MIZZI image manipulation program
================================

Syntax
------

mimp [genaral options] [filter options] <files>.[jpg|ppm]

Example
-------

mimp -oprefix ../mypics/filtered/ -cnr 2 -nr 2 -s 1 -ec 1 101_*.jpg 

filter all immages 101_*.jpg with 
color noise reduction, normal noise reduction, smoth it, 
and apply edge contrast. Finally store all images in 
../mypics/filtered/ 

general options
---------------

--           end of option processing
             specified this option if some files start with a - sign

-oprefix x   set output name prefix to x, default empty string 

-opostfix x  set output name postfix to x, default is a string of 
             the filter parameters

-o x         set output name to x (suffix must be supplied), 
             default empty string, if more than one input file is 
	     given a simple %d will insert an inremented number 
	     starting from 1 for every file, e.g.:
	     mimp -o filtered-%d.jpg -nrl file1.jpg file2.jpg
	     produces filtered-1.jpg filtered-2.jpg as outputname
             if -o is specified -oprefix -opostfix does not have any effect 

-jq n        set jpeg output quality to n 
-jqs         set jpeg output quality to standard 75
-jql         set jpeg output quality to low 90
-jqm         set jpeg output quality to middle 95
-jqh         set jpeg output quality to high 99 (default)

-f           if a file with the same name exist, overwrite it. 


filter options
--------------
optional parameters a placed in [], defaults in ()
all filters will show an output line showing the used paramters.


geometry filters:
-----------------
optional paramters are replaced by the first paramter.

-sc xscale[,yscale]

  Scale image 
  xscale[,yscale] =  0 ... 10 : scale image with scale factor xscale,yscale
  xscale,yscale < 0 : scale image to pixel size -xscale -yscale
  e.g. -sc -1280,-1024 : scale image fit 1280x1024 resolution
 
-cr x1,y1[,x2,y2]

  Crop image 
  x1,y1[,x2,y2] = 0 ... size-of-image : upper x,y point, lower x,y point
  x1,y1 < 0 crop to the ratio of x1,y1
  e.g. -cr -5,-4 -sc -1280,-1024 : crop the image to the ratio 5:4 and scale
  it to fit 1280x1024 resolution. This would be the ideal filter for 
  a background picture. 
  

color filters: 
--------------
optional paramters are replaced by the first paramter.

-cg r[,g,b]
 
  Color gain 
  r,g,b = 0 ... 100 : r,g,b pixel values are multiplied with r,g,b
  this is the normal way to correct under exposed pictures. 
  
-co r[,g,b]

  Color offset shift 
  r,g,b = -1 .. 1 : <0 shift down >0 shift up 
  the r,g,b parameters are subtracted from the all r,g,b values.   
  maximum is 1 (100%) and corresponds to 255 of the pixel r,g,b value.
  e.g.: -co -0.05 -cg 1.05: cut of the lower 5% of the black levels 
  and gain all colors that the white point is corrected to the old white point.
    
-g r[,g,b]

  Gamma for r[,g,b] = 0.1 ... 10 : <1 darker >1 lighter
  gamma correction does not change the black and white color point.
  
-cs r[,g,b]  

  Modify saturation 
  r,g,b = 0 ... 10 :  <1 less sat. >1 more sat. 
  r,g,b=0  make it black and white. 
  
-cc r[,g,b]  

  Decrease or increase contrast of the image.
  r,g,b = 0 ... 10 : <1 lower >1 higher contrast. 
  the contrast filter does not change the black and white and the grey50 color point.
  but if contrast is > 1 pixels between black and grey50 will get darker and pixels 
  between grey50 and white will get brighter. if contrast is < 1 this is done in reverse 
  order.  
  

smoothing and noisereduction filters:
-------------------------------------
optional paramters are replaced by defaults.

-s n[,r,w]   

  Smooth image n (1) times 
  with a blur radius r (1)  
  and the strength w (1)  
  
-nr n[,r,c,s]

  Noise reduction - color and and luminance
  n = 1 ... 10 : number of cycles (default=1). 
  r = 1 ... 100 : radius of noise reduction (default=10)
  c = 0 ... 2 : contour correction gain. 
 
      This parameter corrects blurred bright regions, if the noise 
      reduction has destroyed structures.
      0 is off, 2 high correction, (default is 1), you can set this 
      parameter to 0 for n < 2 or small radius values r < 3 to 
      reduce computing time. at higher noise reduction this paramter
      should be 0.5-1 but note that higher values will destroy moderately 
      the noise reduction. 

  s = 0.01 ... 5 : strength parameter

      for internal use (default is 1), dont change it
   
-cnr n[,r]

  Color noise reduction - remove color noise (crominanz) only. 
  n = 1 ... 10 : number of cycles (default=1). 
  r = 1 ... 100 : radius of noise reduction (default=20)
  e.g.: -cnr 3,30 -nr 3,5 : typical parameters for high ISO values. 
  NOTE: color noise reduction increases B/W noise and should be applied 
  before standard noise reduction. cnr should be used if you need a bigger 
  color noise radius.  

-snr cr,lr,cw,lw

  Simple noise reduction. This filter is much faster than cnr-nr cycles 
  but has nearly the same or better results for low noise images.  
  cr = chrominance radius (default=10). 
  lr = luminance radius (default=cr/4+1)
  cw = chrominance noise reduction strenght (default=1)
  lw = luminance noise reduction strenght (default=3)
  try this filter before any other. It is really perfect (my favorite).
 
-cd n  

  Color dithering
  if your display program is smoothing to much and displays colors show stairs
  you can dither the colors with n (default 1). this randomly add n to one
  color and subtract n from one other. 

-ec g[,r,w]  

  Edge contrast 
  g = 0 ... 10 : the gain of the edge contrast (default 1)
  r = 1 ... 100 : radius of the edge (default 1) s
  w = 0.1 ... 10 : slope of the edge (default 1), dont change it. 
  
-bfc s[,lim,r,inc]
       
  Blue frings correction
  s = strength 1 ... 3 (1) 
  lim = tag limit 0.5 ... 0.99 (0.95)
  r = correction radius 1 .. 50 (dynamic) 
  inc = precision of filter higher values will decrese filter time  

-rfc s[,lim,r,inc]

  Red frings correction
  parameters same as -bfc


Predefined filters:
-------------------
predefined filters combine some filters for simpler usage

-nrl 

  Low noise reduction, should be applied to every image. will improve really 
  everthing, is fast, and removes really no structures. Also apply this filter
  to high quality images which has not visiable noise (iso 50 - iso 100)

-nrm   

  Medium noise filter, this filter is removing almost visable noise, but if you 
  check the images will have some lower contrast in dark structures.
  
-nrh 

  Medium noise reduction, good for higher iso values. 
  
-nru 

  Ultra noise reduction, good for images you must throw away otherwise!
  Also has some nice effect at portraits, try it.
  

Examples
--------

noise reduction only - two simple noise reduction cycles - set postfix and prefix. 

  mimp -oprefix ../filtered/ -opostfix filtered -snr 10,5,1,2 -snr 5,2,1,2  p110*.jpg

enhance contrast and make a simple noise reduction.

  mimp -cc 1.2 -snr 10 p110*.jpg

correct the dark point to -5%,  gain the colors by 5%,  enhance the contrast by 1.2  
and aply noise reduction with raduis 10.

  mimp -co -0.05 -cg 1.05 -cc 1.2 -snr 10 p110*.jpg