|
addpath
| add new path for scripts
|
|
countbook
| count sets of atoms (decompositions) stored in a "book" file
|
|
echo
| echo a string
|
|
fft
| FFT spectrum of signal or residuum
|
|
list
| print command names w/1-line descriptions
|
|
loadasc
| load signal from ASCII file
|
|
loadascbook
| load ASCII file with MP results ("book")
|
|
loadbook
| load binary (v.III) file with MP results ("book")
|
|
loadfloat
| load signal from [multichannel multiplexed] file of floating point numbers
|
|
mp
| run Matching Pursuit on loaded signal, according to
set parameters
|
|
mp_chan_float_rott
| batch analysis: floats, "natural" border conditions
|
|
mp_chan_float_sim
| batch analysis: floats, simmetric border conditions
|
|
mp_chan_int_rott
| batch analysis: integers, "natural" border conditions
|
|
mp_chan_int_sim
| batch analysis: integers, simmetric border conditions
|
|
norm
| signal normalization: sig -> (sig-aver(sig))/var(sig)
|
|
path
| set search path for scripts (*.mps)
|
|
quit
| quit=exit
|
|
reconst
| save signal's reconstruction
|
|
reinit
| reinitialize the dictionary
|
|
reset_sig
| set current signal to zeros
|
|
resetdic
| reset probability distribution of scales to defaults
|
|
save
| save MP results (a "book") in binary (v.III) format
|
|
saveascbook
| save ASCII file with MP results ("book")
|
|
saveoldbook
| save MP results (a "book") in "old" (pre-v.III) format
|
|
set
| settings (decompositon & signal parameters etc.)
|
|
setinfo
| set INFO field for book's file header (v.III binary)
|
|
showbook
| display parameters of fitted atoms (current book)
|
|
showdic
| save dictionary's structure
|
|
showdist
| print density of atoms at scales
|
|
showinfo
| show current INFO field for book's file header
|
|
showtitle
| show currently set printout's title
|
|
tf2gif
| save time-frequency - Wigner(MP) - map, in GIF or raw format
|
|
tf2ps
| time-frequency - Wigner(MP) - map, in PostScript
|
|
title
| set printout's title
|
|
tunedic
| set probability distribution of scales
|
|
type
| type the signal and residuum values
|
|
ver
| program version info
|
|
view
| display t-f representation of current signal's decomposition (java)
|
|
write
| save signal/residuum to disk
|
| -O | filename | |
| -S | | simmetric border conditions |
| -L | | left border condition |
| -R | | right border condition |
| -C | | signal (default) |
| -O | filename | |
| -L | | create current signal from book's reconstruction |
| loadbook -
| load binary (v.III) file with MP results ("book")
|
| -O | filename | |
| -# | val | offset of book in file |
| -L | | create current signal from book's reconstruction |
|
loadfloat -
|
load signal from [multichannel multiplexed] file of floats
i.e. floating point 4-byte numbers (Intel)
|
| -O | filename | |
| -e | val | reference channel #1 |
| -f | val | reference channel #2 |
| -L | | left border condition |
-R | | right border condition |
| -C | | signal (default) |
| -# | val | offset |
| -c | chan | which channel (counting from 1, not 0) |
| -h | num_chan | number of channels in file |
| -s | shift | number of bytes to skip (e.g. file header) |
i.e. read from shift+offset*epoch_size*num_chan+chan
|
loadint -
| load signal from integer [multichannel multiplexed] file
|
| -O | filename | |
| -e | val | reference channel #1 |
| -f | val | reference channel #2 |
| -L | | left border condition |
| -R | | right border condition |
| -C | | signal (default) |
| -# | offset | |
| -c | chan | channel (counting from one, not zero) |
| -h | num_chan | number of channels in file |
| -s | shift | number of bytes to skip (e.g. file header) |
i.e. read from shift+offset*epoch_size*num_chan+chan
|
mp_chan_float_rott -
|
batch analysis:
floating point numbers (4 byte, Intel), "natural" border conditions, subsequent epochs of one channel,
multichannel data multiplexed.
Border conditions for each decomposition use most of the
information about previous and next epochs.
|
| -O | filename | signal file |
| -c | channel | which channel |
| -h | num_chan | number of channels in file |
| -# | offset | start from |
| -M | offset | stop at |
| -s | shift | |
| -I | filename | results (output) file ("book") |
| -R | | reinitialize dictionary before each decomposition |
| -e | r1 | reference channel (electrode) #1 |
| -f | r2 | reference channel (electrode) #2 |
i.e.: if set, x(channel) -> x(channel)-(x(r1)+x(r2))/2
|
mp_chan_float_sim -
|
batch analysis: floating point numbers (4-byte, Intel), simmetric border conditions,
subsequent epochs of one channel, multichannel data multiplexed,
simmetric border conditions.
|
| -O | filename | signal file |
| -c | channel | which channel |
| -h | num_chan | number of channels in file |
| -# | offset | start from |
| -M | offset | stop at |
| -s | shift | |
| -I | filename | results (output) file ("book") |
| -R | | reinitialize dictionary before each decomposition |
| -e | r1 | reference channel (electrode) #1 |
| -f | r2 | reference channel (electrode) #2 |
i.e.: if set, x(channel) -> x(channel)-(x(r1)+x(r2))/2
|
mp_chan_int_rott -
| batch analysis: integers,
"natural" border conditions, subsequent epochs of one channel, multichannel data
multiplexed, border conditions for each decomposition
use most of the information about previous and next epochs
|
| -O | filename | signal file |
| -c | channel | which channel |
| -h | num_chan | number of channels in file |
| -# | offset | start from |
| -M | offset | stop at |
| -s | shift | |
| -I | filename | results (output) file ("book") |
| -R | | reinitialize dictionary before each decomposition |
| -e | r1 | reference channel (electrode) #1 |
| -f | r2 | reference channel (electrode) #2 |
i.e.: if set, x(channel) -> x(channel)-(x(r1)+x(r2))/2
|
mp_chan_int_sim -
| batch analysis:
integers, simmetric border conditions,
subsequent epochs of one channel,
multichannel data multiplexed
|
| -O | filename | signal file |
| -c | channel | which channel |
| -h | num_chan | number of channels in file |
| -# | offset | start from |
| -M | offset | stop at |
| -s | shift | |
| -I | filename | results (output) file ("book") |
| -R | | reinitialize dictionary before each decomposition |
| -e | r1 | reference channel (electrode) #1 |
| -f | r2 | reference channel (electrode) #2 |
i.e.: if set, x(channel) -> x(channel)-(x(r1)+x(r2))/2
|
norm -
| normalization of current signal,
substracting average and dividing by variance:
sig -> (sig-aver(sig))/var(sig)
|
|
path (addpath) -
| set (add to) search path for scripts.
Scripts (in examples named *.mps) - ASCII files,
containing hereby listed commands - can be used to
extend functionality e.g. by calling external programs, or to perform
batch decompositions of large amounts of data.
|
|
reconst -
| save signal's reconstruction
|
| -O | filename | |
| -S | | all calculations (default) |
| -M | | save table with all fitted atoms in time domain |
| -R | | save residuum |
| -m | val | reconstruct atom fitted as number val only |
|
reinit -
| reinitialize the dictionary |
| -R | val | number of (Gabor) atoms |
| -i | + | - | initialize random numbers generator |
| -d | + | - | use dyadic dictionary structure (scale = 2^j) |
|
resetdic -
| reset probability distribution
of scales to defaults, i.e. return to the default distribution of
the parameters of dictionary's atoms
|
|
save -
| save MP results (a "book") in binary (v.III) format
|
| -A | file | add current book to file |
| -a | | add current book to file book.b |
| -S | file | create a new file with current book |
| -s | | create new file 'book.b' with current book |
| -A | file | add current book to file |
| -a | | add current book to file book.b |
| -S | file | create a new file with current book |
| -s | | create new file 'book.b' with current book |
| saveoldbook -
| save MP results (a "book") in "old" (pre-v.III) format |
| -A | file | add current book to file |
| -a | | add current book to file book.b |
| -S | file | create a new file with current book |
| -s | | create new file 'book.b' with current book |
| -M | val | max number of algorithm's iterations |
| -A | val | increase max number of algorithm's iterations by val |
| -E | val | stop iterations after explaining val % of energy |
| -P | +|- | messages on|off (+|-) |
| -I | +|- | decomposition progress info on|off (+|-) |
| -O | val | signal size (val points) |
| -F | val | signal sampling frequency [Hz] |
| -C | val | signal calibration constant [usually points/microVolt] |
| -B | val | book filename |
| -T | | use dyadic dictionary (default stochastic) |
| -S | val | oversampling (for dyadic dictionary) |
| -h | +|- | use heuristic acceleration |
| -N | val | number of atoms for trial dictionary |
| -o | val | range of scales for heuristic |
| -f | val | range of frequencies for heuristic |
| -p | val | range of atom's positions for heuristic |
|
setinfo -
| set text for the INFO field
for book's file header (v.III binary). This ASCII field is stored
in the binary "book" with results of MP decomposition (v.III)
and can be used for
any comments re. decomposed signal etc. For the complete format
of binary files with MP results please consult accompanying examples
of programs on C.
|
|
showdic -
| save dictionary's structure
|
| -O | filename | (default dict.dat)
|
| -r | + | - | histogram |
| -d | + | - | atom's parameters |
|
showinfo -
|
show current text to be written for the INFO field
of book's file header (v.III binary).
This ASCII field is stored in the binary "book"
with results of MP decomposition (v.III) and can be used for
any comments re. decomposed signal etc. For the complete format
of binary files with MP results please consult accompanying examples
of programs in C.
|
| tf2gif -
| - time-frequency - Wigner(MP) - map, in GIF or raw format
|
| -O | file | output GIF filename |
| -F | val | gamma correction coefficient (default 1) |
| -L | | log energy scale |
| -B | file | save "raw" t-f map (floats) in file |
| -b | | save "raw" t-f map (floats) in 'image.raw' |
| -X | val | horizontal dimension (points) |
| -Y | val | vertical dimension (points) |
| -a | file | average map using all the atoms in file |
| -s | | add (currently loaded) signal to the averaged map |
| -d | | map of dispersion (averaged map) |
| -l | val | laplasjan 0--3 |
| -R | | reverse input signal on plot (sig -> -sig) |
| -S | val | (default 1) show lower 1/val of frequencies |
| -c | | color (default) |
| -g | | shades of gray |
| -k | val | cut off high values - may improve dynamics, experimental |
| -x | val | string on printout - default as set by 'title' |
| tf2ps -
| time-frequency - Wigner(MP) - map, in PostScript
|
| -O | file | output PS filename |
| -F | val | gamma correction coefficient (default 1) |
| -L | | log energy scale |
| -B | file | save "raw" t-f map (floats) in file |
| -b | | save "raw" t-f map (floats) in 'image.raw' |
| -X | val | horizontal dimension (points) |
| -Y | val | vertical dimension (points) |
| -a | file | average map using all the atoms in file |
| -s | | add (currently loaded) signal to the averaged map |
| -d | | map of dispersion (averaged map) |
| -l | val | laplasjan 0--3 |
| -R | | reverse input signal on plot (sig -> -sig) |
| -S | val | (default 1) show lower 1/val of frequencies |
| -c | | color (default) |
| -g | | shades of gray |
| -k | val | cut off high values - may improve dynamics, experimental |
| -x | val | string on printout - default as set by 'title' |
| -N | | no blobs numbers on printout |
TO DECIDE:
-i val - (integer) font size in points (default 8)
-p val - number of points per microVolt
-g val - max number of pages on book printout
-F val - gamma correction coeff. (default 1)
-B file - read atoms ("book") from file
| tunedic -
| set probability distribution of scales
|
| -f | filename | distribution from file |
| -o | val,val | distribution for a single element |
|
view -
| display t-f representation of current signal's decomposition (java).
If java binaries (v. 1.2 or newer) is installed in your system,
java application will be called,
allowing e.g saving of the interactively reconstructed signal.
Otherwise, netscape is called with security restrictions.
To use this function please install java binaries (jdk, jre etc.)
or netscape v. 4.6 or newer.
|
| option(s): |
| -v | jdk | jre | hotjava | netscape | auto | (default auto) |
| write -
| save signal/residuum to disk file
|
| -O | filename | (default 'signal.asc') |
| -x | | save original signal (default residuum) |
| -a | | include border conditions |
<<< A SIMPLE EXAMPLE >>>
this example illustrates basic usage of commands for loading
a signal, running MP decomposition and viewing results
in the time-frequency plane
Suppose we have the ASCII file chirp2.txt in the current
directory; let's load it:
loadasc -O chirp2.txt
If the program does not complain about the size of the signal being
different from the basis dimension (for which the dictionary is constructed,
default 512 points, changed with
set -O),
we can run Matching Pursuit (with default simmetric border
conditions) by simply typing
mp
After some time ;-) when the iterations are finished
(the progress is displayed on screen, unless you
set -P- or set -I-)
you may try
view
If java binaries or netscape v. 4.6 or newer is installed on Your machine,
a window should appear, displaying the computed representation, and
letting yoy click&play with reconstructions. That's it!
This representation was computed with the default dictionary of
70000 Gabor functions, drawn from flat distributions of parameters,
plus complete Dirac and Fourier bases (stochastic dictionary),
no heuristic acceleration and settings equivalent to
set -M 32
set -E 95
i.e. stop after 32 iterations or after explaining 95% of signal's energy,
whichever comes first.
To enjoy a more detailed representation,
we can type e.g.:
reinit -R 1000000
set -M 100
set -E 99
mp
view
...which uses dictionary of 10^6 Gabor atoms, and tries to parametrize 99%
of signal's energy, unless more than 100 iterations would be required.