accompanying window. A curve value of 1.0 will yield 100% of the cut buffer; 0.0 will give 100% of the original file, and anything in between will result in a combination of the two. G) Editing Commands: ALTER Menu 1) Phrase multiplies the selected region by amplitude factor. 2) Insert space allows the user to splice in any number of second worth of silence, starting at the insert point. NOTE: This is the only alter- style command that needs an insert point rather than a selected region. 3) Reverse reverses the selected region in time. 4) Transpose transposes the selected region by either an equal-tempered or linear-octave interval, or by any frequency ratio. The resulting length (since this is a time-rate shift) will be written into the old space (which is zeroed), and the remainder (if longer) can be either mixed into the file, destructively written in, or spliced in. 5) Gravy allows independent time dialation/contraction and pitch transposition through the technique of framing. The size of the frames can be varied, as well as the proportion of bevel. The ratio of output length to input length can be set as well as the interval of transposition. The resulting length will be written as discussed in transpose, above. H) Editing Commands: Filter Menu 1) Elliptical filter Selected region is processed by a variable- length elliptical filter, which is capable of very sharp rolloff curves, and may function as a high, low, or bandpass filter with variable stopband ripple and attenuation. NOTE: This filter is not available on systems that do not have a FORTRAN compiler. 2) Resonant filter Selected region is processed by a second-order resonant filter. Center frequency, bandwidth, and gain-mode can be set. 3) Allpass (not yet implemented). 4) Comb Selected region is filtered by a comb filter. Value can be set as either center frequency in CPS or as a loop (delay) time in seconds. The Q of the filter is measured as the time it takes the impulse response to fall by 60 db. 5) Tone filter Selected region is filtered by a first-order lowpass or highpass filter. Cutoff frequency and mode (hi or lopass) can be set. I) Mixview Options: OPTIONS menu 1) Peak rescan determines whether file is rescanned for peak amplitude after an editing operation. This may be set to "On" (automatic), "Off", or "Query" (program will ask for confirmation). NOTE: Be sure to scan the file before rescaling or saving to disk, otherwise serious distortion may occur. 2) Quick display allows large sections of sound to be displayed (albeit VERY sketchily) at a much faster rate. 3) Soundfile Search user has the option of 1) displaying only files that end with the '.snd' (on NeXT), '.au' (on SPARC), or '.short' (BSD) suffix, 2) displaying all soundfiles with valid header, or 3) displaying all files in the selected SFDIR directory. J) Global Commands: HEADER menu 1) Soundfile Information displays a window with all current data about soundfile: sfdir, name, sample rate, length, file size in Megabytes, soundfile type, and number of channels. 2) Sample format allows user to convert file from one format to another, i.e., floating point to short int or short int to 8-bit mu-law. NOTE: Do not edit soundfiles while in mu-law format. It is only for playing and saving to disk. 3) Sampling rate allows user to change the sampling rate of the soundfile. 4) Rescale sets the peak amplitude of the file to +-32767 (for short samps) and +-32676.0 (for float samps). 5) Change length allows the user to lengthen or shorten the soundfile to any size (in seconds). It will give a warning if insufficient memory is available. 6) Change sf comment pops up a text window into which the user may type up to 512 chars worth of text, to be stored in the soundfile header as a comment. There are various text-editing commands available from a small menu that may be seen by holding down the right mouse button in the text window. To save the comment, select "Save" and under it, the "/tmp/comment.***" item, then select "Destroy" to remove the text window. K) Analysis Commands: 1) Histogram (not yet implemented) 2) FFT user can do a Fast-Fourier Transform on a selected region at various resolutions. An output datafile will be created (default is soundfile name with ".fft" suffix added). If "Display data?" is set to yes, a data- display window will be mapped on the screen with the amplitude (either linear or decibel) and phase displayed as graphs. 3) LPC allows user to run a Linear-Predictive Coding analysis of the selected region. This uses the LPC program developed by Paul Lansky at Princeton. A dialog window will let you set the usual parameters (see man page for LPC) and select an output datafile name (default is soundfile name with ".lpc" suffix added). If "Display data?" is set to yes, a data- display window will be mapped on the screen with the first four lpc values (RMS1, RMS2, error, and pitch) displayed as graphs. 4) Pitch Track allows user to run a pitch tracking analysis of the selected region, storing it in a datafile of the user's choice (default is soundfile name with ".pt" suffix added). A dialog window will let you set the usual parameters. If "Display data?" is set to yes, a data-display window will be mapped on the screen with the pitch and amplitude data displayed as graphs. L) File Commands: FILE Menu 1) New allows user to create an empty file in memory (not written to disk) for scratch use or for building up a file from external buffers (see below). Sample rate, number of channels, and sample size can be set - as well as the name. 2) Change Sfdir prompts the user for a new soundfile directory, which will then be the base menu in the "Open" item (next). This is also used to rescan the current directory for new entries (those created since the last call to this command). 3) Open -> opens a new (but already existing) file from disk. Select file from menus that can be dragged to right of -> items. It will ask for confirmation before closing a previous unsaved file. 4) Save saves current file to disk under its original soundfile name. This does not ask for confirmation. 5) Save To saves current file to another name, and will ask for confirmation on overwrites. NOTE: The currently open soundfile will retain its name and identity. To edit the newly-saved file, you must close the current file, rescan the soundfile directory (see above), and open the new file. NOTE: To save an 8-bit file (that has been converted to 16-bit) back to an 8-bit version, you must convert the sample format back to 8-bit first, and then save. Soundfiles may be saved as BSD-style or native (on NeXT and SPARC). 6) Close closes current file. It will ask for confirmation if the file has been altered since the last save or save-to. 7) Revert reverts to last saved version of file. It will ask for confirmation if file has been altered since the last save or save-to. 8) Print Version # displays the current version number of mixview. 9) Exit functions the same as the "Exit" button, but is the only way to exit when there is no current soundfile open (the panels are inactive at that time). M) Transferring Data Between Windows: Selecting a region: This assumes you already have two mixview applications currently open, with an open file in each. The files can be new or pre-existing. Highlight the section of the file you wish to transfer in the window displaying that file. Setting up the transfer: Move to the window into which you wish to transfer the selection. Open the EDIT menu and select Buffer, dragging the pullright menu out to display "Internal" and "External", and select "External". Transferring the selection: Set an insert point as you would for an internal operation, and chose any of the edit operations that are active. The highlighted rectangle in the first window will disappear, indicating that the region has been copied into the buffer. After a pause, the selection will be mixed, written, or spliced into the new file. This buffer is now available as a local buffer until a new copy, erase, etc., is performed. To repeat an external transfer, "External" must be chosen again. NOTE: You can also move sound buffers from file to file by doing the following: 1) COPY, REMOVE, or SPLICE OUT a section of one file, 2) Close that file (but do not quit mixview). 3) Select OPEN from the FILE menu and open another file, or select NEW and create a new file. 4) The buffer from the previous file can now be combined with this file -- as long as the sample size and number of channels are the same. M) Playing and Recording Soundfiles: Soundfiles may be played via DACs on NeXT workstations and a selected few others. If a section of the file is selected (via highlighting), it will be played. If no region is selected, the portion visible on the screen will be heard. If the vertical hairline is set to any time in the visible window, the file will be played from that point to the end of the window. NOTE: CODEC, short integer and floating point files may be played. The latter takes a bit longer, as mixview must rescale the sound first. The record feature in mixview is not yet implemented. On SPARCStations with the SOUND software installed, the volume and output (speaker or headphone) may be set using the slider and switch on the control panel (these are only visible on these systems). N) Dialog Panels: Just a few notes: For any panel that allows test entry (either numbers or text strings), the default button process (the button with all CAPITAL letters) can be activated via . This is not yet so for the other panels. To move from one text-entry to the next in the case of multiple texts, you may either use the mouse (clicking on new entry) or cycle through using . For more information on how these things work, see the lxt.ms file in the lib directory. O) Data Windows: These are under development to replace all other display windows in the program. Currently they are display-only, and show plots of the datafiles produced by fft, ptrack and lpc. The windows are resizable to any dimensions, and up to 16 may be displayed simultaneously. Several commands are available through a menu that can be popped up via the right mouse button if it is pressed while in the display area: Close - Closes the currently selected data window. Set vert scale - Change scaling for selected channel of data. Set horiz scale - Change horiz scaling for all channels. Edit - (Not functional yet) Save changes - Writes file out (again) to disk. Save to new - (Not functional yet) Read file - (Not functional yet) There are also 2 buttons in association with the panels. The first simply redraws the graphs, the second attempts to fit all the data into the smallest horizontal space possible, for viewing. P) Other keyboard equivalents: While in the soundfile display window, many commands have keystroke equivalents. The editing ones are mentioned above. Here are some others. - S Save file. - C Close file. - Q Exit mixview. - V Show current version number. - P Play current selection or display. - A Select entire file. - 6 (carat) Zoom in on selected region (and unzoom). - 8 (asterisk) Reset display. Left arrow Halve the current horizontal resolution. Right arrow Double " " " ". Down arrow Halve the current vertical resolution. Up arrow Double " " " ".