AIPS HELP file for QMSPL in 31DEC09
As of Sun Nov 22 17:17:41 2009
QMSPL: Task to send a plot file to the QMS printer/plotter
INPUTS
INNAME Image name (name)
INCLASS Image name (class)
INSEQ 0.0 9999.0 Image name (seq. #)
INDISK 0.0 9.0 Disk drive #
PLVER 0.0 46655.0 Version # of PL file. 0=>last
ASPMM 0.0 Arc sec. per mm. 0=self scale
COPIES 0.0 200.0 Number of copies needed.
LPEN 0.0 20.0 Pen width (dots). 0 => 3.
OUTFILE
Save output in file named ..
' ' => print and delete
OPTYPE Type of grey scale encoding
'NONE','QRAN','MRAN','RAND',
'BITS' -- else 'MBIT'
FUNCTYPE 'NE','LG','NG','PL','NL'
else linear
FACTOR Power law index for use with
'PL' or 'NL' FUNCTYPE's
AXVAL (1) Input and (2) output
values at inflection point
DPARM (1,2) Clip recorded greys
before FUNCTYPE (0 to 1)
(3,4) After FUNCTYPE scale
by g*DPARM(3) + DPARM(4)
Parms for .BIT. modes:
(5) 0, 2 => ignore DPARM
1, 3 => use DPARM
0, 1 => Gaussian mask
2, 3 => Sine wave mask
(6) size of mask in dots
(7) Linear scale factor
(8) x-shift of function
(9) y-shift of function
HELP SECTION
QMSPL
Type: Task
Use: QMSPL will find an extension file of type PL and use the
graphics commands in that file to display the image on
the QMS Lasergaphix 800 or 1200. The print file produced
by QMSPL cannot be printed on the laser unless it is
complete. Thus, it is necessary to stop QMSPL with TELL
operation QUIT if you want to see the partial output
file.
Adverbs:
INNAME......Image name(name). blank => any
INCLASS.....Image name(class). blank => any
INSEQ.......Image name(seq. #). 0 => any
INDISK......Disk drive # of image. 0 => any
PLVER.......Version # of the PL file. 0 => largest version #
ASPMM.......Arc seconds per millimeter. 0 => self scale.
COPIES......Number of copies to make. 0 => 1.
LPEN........Width of lines in dots. 0 => 3 (less than 3 is
not recommended for copying).
OUTFILE.....If not blank, write the output print file to the
specified file name and print it. If blank, print
and delete the output.
OPTYPE......Determines the method by which grey scale images
are converted to dots on the paper: 'NONE' means
to suppress the grey scales. 'MBIT', 'QRAN',
and 'MRAN' use "fonts" loaded into the QMS, which
is fast but may have bad repetitive dot patterns.
QRAN uses a random number generator to generate 26
levels and a random number to choose which of 3
equal versions of that level are used for each
pixel. 'MRAN' uses 78 levels instead but uses a
fixed "character" for each level. 'MBIT' is like
'MRAN' but instead of a random number font, it
uses one generated by the Starlink algorithm.
'RAND' uses a random number generator to set each
dot and 'BITS' uses a STARLINK method of comparing
the value with a mask pattern. The last two are
very slow.
FUNCTYPE....Convert the recorded grey levels using inverse
linear ('NE'), logarithmic ('LG'), inverse log
('NG'), power law ('PL'), inverse power law
('NL') or linear (anything else). Note that
linear produces "negative" images --- ones in
which the most intense sources are the blackest.
FACTOR......Specifies the power law index used in converting
the recorded grey levels when FUNCTYPE 'PL' or
'NL' is specified. < 0.001 or > 999. => 1.0
AXVAL.......Specifies the input and output values of the
inflection point in the transfer function between
a 1./FACTOR and FACTOR power law. (0,0) => simple
FACTOR power law. If either value is outside the
range 0.001 - 0.999, a simple power law is used.
DPARM.......(1,2) The recorded grey values (0 to 1.0) may be
clipped and linearly rescaled before applying
FUNCTYPE, where 0 <= DPARM(1) < DPARM(2) <= 1.
Default is DPARM(1) = 0.0, DPARM(2) = 1.0.
(3,4) After FUNCTYPE has been applied, the values
range from 0 to 1. These may be shifted,
scaled, and reclipped to 0,1 by g*DPARM(3) +
DPARM(4). Defaults are 1.0 and 0.0 from limits
0.01 < DPARM(3) < 100, -0.9 < DPARM(4) < 0.9.
When using 'NE' or 'NG' functions, it is often
good to limit the darkest areas if they are
large to grey (i.e. DPARM(3) around .85).
Parameters which control the mask generation for .BIT. OPTYPEs
DPARM(5): = 0 -> use Gaussian with defaults
= 1 -> use Gaussian with DPARM parms
= 2 -> use Sine with defaults
= 3 -> use Sine with DPARM parameters
< 0 or > 3 => 0.
DPARM(6): mask size in dots, default = 8.
(< 0 for 'MDOT' => force to this, else
MDOT uses between 7 and 13 to come close
to the desired ASPMM.) Let mx, my =
mask size in x, y (in dots). Then
DPARM(7): G = e ** (-D(2) * x * x/ ((mx/4)**2))
S = sin (2PI * D(2) * x / mx)
default = 1.0
DPARM(8): X-shift of function wrt dots in dots
default = 0.25 * (mx/8.0), limit < 0.9.
DPARM(9): Y-shift of function wrt dots in dots
default = 0.133 * (my/8.0), limit < 0.9.
EXPLAIN SECTION
Explanation of QMSPL
Notes on scaling grey levels:
(1) GREYS puts the user-specified image intensities into the
plot file in a form which is converted to the range 0.0
to 1.0 by QMSPL.
(2) Then QMSPL clips these intensities, setting any < DPARM(1)
to DPARM(1) and any > DPARM(2) to DPARM(2) where 0 <=
DPARM(1) < DPARM(2) <= 1.0. The range DPARM(1) - DPARM(2)
is then rescaled to 0 - 1.
(3) Then QMSPL applies a transfer function which can be linear
(do nothing), negative linear (v = 1-v), logarithmic
(v = log10(30*v+1)/log10(30.42) - .007), negative log
(v=1-v, then log), power law (v = v ** FACTOR) or negative
power law (v=1-v, then power law). A two part power law
with inflection point is obtained with nonzero values of
AXVAL yielding v = AXVAL(2)*(v/AXVAL(1))**(1/FACTOR)
for v < AXVAL(1), and v = AXVAL(2)+(1-AXVAL(2))*
((v-AXVAL(1))/(1-AXVAL(1)))**FACTOR for v > or = AXVAL(1).
These functions convert the range 0 - 1 into the range 0 - 1
(4) Then QMSPL rescales the image data as v = DPARM(3) * v +
DPARM(4) and clips it at 0 and 1.
The results of all this are converted into a grey level on the
paper.