• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1Advanced usage instructions for the Independent JPEG Group's JPEG software
2==========================================================================
3
4This file describes cjpeg's "switches for wizards".
5
6The "wizard" switches are intended for experimentation with JPEG by persons
7who are reasonably knowledgeable about the JPEG standard.  If you don't know
8what you are doing, DON'T USE THESE SWITCHES.  You'll likely produce files
9with worse image quality and/or poorer compression than you'd get from the
10default settings.  Furthermore, these switches must be used with caution
11when making files intended for general use, because not all JPEG decoders
12will support unusual JPEG parameter settings.
13
14
15Quantization Table Adjustment
16-----------------------------
17
18Ordinarily, cjpeg starts with a default set of tables (the same ones given
19as examples in the JPEG standard) and scales them up or down according to
20the -quality setting.  The details of the scaling algorithm can be found in
21jcparam.c.  At very low quality settings, some quantization table entries
22can get scaled up to values exceeding 255.  Although 2-byte quantization
23values are supported by the IJG software, this feature is not in baseline
24JPEG and is not supported by all implementations.  If you need to ensure
25wide compatibility of low-quality files, you can constrain the scaled
26quantization values to no more than 255 by giving the -baseline switch.
27Note that use of -baseline will result in poorer quality for the same file
28size, since more bits than necessary are expended on higher AC coefficients.
29
30You can substitute a different set of quantization values by using the
31-qtables switch:
32
33        -qtables file   Use the quantization tables given in the named file.
34
35The specified file should be a text file containing decimal quantization
36values.  The file should contain one to four tables, each of 64 elements.
37The tables are implicitly numbered 0,1,etc. in order of appearance.  Table
38entries appear in normal array order (NOT in the zigzag order in which they
39will be stored in the JPEG file).
40
41Quantization table files are free format, in that arbitrary whitespace can
42appear between numbers.  Also, comments can be included: a comment starts
43with '#' and extends to the end of the line.  Here is an example file that
44duplicates the default quantization tables:
45
46        # Quantization tables given in Annex K (Clause K.1) of
47        # Recommendation ITU-T T.81 (1992) | ISO/IEC 10918-1:1994.
48
49        # This is table 0 (the luminance table):
50          16  11  10  16  24  40  51  61
51          12  12  14  19  26  58  60  55
52          14  13  16  24  40  57  69  56
53          14  17  22  29  51  87  80  62
54          18  22  37  56  68 109 103  77
55          24  35  55  64  81 104 113  92
56          49  64  78  87 103 121 120 101
57          72  92  95  98 112 100 103  99
58
59        # This is table 1 (the chrominance table):
60          17  18  24  47  99  99  99  99
61          18  21  26  66  99  99  99  99
62          24  26  56  99  99  99  99  99
63          47  66  99  99  99  99  99  99
64          99  99  99  99  99  99  99  99
65          99  99  99  99  99  99  99  99
66          99  99  99  99  99  99  99  99
67          99  99  99  99  99  99  99  99
68
69If the -qtables switch is used without -quality, then the specified tables
70are used exactly as-is.  If both -qtables and -quality are used, then the
71tables taken from the file are scaled in the same fashion that the default
72tables would be scaled for that quality setting.  If -baseline appears, then
73the quantization values are constrained to the range 1-255.
74
75By default, cjpeg will use quantization table 0 for luminance components and
76table 1 for chrominance components.  To override this choice, use the -qslots
77switch:
78
79        -qslots N[,...]         Select which quantization table to use for
80                                each color component.
81
82The -qslots switch specifies a quantization table number for each color
83component, in the order in which the components appear in the JPEG SOF marker.
84For example, to create a separate table for each of Y,Cb,Cr, you could
85provide a -qtables file that defines three quantization tables and say
86"-qslots 0,1,2".  If -qslots gives fewer table numbers than there are color
87components, then the last table number is repeated as necessary.
88
89
90Sampling Factor Adjustment
91--------------------------
92
93By default, cjpeg uses 2:1 horizontal and vertical downsampling when
94compressing YCbCr data, and no downsampling for all other color spaces.
95You can override this default with the -sample switch:
96
97        -sample HxV[,...]       Set JPEG sampling factors for each color
98                                component.
99
100The -sample switch specifies the JPEG sampling factors for each color
101component, in the order in which they appear in the JPEG SOF marker.
102If you specify fewer HxV pairs than there are components, the remaining
103components are set to 1x1 sampling.  For example, the default YCbCr setting
104is equivalent to "-sample 2x2,1x1,1x1", which can be abbreviated to
105"-sample 2x2".
106
107There are still some JPEG decoders in existence that support only 2x1
108sampling (also called 4:2:2 sampling).  Compatibility with such decoders can
109be achieved by specifying "-sample 2x1".  This is not recommended unless
110really necessary, since it increases file size and encoding/decoding time
111with very little quality gain.
112
113
114Multiple Scan / Progression Control
115-----------------------------------
116
117By default, cjpeg emits a single-scan sequential JPEG file.  The
118-progressive switch generates a progressive JPEG file using a default series
119of progression parameters.  You can create multiple-scan sequential JPEG
120files or progressive JPEG files with custom progression parameters by using
121the -scans switch:
122
123        -scans file     Use the scan sequence given in the named file.
124
125The specified file should be a text file containing a "scan script".
126The script specifies the contents and ordering of the scans to be emitted.
127Each entry in the script defines one scan.  A scan definition specifies
128the components to be included in the scan, and for progressive JPEG it also
129specifies the progression parameters Ss,Se,Ah,Al for the scan.  Scan
130definitions are separated by semicolons (';').  A semicolon after the last
131scan definition is optional.
132
133Each scan definition contains one to four component indexes, optionally
134followed by a colon (':') and the four progressive-JPEG parameters.  The
135component indexes denote which color component(s) are to be transmitted in
136the scan.  Components are numbered in the order in which they appear in the
137JPEG SOF marker, with the first component being numbered 0.  (Note that these
138indexes are not the "component ID" codes assigned to the components, just
139positional indexes.)
140
141The progression parameters for each scan are:
142        Ss      Zigzag index of first coefficient included in scan
143        Se      Zigzag index of last coefficient included in scan
144        Ah      Zero for first scan of a coefficient, else Al of prior scan
145        Al      Successive approximation low bit position for scan
146If the progression parameters are omitted, the values 0,63,0,0 are used,
147producing a sequential JPEG file.  cjpeg automatically determines whether
148the script represents a progressive or sequential file, by observing whether
149Ss and Se values other than 0 and 63 appear.  (The -progressive switch is
150not needed to specify this; in fact, it is ignored when -scans appears.)
151The scan script must meet the JPEG restrictions on progression sequences.
152(cjpeg checks that the spec's requirements are obeyed.)
153
154Scan script files are free format, in that arbitrary whitespace can appear
155between numbers and around punctuation.  Also, comments can be included: a
156comment starts with '#' and extends to the end of the line.  For additional
157legibility, commas or dashes can be placed between values.  (Actually, any
158single punctuation character other than ':' or ';' can be inserted.)  For
159example, the following two scan definitions are equivalent:
160        0 1 2: 0 63 0 0;
161        0,1,2 : 0-63, 0,0 ;
162
163Here is an example of a scan script that generates a partially interleaved
164sequential JPEG file:
165
166        0;                      # Y only in first scan
167        1 2;                    # Cb and Cr in second scan
168
169Here is an example of a progressive scan script using only spectral selection
170(no successive approximation):
171
172        # Interleaved DC scan for Y,Cb,Cr:
173        0,1,2: 0-0,   0, 0 ;
174        # AC scans:
175        0:     1-2,   0, 0 ;    # First two Y AC coefficients
176        0:     3-5,   0, 0 ;    # Three more
177        1:     1-63,  0, 0 ;    # All AC coefficients for Cb
178        2:     1-63,  0, 0 ;    # All AC coefficients for Cr
179        0:     6-9,   0, 0 ;    # More Y coefficients
180        0:     10-63, 0, 0 ;    # Remaining Y coefficients
181
182Here is an example of a successive-approximation script.  This is equivalent
183to the default script used by "cjpeg -progressive" for YCbCr images:
184
185        # Initial DC scan for Y,Cb,Cr (lowest bit not sent)
186        0,1,2: 0-0,   0, 1 ;
187        # First AC scan: send first 5 Y AC coefficients, minus 2 lowest bits:
188        0:     1-5,   0, 2 ;
189        # Send all Cr,Cb AC coefficients, minus lowest bit:
190        # (chroma data is usually too small to be worth subdividing further;
191        #  but note we send Cr first since eye is least sensitive to Cb)
192        2:     1-63,  0, 1 ;
193        1:     1-63,  0, 1 ;
194        # Send remaining Y AC coefficients, minus 2 lowest bits:
195        0:     6-63,  0, 2 ;
196        # Send next-to-lowest bit of all Y AC coefficients:
197        0:     1-63,  2, 1 ;
198        # At this point we've sent all but the lowest bit of all coefficients.
199        # Send lowest bit of DC coefficients
200        0,1,2: 0-0,   1, 0 ;
201        # Send lowest bit of AC coefficients
202        2:     1-63,  1, 0 ;
203        1:     1-63,  1, 0 ;
204        # Y AC lowest bit scan is last; it's usually the largest scan
205        0:     1-63,  1, 0 ;
206
207It may be worth pointing out that this script is tuned for quality settings
208of around 50 to 75.  For lower quality settings, you'd probably want to use
209a script with fewer stages of successive approximation (otherwise the
210initial scans will be really bad).  For higher quality settings, you might
211want to use more stages of successive approximation (so that the initial
212scans are not too large).
213