** PURPOSE OF MANAUTON **

The purpose of Manauton is to provide digitally recording of sound in
either a manual or autonomous mode, hence the name Manauton.

When in manual mode, recording can be paused and un-paused with a
key-press or remote control. This is similar to a tape recorder, but
with a difference, there is no latency! As a matter of fact, Manauton
works with negative latency. This negative latency cancels out the
effect of the human latency associated with hearing a sound
interpreting it and reacting to the sound. Negative latency is
accomplished by buffering the sound in memory prior to recording to
disk.

When operating in autonomous mode, Manauton starts and stops the
recording automatically based on the sound


  ** RUNNING MANAUTON LIKE A STANDARD RECORDER ** 

manauton filename.wav 

Will cause Manauton to function similar to a standard tape recorder.
To start recording hit the r key to pause recording hit p. You can
quit recording use the q key, Manauton allows you to quit in when in
pause or record modes.


  ** THE CLICK-BEEP **

By default whenever you start record a section of sound the section is
preceded with a click-beep.  This goes for the first section of sound
along with all subsequent sections.  The click-beep is a artifact that
will allow you to distinguish sections that were recorded
separately. Using a digital sound editor the shape of the click-beep
should easy distinguished. The click-beep also includes the encoded
time and date when the section started at. A click-beep interpreter
should be released shortly that will display all times and dates
embedded within a sound sample. If you wish to disable the click-beep
use the --tag-off option.


  ** CHAINED or UNCHAINED ** 

By default all sections of sound you record in a single session of
Manauton will be chained into a single wave file. If you would like
Manauton to place each section in a separate file use the --unchained
option.


  ** RUNNING MANAUTON WITH NEGATIVE LATENCY ** 

manauton --latency=2.0 filename.wav

This command causes Manauton to run with two seconds of negative
latency.  The 2.0 can of course be any positive floating point number.
The amount of negative latency should should be limited by computers
RAM (approximately 11 seconds per megabyte).

Once Manauton has been started in this way, it will record sounds from
two seconds before you hit r to the time you hit p or q.  This is
beneficial because when you hear a sound it will take time for your
brain to interpret the sound and tell your finger to press the
button. Negative latency allows you time to consider the sound and
decide whether or not to capture it.

Manauton can't record anything that happened before the program has
started.  In the event that r is pressed before the latency time has
passed, Manauton will record from time that it was started until you
press q or p.

In the event that you pause the recording then restart recording
before latency time has expired, Manauton will join the sound in a
single continuous stream without the inclusion of a click-beep.


  ** CHANGING SAMPLE RATE AND FORMAT OF SAMPLE **

The default format for the sample is 44.1K samples/second 16 bit
signed.  use --rate=# to change rate. the option -8 will cause the
sample to be stored in 8-bit unsigned format (use --help to see if
this feature has been implemented yet).


  ** USING AN EXTERNAL CONTROLLER **

manauton --controller=smsmouse filename

This command will allow you to use an external controller with
Manauton. In this case the controller is a Microsoft mouse in serial
port known to Linux as /dev/ttyS0.  The right button is record, left
is pause and middle is quit.  It should be noted that this will not
work if the same mouse is used for the X-window system or GSM, these
must either be disabled or a second mouse must be attached.

If you wish to use the other serial port you can use
"--controller=smsmouse1". If you wish to re-map the function of the
mouse buttons you can use the command "--controller=smsmouse#,r,p,q"
where # is replaced by the serial port ether 0 or 1. The letters r,p,q
are replaced each by the button values for record,pause,quit. Key
values are 0-none, 1-left, 2-middle, 4-right, 8-wheel_down,
16-Wheel_up.  By adding the values together you can signify a button
combination.

In the future new controllers may be added. Use the command "manauton
--controller" to get a list of available controllers.


Example 1 : --controller=smsmouse0,1,4,2
/dev/ttys0, record on left, pause on right, quit on middle 

Example 2 : --controller=smsmouse1,4,0,1
/dev/ttys1, record on right , pause on release, quit on left 

Example 3 : --controller=smsmouse0,1,4,5
/dev/ttys0, record on left, pause on right, hold both to quit


  ** USING AUTONOMOUS MODE ** 

manauton --criteria=sspike --pdelay==2.0 filename.wav

This command in used to put Manauton in autonomous mode using the
single spike algorithm.  Manauton will start recording when a single
sound peek reaches 50% of the maximum value (or greater ). As long as
sound spikes reach 50% (or greater) Manauton will continue recording.
Once the sound spikes no longer reach the 50% mark Manauton will wait
2.0 seconds then go into pause mode, where it will wait to be
re-triggering.

With the option "--criteria=sspikex,y", x and y can each be a number
from 0 to 100, where x in the percentage to start recording and y is
the percentages to continue recording. If only a single number is
given it will represent both the percentage to start and continue
recording.

In the future new autonomous algorithms may be added. Use the command
"manauton --criteria" to get a list of available algorithms.

The "--pdelay" option can be left off, however this will result in a
choppy sample. When the option "--latency=" is used Manauton will
record sound prior to the trigger. The option "--pdelay" can also be
used in manual mode, it will cause Manauton to continue to record
after you hit pause.

A new trigger mode with the form "--criteria=hitcountx,y" has been
added to manauton. The purpose of this new option is to avoid
triggering due to extremely brief sounds, such as pops or clicks.  The
number y is the target amplitude which is expressed as a percentage of
the maximum obtainable value.  The value x indicates the percentage of
samples within the frame which must have an amplitude exceeding or
matching the target value. I have found that a value of 10,30 works
best value to begin an experiment with. If you wish to adjust sound
discrimination based on time length, you should adjust the first
value.  If you wish to adjust sound discrimination based on volume,
you should adjust the second value.

When used with feedback (see next section), the master trigger level
indicates the relative number of times per frame in which the sample
exceeds the target amplitude. Level 2 indicates the maximum amplitude
of all samples within the frame relative to the target amplitude. When
level 2 reaches the middle level there is at least one sample that has
reached the target amplitude. Triggering will only take place if the
Master level reaches the center mark.


  ** VIEWING SOUND LEVEL FEEDBACK **

If you wish to use the curses feedback mode you will need to use the
two options "--feedback=curses" and "--controller=curses" unless you
wish to use an external controller. If you do use an external
controller, then substitute "--controller=curses" with the appropriate
controller mode. Because of the architecture of Manauton it is
necessary to use the controller command when in curses mode. Future
versions may not require it.

The meter labeled channel 1 is proportional to the amplitude of the
sound being recorded. A full meter indicates that the sound level has
"maxed out".


  ** FEEDBACK DURING AUTONOMOUS MODE ** 

The meter labeled level 1 is the master trigger level.  The center of
the meter is marked by a carrot symbol "^". When the level bar reaches
or exceeds the center mark the trigger has been activated. Some
activation modes may have multiple trigger level indicators. In this
case, only the master trigger level will indicate that triggering has
taken place. It is possible that one or more of the sub-triggers reach
the center without the master trigger reaching the center, in which
case the Manauton auto recording will not be triggered. As a rule of
thumb when the master trigger or sub-trigger reaches the center it
indicates that it has obtained 100% of the quantity needed for
triggering. When the bar within the (sub)trigger indicator is full it
indicates that 200% of the triggering value has been reached.


 LocalWords:  manauton filename wav pdelay smsmouse sspike sspikex hitcountx
 LocalWords:  GSM