Documentation/DocBook/v4l/pixfmt.xml

   1   <title>Image Formats</title>
   2
   3   <para>The V4L2 API was primarily designed for devices exchanging
   4 image data with applications. The
   5 <structname>v4l2_pix_format</structname> and <structname>v4l2_pix_format_mplane
   6 </structname> structures define the format and layout of an image in memory.
   7 The former is used with the single-planar API, while the latter is used with the
   8 multi-planar version (see <xref linkend="planar-apis"/>). Image formats are
   9 negotiated with the &VIDIOC-S-FMT; ioctl. (The explanations here focus on video
  10 capturing and output, for overlay frame buffer formats see also
  11 &VIDIOC-G-FBUF;.)</para>
  12
  13 <section>
  14   <title>Single-planar format structure</title>
  15   <table pgwide="1" frame="none" id="v4l2-pix-format">
  16     <title>struct <structname>v4l2_pix_format</structname></title>
  17     <tgroup cols="3">
  18       &cs-str;
  19       <tbody valign="top">
  20         <row>
  21           <entry>__u32</entry>
  22           <entry><structfield>width</structfield></entry>
  23           <entry>Image width in pixels.</entry>
  24         </row>
  25         <row>
  26           <entry>__u32</entry>
  27           <entry><structfield>height</structfield></entry>
  28           <entry>Image height in pixels.</entry>
  29         </row>
  30         <row>
  31           <entry spanname="hspan">Applications set these fields to
  32 request an image size, drivers return the closest possible values. In
  33 case of planar formats the <structfield>width</structfield> and
  34 <structfield>height</structfield> applies to the largest plane. To
  35 avoid ambiguities drivers must return values rounded up to a multiple
  36 of the scale factor of any smaller planes. For example when the image
  37 format is YUV 4:2:0, <structfield>width</structfield> and
  38 <structfield>height</structfield> must be multiples of two.</entry>
  39         </row>
  40         <row>
  41           <entry>__u32</entry>
  42           <entry><structfield>pixelformat</structfield></entry>
  43           <entry>The pixel format or type of compression, set by the
  44 application. This is a little endian <link
  45 linkend="v4l2-fourcc">four character code</link>. V4L2 defines
  46 standard RGB formats in <xref linkend="rgb-formats" />, YUV formats in <xref
  47 linkend="yuv-formats" />, and reserved codes in <xref
  48 linkend="reserved-formats" /></entry>
  49         </row>
  50         <row>
  51           <entry>&v4l2-field;</entry>
  52           <entry><structfield>field</structfield></entry>
  53           <entry>Video images are typically interlaced. Applications
  54 can request to capture or output only the top or bottom field, or both
  55 fields interlaced or sequentially stored in one buffer or alternating
  56 in separate buffers. Drivers return the actual field order selected.
  57 For details see <xref linkend="field-order" />.</entry>
  58         </row>
  59         <row>
  60           <entry>__u32</entry>
  61           <entry><structfield>bytesperline</structfield></entry>
  62           <entry>Distance in bytes between the leftmost pixels in two
  63 adjacent lines.</entry>
  64         </row>
  65         <row>
  66           <entry spanname="hspan"><para>Both applications and drivers
  67 can set this field to request padding bytes at the end of each line.
  68 Drivers however may ignore the value requested by the application,
  69 returning <structfield>width</structfield> times bytes per pixel or a
  70 larger value required by the hardware. That implies applications can
  71 just set this field to zero to get a reasonable
  72 default.</para><para>Video hardware may access padding bytes,
  73 therefore they must reside in accessible memory. Consider cases where
  74 padding bytes after the last line of an image cross a system page
  75 boundary. Input devices may write padding bytes, the value is
  76 undefined. Output devices ignore the contents of padding
  77 bytes.</para><para>When the image format is planar the
  78 <structfield>bytesperline</structfield> value applies to the largest
  79 plane and is divided by the same factor as the
  80 <structfield>width</structfield> field for any smaller planes. For
  81 example the Cb and Cr planes of a YUV 4:2:0 image have half as many
  82 padding bytes following each line as the Y plane. To avoid ambiguities
  83 drivers must return a <structfield>bytesperline</structfield> value
  84 rounded up to a multiple of the scale factor.</para></entry>
  85         </row>
  86         <row>
  87           <entry>__u32</entry>
  88           <entry><structfield>sizeimage</structfield></entry>
  89           <entry>Size in bytes of the buffer to hold a complete image,
  90 set by the driver. Usually this is
  91 <structfield>bytesperline</structfield> times
  92 <structfield>height</structfield>. When the image consists of variable
  93 length compressed data this is the maximum number of bytes required to
  94 hold an image.</entry>
  95         </row>
  96         <row>
  97           <entry>&v4l2-colorspace;</entry>
  98           <entry><structfield>colorspace</structfield></entry>
  99           <entry>This information supplements the
 100 <structfield>pixelformat</structfield> and must be set by the driver,
 101 see <xref linkend="colorspaces" />.</entry>
 102         </row>
 103         <row>
 104           <entry>__u32</entry>
 105           <entry><structfield>priv</structfield></entry>
 106           <entry>Reserved for custom (driver defined) additional
 107 information about formats. When not used drivers and applications must
 108 set this field to zero.</entry>
 109         </row>
 110       </tbody>
 111     </tgroup>
 112   </table>
 113 </section>
 114
 115 <section>
 116   <title>Multi-planar format structures</title>
 117   <para>The <structname>v4l2_plane_pix_format</structname> structures define
 118     size and layout for each of the planes in a multi-planar format.
 119     The <structname>v4l2_pix_format_mplane</structname> structure contains
 120     information common to all planes (such as image width and height) and
 121     an array of <structname>v4l2_plane_pix_format</structname> structures,
 122     describing all planes of that format.</para>
 123   <table pgwide="1" frame="none" id="v4l2-plane-pix-format">
 124     <title>struct <structname>vl42_plane_pix_format</structname></title>
 125     <tgroup cols="3">
 126       &cs-str;
 127       <tbody valign="top">
 128         <row>
 129           <entry>__u32</entry>
 130           <entry><structfield>sizeimage</structfield></entry>
 131           <entry>Maximum size in bytes required for image data in this plane.
 132           </entry>
 133         </row>
 134         <row>
 135           <entry>__u16</entry>
 136           <entry><structfield>bytesperline</structfield></entry>
 137           <entry>Distance in bytes between the leftmost pixels in two adjacent
 138             lines.</entry>
 139         </row>
 140         <row>
 141           <entry>__u16</entry>
 142           <entry><structfield>reserved[7]</structfield></entry>
 143           <entry>Reserved for future extensions. Should be zeroed by the
 144            application.</entry>
 145         </row>
 146       </tbody>
 147     </tgroup>
 148   </table>
 149   <table pgwide="1" frame="none" id="v4l2-pix-format-mplane">
 150     <title>struct <structname>v4l2_pix_format_mplane</structname></title>
 151     <tgroup cols="3">
 152       &cs-str;
 153       <tbody valign="top">
 154         <row>
 155           <entry>__u32</entry>
 156           <entry><structfield>width</structfield></entry>
 157           <entry>Image width in pixels.</entry>
 158         </row>
 159         <row>
 160           <entry>__u32</entry>
 161           <entry><structfield>height</structfield></entry>
 162           <entry>Image height in pixels.</entry>
 163         </row>
 164         <row>
 165           <entry>__u32</entry>
 166           <entry><structfield>pixelformat</structfield></entry>
 167           <entry>The pixel format. Both single- and multi-planar four character
 168 codes can be used.</entry>
 169         </row>
 170         <row>
 171           <entry>&v4l2-field;</entry>
 172           <entry><structfield>field</structfield></entry>
 173           <entry>See &v4l2-pix-format;.</entry>
 174         </row>
 175         <row>
 176           <entry>&v4l2-colorspace;</entry>
 177           <entry><structfield>colorspace</structfield></entry>
 178           <entry>See &v4l2-pix-format;.</entry>
 179         </row>
 180         <row>
 181           <entry>&v4l2-plane-pix-format;</entry>
 182           <entry><structfield>plane_fmt[VIDEO_MAX_PLANES]</structfield></entry>
 183           <entry>An array of structures describing format of each plane this
 184           pixel format consists of. The number of valid entries in this array
 185           has to be put in the <structfield>num_planes</structfield>
 186           field.</entry>
 187         </row>
 188         <row>
 189           <entry>__u8</entry>
 190           <entry><structfield>num_planes</structfield></entry>
 191           <entry>Number of planes (i.e. separate memory buffers) for this format
 192           and the number of valid entries in the
 193           <structfield>plane_fmt</structfield> array.</entry>
 194         </row>
 195         <row>
 196           <entry>__u8</entry>
 197           <entry><structfield>reserved[11]</structfield></entry>
 198           <entry>Reserved for future extensions. Should be zeroed by the
 199            application.</entry>
 200         </row>
 201       </tbody>
 202     </tgroup>
 203   </table>
 204 </section>
 205
 206   <section>
 207     <title>Standard Image Formats</title>
 208
 209     <para>In order to exchange images between drivers and
 210 applications, it is necessary to have standard image data formats
 211 which both sides will interpret the same way. V4L2 includes several
 212 such formats, and this section is intended to be an unambiguous
 213 specification of the standard image data formats in V4L2.</para>
 214
 215     <para>V4L2 drivers are not limited to these formats, however.
 216 Driver-specific formats are possible. In that case the application may
 217 depend on a codec to convert images to one of the standard formats
 218 when needed. But the data can still be stored and retrieved in the
 219 proprietary format. For example, a device may support a proprietary
 220 compressed format. Applications can still capture and save the data in
 221 the compressed format, saving much disk space, and later use a codec
 222 to convert the images to the X Windows screen format when the video is
 223 to be displayed.</para>
 224
 225     <para>Even so, ultimately, some standard formats are needed, so
 226 the V4L2 specification would not be complete without well-defined
 227 standard formats.</para>
 228
 229     <para>The V4L2 standard formats are mainly uncompressed formats. The
 230 pixels are always arranged in memory from left to right, and from top
 231 to bottom. The first byte of data in the image buffer is always for
 232 the leftmost pixel of the topmost row. Following that is the pixel
 233 immediately to its right, and so on until the end of the top row of
 234 pixels. Following the rightmost pixel of the row there may be zero or
 235 more bytes of padding to guarantee that each row of pixel data has a
 236 certain alignment. Following the pad bytes, if any, is data for the
 237 leftmost pixel of the second row from the top, and so on. The last row
 238 has just as many pad bytes after it as the other rows.</para>
 239
 240     <para>In V4L2 each format has an identifier which looks like
 241 <constant>PIX_FMT_XXX</constant>, defined in the <link
 242 linkend="videodev">videodev.h</link> header file. These identifiers
 243 represent <link linkend="v4l2-fourcc">four character (FourCC) codes</link>
 244 which are also listed below, however they are not the same as those
 245 used in the Windows world.</para>
 246
 247     <para>For some formats, data is stored in separate, discontiguous
 248 memory buffers. Those formats are identified by a separate set of FourCC codes
 249 and are referred to as "multi-planar formats". For example, a YUV422 frame is
 250 normally stored in one memory buffer, but it can also be placed in two or three
 251 separate buffers, with Y component in one buffer and CbCr components in another
 252 in the 2-planar version or with each component in its own buffer in the
 253 3-planar case. Those sub-buffers are referred to as "planes".</para>
 254   </section>
 255
 256   <section id="colorspaces">
 257     <title>Colorspaces</title>
 258
 259     <para>[intro]</para>
 260
 261     <!-- See proposal by Billy Biggs, video4linux-list@redhat.com
 262 on 11 Oct 2002, subject: "Re: [V4L] Re: v4l2 api", and
 263 http://vektor.theorem.ca/graphics/ycbcr/ and
 264 http://www.poynton.com/notes/colour_and_gamma/ColorFAQ.html -->
 265
 266     <para>
 267       <variablelist>
 268         <varlistentry>
 269           <term>Gamma Correction</term>
 270           <listitem>
 271             <para>[to do]</para>
 272             <para>E'<subscript>R</subscript> = f(R)</para>
 273             <para>E'<subscript>G</subscript> = f(G)</para>
 274             <para>E'<subscript>B</subscript> = f(B)</para>
 275           </listitem>
 276         </varlistentry>
 277         <varlistentry>
 278           <term>Construction of luminance and color-difference
 279 signals</term>
 280           <listitem>
 281             <para>[to do]</para>
 282             <para>E'<subscript>Y</subscript> =
 283 Coeff<subscript>R</subscript> E'<subscript>R</subscript>
 284 + Coeff<subscript>G</subscript> E'<subscript>G</subscript>
 285 + Coeff<subscript>B</subscript> E'<subscript>B</subscript></para>
 286             <para>(E'<subscript>R</subscript> - E'<subscript>Y</subscript>) = E'<subscript>R</subscript>
 287 - Coeff<subscript>R</subscript> E'<subscript>R</subscript>
 288 - Coeff<subscript>G</subscript> E'<subscript>G</subscript>
 289 - Coeff<subscript>B</subscript> E'<subscript>B</subscript></para>
 290             <para>(E'<subscript>B</subscript> - E'<subscript>Y</subscript>) = E'<subscript>B</subscript>
 291 - Coeff<subscript>R</subscript> E'<subscript>R</subscript>
 292 - Coeff<subscript>G</subscript> E'<subscript>G</subscript>
 293 - Coeff<subscript>B</subscript> E'<subscript>B</subscript></para>
 294           </listitem>
 295         </varlistentry>
 296         <varlistentry>
 297           <term>Re-normalized color-difference signals</term>
 298           <listitem>
 299             <para>The color-difference signals are scaled back to unity
 300 range [-0.5;+0.5]:</para>
 301             <para>K<subscript>B</subscript> = 0.5 / (1 - Coeff<subscript>B</subscript>)</para>
 302             <para>K<subscript>R</subscript> = 0.5 / (1 - Coeff<subscript>R</subscript>)</para>
 303             <para>P<subscript>B</subscript> =
 304 K<subscript>B</subscript> (E'<subscript>B</subscript> - E'<subscript>Y</subscript>) =
 305   0.5 (Coeff<subscript>R</subscript> / Coeff<subscript>B</subscript>) E'<subscript>R</subscript>
 306 + 0.5 (Coeff<subscript>G</subscript> / Coeff<subscript>B</subscript>) E'<subscript>G</subscript>
 307 + 0.5 E'<subscript>B</subscript></para>
 308             <para>P<subscript>R</subscript> =
 309 K<subscript>R</subscript> (E'<subscript>R</subscript> - E'<subscript>Y</subscript>) =
 310   0.5 E'<subscript>R</subscript>
 311 + 0.5 (Coeff<subscript>G</subscript> / Coeff<subscript>R</subscript>) E'<subscript>G</subscript>
 312 + 0.5 (Coeff<subscript>B</subscript> / Coeff<subscript>R</subscript>) E'<subscript>B</subscript></para>
 313           </listitem>
 314         </varlistentry>
 315         <varlistentry>
 316           <term>Quantization</term>
 317           <listitem>
 318             <para>[to do]</para>
 319             <para>Y' = (Lum. Levels - 1) &middot; E'<subscript>Y</subscript> + Lum. Offset</para>
 320             <para>C<subscript>B</subscript> = (Chrom. Levels - 1)
 321 &middot; P<subscript>B</subscript> + Chrom. Offset</para>
 322             <para>C<subscript>R</subscript> = (Chrom. Levels - 1)
 323 &middot; P<subscript>R</subscript> + Chrom. Offset</para>
 324             <para>Rounding to the nearest integer and clamping to the range
 325 [0;255] finally yields the digital color components Y'CbCr
 326 stored in YUV images.</para>
 327           </listitem>
 328         </varlistentry>
 329       </variablelist>
 330     </para>
 331
 332     <example>
 333       <title>ITU-R Rec. BT.601 color conversion</title>
 334
 335       <para>Forward Transformation</para>
 336
 337       <programlisting>
 338 int ER, EG, EB;         /* gamma corrected RGB input [0;255] */
 339 int Y1, Cb, Cr;         /* output [0;255] */
 340
 341 double r, g, b;         /* temporaries */
 342 double y1, pb, pr;
 343
 344 int
 345 clamp (double x)
 346 {
 347         int r = x;      /* round to nearest */
 348
 349         if (r &lt; 0)         return 0;
 350         else if (r &gt; 255)  return 255;
 351         else               return r;
 352 }
 353
 354 r = ER / 255.0;
 355 g = EG / 255.0;
 356 b = EB / 255.0;
 357
 358 y1  =  0.299  * r + 0.587 * g + 0.114  * b;
 359 pb  = -0.169  * r - 0.331 * g + 0.5    * b;
 360 pr  =  0.5    * r - 0.419 * g - 0.081  * b;
 361
 362 Y1 = clamp (219 * y1 + 16);
 363 Cb = clamp (224 * pb + 128);
 364 Cr = clamp (224 * pr + 128);
 365
 366 /* or shorter */
 367
 368 y1 = 0.299 * ER + 0.587 * EG + 0.114 * EB;
 369
 370 Y1 = clamp ( (219 / 255.0)                    *       y1  + 16);
 371 Cb = clamp (((224 / 255.0) / (2 - 2 * 0.114)) * (EB - y1) + 128);
 372 Cr = clamp (((224 / 255.0) / (2 - 2 * 0.299)) * (ER - y1) + 128);
 373       </programlisting>
 374
 375       <para>Inverse Transformation</para>
 376
 377       <programlisting>
 378 int Y1, Cb, Cr;         /* gamma pre-corrected input [0;255] */
 379 int ER, EG, EB;         /* output [0;255] */
 380
 381 double r, g, b;         /* temporaries */
 382 double y1, pb, pr;
 383
 384 int
 385 clamp (double x)
 386 {
 387         int r = x;      /* round to nearest */
 388
 389         if (r &lt; 0)         return 0;
 390         else if (r &gt; 255)  return 255;
 391         else               return r;
 392 }
 393
 394 y1 = (255 / 219.0) * (Y1 - 16);
 395 pb = (255 / 224.0) * (Cb - 128);
 396 pr = (255 / 224.0) * (Cr - 128);
 397
 398 r = 1.0 * y1 + 0     * pb + 1.402 * pr;
 399 g = 1.0 * y1 - 0.344 * pb - 0.714 * pr;
 400 b = 1.0 * y1 + 1.772 * pb + 0     * pr;
 401
 402 ER = clamp (r * 255); /* [ok? one should prob. limit y1,pb,pr] */
 403 EG = clamp (g * 255);
 404 EB = clamp (b * 255);
 405       </programlisting>
 406     </example>
 407
 408     <table pgwide="1" id="v4l2-colorspace" orient="land">
 409       <title>enum v4l2_colorspace</title>
 410       <tgroup cols="11" align="center">
 411         <colspec align="left" />
 412         <colspec align="center" />
 413         <colspec align="left" />
 414         <colspec colname="cr" />
 415         <colspec colname="cg" />
 416         <colspec colname="cb" />
 417         <colspec colname="wp" />
 418         <colspec colname="gc" />
 419         <colspec colname="lum" />
 420         <colspec colname="qy" />
 421         <colspec colname="qc" />
 422         <spanspec namest="cr" nameend="cb" spanname="chrom" />
 423         <spanspec namest="qy" nameend="qc" spanname="quant" />
 424         <spanspec namest="lum" nameend="qc" spanname="spam" />
 425         <thead>
 426           <row>
 427             <entry morerows="1">Identifier</entry>
 428             <entry morerows="1">Value</entry>
 429             <entry morerows="1">Description</entry>
 430             <entry spanname="chrom">Chromaticities<footnote>
 431                 <para>The coordinates of the color primaries are
 432 given in the CIE system (1931)</para>
 433               </footnote></entry>
 434             <entry morerows="1">White Point</entry>
 435             <entry morerows="1">Gamma Correction</entry>
 436             <entry morerows="1">Luminance E'<subscript>Y</subscript></entry>
 437             <entry spanname="quant">Quantization</entry>
 438           </row>
 439           <row>
 440             <entry>Red</entry>
 441             <entry>Green</entry>
 442             <entry>Blue</entry>
 443             <entry>Y'</entry>
 444             <entry>Cb, Cr</entry>
 445           </row>
 446         </thead>
 447         <tbody valign="top">
 448           <row>
 449             <entry><constant>V4L2_COLORSPACE_SMPTE170M</constant></entry>
 450             <entry>1</entry>
 451             <entry>NTSC/PAL according to <xref linkend="smpte170m" />,
 452 <xref linkend="itu601" /></entry>
 453             <entry>x&nbsp;=&nbsp;0.630, y&nbsp;=&nbsp;0.340</entry>
 454             <entry>x&nbsp;=&nbsp;0.310, y&nbsp;=&nbsp;0.595</entry>
 455             <entry>x&nbsp;=&nbsp;0.155, y&nbsp;=&nbsp;0.070</entry>
 456             <entry>x&nbsp;=&nbsp;0.3127, y&nbsp;=&nbsp;0.3290,
 457             Illuminant D<subscript>65</subscript></entry>
 458             <entry>E' = 4.5&nbsp;I&nbsp;for&nbsp;I&nbsp;&le;0.018,
 459 1.099&nbsp;I<superscript>0.45</superscript>&nbsp;-&nbsp;0.099&nbsp;for&nbsp;0.018&nbsp;&lt;&nbsp;I</entry>
 460             <entry>0.299&nbsp;E'<subscript>R</subscript>
 461 +&nbsp;0.587&nbsp;E'<subscript>G</subscript>
 462 +&nbsp;0.114&nbsp;E'<subscript>B</subscript></entry>
 463             <entry>219&nbsp;E'<subscript>Y</subscript>&nbsp;+&nbsp;16</entry>
 464             <entry>224&nbsp;P<subscript>B,R</subscript>&nbsp;+&nbsp;128</entry>
 465           </row>
 466           <row>
 467             <entry><constant>V4L2_COLORSPACE_SMPTE240M</constant></entry>
 468             <entry>2</entry>
 469             <entry>1125-Line (US) HDTV, see <xref
 470 linkend="smpte240m" /></entry>
 471             <entry>x&nbsp;=&nbsp;0.630, y&nbsp;=&nbsp;0.340</entry>
 472             <entry>x&nbsp;=&nbsp;0.310, y&nbsp;=&nbsp;0.595</entry>
 473             <entry>x&nbsp;=&nbsp;0.155, y&nbsp;=&nbsp;0.070</entry>
 474             <entry>x&nbsp;=&nbsp;0.3127, y&nbsp;=&nbsp;0.3290,
 475             Illuminant D<subscript>65</subscript></entry>
 476             <entry>E' = 4&nbsp;I&nbsp;for&nbsp;I&nbsp;&le;0.0228,
 477 1.1115&nbsp;I<superscript>0.45</superscript>&nbsp;-&nbsp;0.1115&nbsp;for&nbsp;0.0228&nbsp;&lt;&nbsp;I</entry>
 478             <entry>0.212&nbsp;E'<subscript>R</subscript>
 479 +&nbsp;0.701&nbsp;E'<subscript>G</subscript>
 480 +&nbsp;0.087&nbsp;E'<subscript>B</subscript></entry>
 481             <entry>219&nbsp;E'<subscript>Y</subscript>&nbsp;+&nbsp;16</entry>
 482             <entry>224&nbsp;P<subscript>B,R</subscript>&nbsp;+&nbsp;128</entry>
 483           </row>
 484           <row>
 485             <entry><constant>V4L2_COLORSPACE_REC709</constant></entry>
 486             <entry>3</entry>
 487             <entry>HDTV and modern devices, see <xref
 488 linkend="itu709" /></entry>
 489             <entry>x&nbsp;=&nbsp;0.640, y&nbsp;=&nbsp;0.330</entry>
 490             <entry>x&nbsp;=&nbsp;0.300, y&nbsp;=&nbsp;0.600</entry>
 491             <entry>x&nbsp;=&nbsp;0.150, y&nbsp;=&nbsp;0.060</entry>
 492             <entry>x&nbsp;=&nbsp;0.3127, y&nbsp;=&nbsp;0.3290,
 493             Illuminant D<subscript>65</subscript></entry>
 494             <entry>E' = 4.5&nbsp;I&nbsp;for&nbsp;I&nbsp;&le;0.018,
 495 1.099&nbsp;I<superscript>0.45</superscript>&nbsp;-&nbsp;0.099&nbsp;for&nbsp;0.018&nbsp;&lt;&nbsp;I</entry>
 496             <entry>0.2125&nbsp;E'<subscript>R</subscript>
 497 +&nbsp;0.7154&nbsp;E'<subscript>G</subscript>
 498 +&nbsp;0.0721&nbsp;E'<subscript>B</subscript></entry>
 499             <entry>219&nbsp;E'<subscript>Y</subscript>&nbsp;+&nbsp;16</entry>
 500             <entry>224&nbsp;P<subscript>B,R</subscript>&nbsp;+&nbsp;128</entry>
 501           </row>
 502           <row>
 503             <entry><constant>V4L2_COLORSPACE_BT878</constant></entry>
 504             <entry>4</entry>
 505             <entry>Broken Bt878 extents<footnote>
 506                 <para>The ubiquitous Bt878 video capture chip
 507 quantizes E'<subscript>Y</subscript> to 238 levels, yielding a range
 508 of Y' = 16 &hellip; 253, unlike Rec. 601 Y' = 16 &hellip;
 509 235. This is not a typo in the Bt878 documentation, it has been
 510 implemented in silicon. The chroma extents are unclear.</para>
 511               </footnote>, <xref linkend="itu601" /></entry>
 512             <entry>?</entry>
 513             <entry>?</entry>
 514             <entry>?</entry>
 515             <entry>?</entry>
 516             <entry>?</entry>
 517             <entry>0.299&nbsp;E'<subscript>R</subscript>
 518 +&nbsp;0.587&nbsp;E'<subscript>G</subscript>
 519 +&nbsp;0.114&nbsp;E'<subscript>B</subscript></entry>
 520             <entry><emphasis>237</emphasis>&nbsp;E'<subscript>Y</subscript>&nbsp;+&nbsp;16</entry>
 521             <entry>224&nbsp;P<subscript>B,R</subscript>&nbsp;+&nbsp;128 (probably)</entry>
 522           </row>
 523           <row>
 524             <entry><constant>V4L2_COLORSPACE_470_SYSTEM_M</constant></entry>
 525             <entry>5</entry>
 526             <entry>M/NTSC<footnote>
 527                 <para>No identifier exists for M/PAL which uses
 528 the chromaticities of M/NTSC, the remaining parameters are equal to B and
 529 G/PAL.</para>
 530               </footnote> according to <xref linkend="itu470" />, <xref
 531                 linkend="itu601" /></entry>
 532             <entry>x&nbsp;=&nbsp;0.67, y&nbsp;=&nbsp;0.33</entry>
 533             <entry>x&nbsp;=&nbsp;0.21, y&nbsp;=&nbsp;0.71</entry>
 534             <entry>x&nbsp;=&nbsp;0.14, y&nbsp;=&nbsp;0.08</entry>
 535             <entry>x&nbsp;=&nbsp;0.310, y&nbsp;=&nbsp;0.316, Illuminant C</entry>
 536             <entry>?</entry>
 537             <entry>0.299&nbsp;E'<subscript>R</subscript>
 538 +&nbsp;0.587&nbsp;E'<subscript>G</subscript>
 539 +&nbsp;0.114&nbsp;E'<subscript>B</subscript></entry>
 540             <entry>219&nbsp;E'<subscript>Y</subscript>&nbsp;+&nbsp;16</entry>
 541             <entry>224&nbsp;P<subscript>B,R</subscript>&nbsp;+&nbsp;128</entry>
 542           </row>
 543           <row>
 544             <entry><constant>V4L2_COLORSPACE_470_SYSTEM_BG</constant></entry>
 545             <entry>6</entry>
 546             <entry>625-line PAL and SECAM systems according to <xref
 547 linkend="itu470" />, <xref linkend="itu601" /></entry>
 548             <entry>x&nbsp;=&nbsp;0.64, y&nbsp;=&nbsp;0.33</entry>
 549             <entry>x&nbsp;=&nbsp;0.29, y&nbsp;=&nbsp;0.60</entry>
 550             <entry>x&nbsp;=&nbsp;0.15, y&nbsp;=&nbsp;0.06</entry>
 551             <entry>x&nbsp;=&nbsp;0.313, y&nbsp;=&nbsp;0.329,
 552 Illuminant D<subscript>65</subscript></entry>
 553             <entry>?</entry>
 554             <entry>0.299&nbsp;E'<subscript>R</subscript>
 555 +&nbsp;0.587&nbsp;E'<subscript>G</subscript>
 556 +&nbsp;0.114&nbsp;E'<subscript>B</subscript></entry>
 557             <entry>219&nbsp;E'<subscript>Y</subscript>&nbsp;+&nbsp;16</entry>
 558             <entry>224&nbsp;P<subscript>B,R</subscript>&nbsp;+&nbsp;128</entry>
 559           </row>
 560           <row>
 561             <entry><constant>V4L2_COLORSPACE_JPEG</constant></entry>
 562             <entry>7</entry>
 563             <entry>JPEG Y'CbCr, see <xref linkend="jfif" />, <xref linkend="itu601" /></entry>
 564             <entry>?</entry>
 565             <entry>?</entry>
 566             <entry>?</entry>
 567             <entry>?</entry>
 568             <entry>?</entry>
 569             <entry>0.299&nbsp;E'<subscript>R</subscript>
 570 +&nbsp;0.587&nbsp;E'<subscript>G</subscript>
 571 +&nbsp;0.114&nbsp;E'<subscript>B</subscript></entry>
 572             <entry>256&nbsp;E'<subscript>Y</subscript>&nbsp;+&nbsp;16<footnote>
 573                 <para>Note JFIF quantizes
 574 Y'P<subscript>B</subscript>P<subscript>R</subscript> in range [0;+1] and
 575 [-0.5;+0.5] to <emphasis>257</emphasis> levels, however Y'CbCr signals
 576 are still clamped to [0;255].</para>
 577               </footnote></entry>
 578             <entry>256&nbsp;P<subscript>B,R</subscript>&nbsp;+&nbsp;128</entry>
 579           </row>
 580           <row>
 581             <entry><constant>V4L2_COLORSPACE_SRGB</constant></entry>
 582             <entry>8</entry>
 583             <entry>[?]</entry>
 584             <entry>x&nbsp;=&nbsp;0.640, y&nbsp;=&nbsp;0.330</entry>
 585             <entry>x&nbsp;=&nbsp;0.300, y&nbsp;=&nbsp;0.600</entry>
 586             <entry>x&nbsp;=&nbsp;0.150, y&nbsp;=&nbsp;0.060</entry>
 587             <entry>x&nbsp;=&nbsp;0.3127, y&nbsp;=&nbsp;0.3290,
 588             Illuminant D<subscript>65</subscript></entry>
 589             <entry>E' = 4.5&nbsp;I&nbsp;for&nbsp;I&nbsp;&le;0.018,
 590 1.099&nbsp;I<superscript>0.45</superscript>&nbsp;-&nbsp;0.099&nbsp;for&nbsp;0.018&nbsp;&lt;&nbsp;I</entry>
 591             <entry spanname="spam">n/a</entry>
 592           </row>
 593         </tbody>
 594       </tgroup>
 595     </table>
 596   </section>
 597
 598   <section id="pixfmt-indexed">
 599     <title>Indexed Format</title>
 600
 601     <para>In this format each pixel is represented by an 8 bit index
 602 into a 256 entry ARGB palette. It is intended for <link
 603 linkend="osd">Video Output Overlays</link> only. There are no ioctls to
 604 access the palette, this must be done with ioctls of the Linux framebuffer API.</para>
 605
 606     <table pgwide="0" frame="none">
 607       <title>Indexed Image Format</title>
 608       <tgroup cols="37" align="center">
 609         <colspec colname="id" align="left" />
 610         <colspec colname="fourcc" />
 611         <colspec colname="bit" />
 612
 613         <colspec colnum="4" colname="b07" align="center" />
 614         <colspec colnum="5" colname="b06" align="center" />
 615         <colspec colnum="6" colname="b05" align="center" />
 616         <colspec colnum="7" colname="b04" align="center" />
 617         <colspec colnum="8" colname="b03" align="center" />
 618         <colspec colnum="9" colname="b02" align="center" />
 619         <colspec colnum="10" colname="b01" align="center" />
 620         <colspec colnum="11" colname="b00" align="center" />
 621
 622         <spanspec namest="b07" nameend="b00" spanname="b0" />
 623         <spanspec namest="b17" nameend="b10" spanname="b1" />
 624         <spanspec namest="b27" nameend="b20" spanname="b2" />
 625         <spanspec namest="b37" nameend="b30" spanname="b3" />
 626         <thead>
 627           <row>
 628             <entry>Identifier</entry>
 629             <entry>Code</entry>
 630             <entry>&nbsp;</entry>
 631             <entry spanname="b0">Byte&nbsp;0</entry>
 632           </row>
 633           <row>
 634             <entry>&nbsp;</entry>
 635             <entry>&nbsp;</entry>
 636             <entry>Bit</entry>
 637             <entry>7</entry>
 638             <entry>6</entry>
 639             <entry>5</entry>
 640             <entry>4</entry>
 641             <entry>3</entry>
 642             <entry>2</entry>
 643             <entry>1</entry>
 644             <entry>0</entry>
 645           </row>
 646         </thead>
 647         <tbody valign="top">
 648           <row id="V4L2-PIX-FMT-PAL8">
 649             <entry><constant>V4L2_PIX_FMT_PAL8</constant></entry>
 650             <entry>'PAL8'</entry>
 651             <entry></entry>
 652             <entry>i<subscript>7</subscript></entry>
 653             <entry>i<subscript>6</subscript></entry>
 654             <entry>i<subscript>5</subscript></entry>
 655             <entry>i<subscript>4</subscript></entry>
 656             <entry>i<subscript>3</subscript></entry>
 657             <entry>i<subscript>2</subscript></entry>
 658             <entry>i<subscript>1</subscript></entry>
 659             <entry>i<subscript>0</subscript></entry>
 660           </row>
 661         </tbody>
 662       </tgroup>
 663     </table>
 664   </section>
 665
 666   <section id="pixfmt-rgb">
 667     <title>RGB Formats</title>
 668
 669     &sub-packed-rgb;
 670     &sub-sbggr8;
 671     &sub-sgbrg8;
 672     &sub-sgrbg8;
 673     &sub-srggb8;
 674     &sub-sbggr16;
 675     &sub-srggb10;
 676   </section>
 677
 678   <section id="yuv-formats">
 679     <title>YUV Formats</title>
 680
 681     <para>YUV is the format native to TV broadcast and composite video
 682 signals. It separates the brightness information (Y) from the color
 683 information (U and V or Cb and Cr). The color information consists of
 684 red and blue <emphasis>color difference</emphasis> signals, this way
 685 the green component can be reconstructed by subtracting from the
 686 brightness component. See <xref linkend="colorspaces" /> for conversion
 687 examples. YUV was chosen because early television would only transmit
 688 brightness information. To add color in a way compatible with existing
 689 receivers a new signal carrier was added to transmit the color
 690 difference signals. Secondary in the YUV format the U and V components
 691 usually have lower resolution than the Y component. This is an analog
 692 video compression technique taking advantage of a property of the
 693 human visual system, being more sensitive to brightness
 694 information.</para>
 695
 696     &sub-packed-yuv;
 697     &sub-grey;
 698     &sub-y10;
 699     &sub-y12;
 700     &sub-y16;
 701     &sub-yuyv;
 702     &sub-uyvy;
 703     &sub-yvyu;
 704     &sub-vyuy;
 705     &sub-y41p;
 706     &sub-yuv420;
 707     &sub-yuv420m;
 708     &sub-yuv410;
 709     &sub-yuv422p;
 710     &sub-yuv411p;
 711     &sub-nv12;
 712     &sub-nv12m;
 713     &sub-nv12mt;
 714     &sub-nv16;
 715   </section>
 716
 717   <section>
 718     <title>Compressed Formats</title>
 719
 720     <table pgwide="1" frame="none" id="compressed-formats">
 721       <title>Compressed Image Formats</title>
 722       <tgroup cols="3" align="left">
 723         &cs-def;
 724         <thead>
 725           <row>
 726             <entry>Identifier</entry>
 727             <entry>Code</entry>
 728             <entry>Details</entry>
 729           </row>
 730         </thead>
 731         <tbody valign="top">
 732          <row id="V4L2-PIX-FMT-JPEG">
 733             <entry><constant>V4L2_PIX_FMT_JPEG</constant></entry>
 734             <entry>'JPEG'</entry>
 735             <entry>TBD. See also &VIDIOC-G-JPEGCOMP;,
 736             &VIDIOC-S-JPEGCOMP;.</entry>
 737           </row>
 738           <row id="V4L2-PIX-FMT-MPEG">
 739             <entry><constant>V4L2_PIX_FMT_MPEG</constant></entry>
 740             <entry>'MPEG'</entry>
 741             <entry>MPEG stream. The actual format is determined by
 742 extended control <constant>V4L2_CID_MPEG_STREAM_TYPE</constant>, see
 743 <xref linkend="mpeg-control-id" />.</entry>
 744           </row>
 745         </tbody>
 746       </tgroup>
 747     </table>
 748   </section>
 749
 750   <section id="pixfmt-reserved">
 751     <title>Reserved Format Identifiers</title>
 752
 753     <para>These formats are not defined by this specification, they
 754 are just listed for reference and to avoid naming conflicts. If you
 755 want to register your own format, send an e-mail to the linux-media mailing
 756 list &v4l-ml; for inclusion in the <filename>videodev2.h</filename>
 757 file. If you want to share your format with other developers add a
 758 link to your documentation and send a copy to the linux-media mailing list
 759 for inclusion in this section. If you think your format should be listed
 760 in a standard format section please make a proposal on the linux-media mailing
 761 list.</para>
 762
 763     <table pgwide="1" frame="none" id="reserved-formats">
 764       <title>Reserved Image Formats</title>
 765       <tgroup cols="3" align="left">
 766         &cs-def;
 767         <thead>
 768           <row>
 769             <entry>Identifier</entry>
 770             <entry>Code</entry>
 771             <entry>Details</entry>
 772           </row>
 773         </thead>
 774         <tbody valign="top">
 775           <row id="V4L2-PIX-FMT-DV">
 776             <entry><constant>V4L2_PIX_FMT_DV</constant></entry>
 777             <entry>'dvsd'</entry>
 778             <entry>unknown</entry>
 779           </row>
 780           <row id="V4L2-PIX-FMT-ET61X251">
 781             <entry><constant>V4L2_PIX_FMT_ET61X251</constant></entry>
 782             <entry>'E625'</entry>
 783             <entry>Compressed format of the ET61X251 driver.</entry>
 784           </row>
 785           <row id="V4L2-PIX-FMT-HI240">
 786             <entry><constant>V4L2_PIX_FMT_HI240</constant></entry>
 787             <entry>'HI24'</entry>
 788             <entry><para>8 bit RGB format used by the BTTV driver.</para></entry>
 789           </row>
 790           <row id="V4L2-PIX-FMT-HM12">
 791             <entry><constant>V4L2_PIX_FMT_HM12</constant></entry>
 792             <entry>'HM12'</entry>
 793             <entry><para>YUV 4:2:0 format used by the
 794 IVTV driver, <ulink url="http://www.ivtvdriver.org/">
 795 http://www.ivtvdriver.org/</ulink></para><para>The format is documented in the
 796 kernel sources in the file <filename>Documentation/video4linux/cx2341x/README.hm12</filename>
 797 </para></entry>
 798           </row>
 799           <row id="V4L2-PIX-FMT-CPIA1">
 800             <entry><constant>V4L2_PIX_FMT_CPIA1</constant></entry>
 801             <entry>'CPIA'</entry>
 802             <entry>YUV format used by the gspca cpia1 driver.</entry>
 803           </row>
 804           <row id="V4L2-PIX-FMT-SPCA501">
 805             <entry><constant>V4L2_PIX_FMT_SPCA501</constant></entry>
 806             <entry>'S501'</entry>
 807             <entry>YUYV per line used by the gspca driver.</entry>
 808           </row>
 809           <row id="V4L2-PIX-FMT-SPCA505">
 810             <entry><constant>V4L2_PIX_FMT_SPCA505</constant></entry>
 811             <entry>'S505'</entry>
 812             <entry>YYUV per line used by the gspca driver.</entry>
 813           </row>
 814           <row id="V4L2-PIX-FMT-SPCA508">
 815             <entry><constant>V4L2_PIX_FMT_SPCA508</constant></entry>
 816             <entry>'S508'</entry>
 817             <entry>YUVY per line used by the gspca driver.</entry>
 818           </row>
 819           <row id="V4L2-PIX-FMT-SPCA561">
 820             <entry><constant>V4L2_PIX_FMT_SPCA561</constant></entry>
 821             <entry>'S561'</entry>
 822             <entry>Compressed GBRG Bayer format used by the gspca driver.</entry>
 823           </row>
 824           <row id="V4L2-PIX-FMT-SGRBG10DPCM8">
 825             <entry><constant>V4L2_PIX_FMT_SGRBG10DPCM8</constant></entry>
 826             <entry>'DB10'</entry>
 827             <entry>10 bit raw Bayer DPCM compressed to 8 bits.</entry>
 828           </row>
 829           <row id="V4L2-PIX-FMT-PAC207">
 830             <entry><constant>V4L2_PIX_FMT_PAC207</constant></entry>
 831             <entry>'P207'</entry>
 832             <entry>Compressed BGGR Bayer format used by the gspca driver.</entry>
 833           </row>
 834           <row id="V4L2-PIX-FMT-MR97310A">
 835             <entry><constant>V4L2_PIX_FMT_MR97310A</constant></entry>
 836             <entry>'M310'</entry>
 837             <entry>Compressed BGGR Bayer format used by the gspca driver.</entry>
 838           </row>
 839           <row id="V4L2-PIX-FMT-OV511">
 840             <entry><constant>V4L2_PIX_FMT_OV511</constant></entry>
 841             <entry>'O511'</entry>
 842             <entry>OV511 JPEG format used by the gspca driver.</entry>
 843           </row>
 844           <row id="V4L2-PIX-FMT-OV518">
 845             <entry><constant>V4L2_PIX_FMT_OV518</constant></entry>
 846             <entry>'O518'</entry>
 847             <entry>OV518 JPEG format used by the gspca driver.</entry>
 848           </row>
 849           <row id="V4L2-PIX-FMT-PJPG">
 850             <entry><constant>V4L2_PIX_FMT_PJPG</constant></entry>
 851             <entry>'PJPG'</entry>
 852             <entry>Pixart 73xx JPEG format used by the gspca driver.</entry>
 853           </row>
 854           <row id="V4L2-PIX-FMT-SQ905C">
 855             <entry><constant>V4L2_PIX_FMT_SQ905C</constant></entry>
 856             <entry>'905C'</entry>
 857             <entry>Compressed RGGB bayer format used by the gspca driver.</entry>
 858           </row>
 859           <row id="V4L2-PIX-FMT-MJPEG">
 860             <entry><constant>V4L2_PIX_FMT_MJPEG</constant></entry>
 861             <entry>'MJPG'</entry>
 862             <entry>Compressed format used by the Zoran driver</entry>
 863           </row>
 864           <row id="V4L2-PIX-FMT-PWC1">
 865             <entry><constant>V4L2_PIX_FMT_PWC1</constant></entry>
 866             <entry>'PWC1'</entry>
 867             <entry>Compressed format of the PWC driver.</entry>
 868           </row>
 869           <row id="V4L2-PIX-FMT-PWC2">
 870             <entry><constant>V4L2_PIX_FMT_PWC2</constant></entry>
 871             <entry>'PWC2'</entry>
 872             <entry>Compressed format of the PWC driver.</entry>
 873           </row>
 874           <row id="V4L2-PIX-FMT-SN9C10X">
 875             <entry><constant>V4L2_PIX_FMT_SN9C10X</constant></entry>
 876             <entry>'S910'</entry>
 877             <entry>Compressed format of the SN9C102 driver.</entry>
 878           </row>
 879           <row id="V4L2-PIX-FMT-SN9C20X-I420">
 880             <entry><constant>V4L2_PIX_FMT_SN9C20X_I420</constant></entry>
 881             <entry>'S920'</entry>
 882             <entry>YUV 4:2:0 format of the gspca sn9c20x driver.</entry>
 883           </row>
 884           <row id="V4L2-PIX-FMT-SN9C2028">
 885             <entry><constant>V4L2_PIX_FMT_SN9C2028</constant></entry>
 886             <entry>'SONX'</entry>
 887             <entry>Compressed GBRG bayer format of the gspca sn9c2028 driver.</entry>
 888           </row>
 889           <row id="V4L2-PIX-FMT-STV0680">
 890             <entry><constant>V4L2_PIX_FMT_STV0680</constant></entry>
 891             <entry>'S680'</entry>
 892             <entry>Bayer format of the gspca stv0680 driver.</entry>
 893           </row>
 894           <row id="V4L2-PIX-FMT-WNVA">
 895             <entry><constant>V4L2_PIX_FMT_WNVA</constant></entry>
 896             <entry>'WNVA'</entry>
 897             <entry><para>Used by the Winnov Videum driver, <ulink
 898 url="http://www.thedirks.org/winnov/">
 899 http://www.thedirks.org/winnov/</ulink></para></entry>
 900           </row>
 901           <row id="V4L2-PIX-FMT-TM6000">
 902             <entry><constant>V4L2_PIX_FMT_TM6000</constant></entry>
 903             <entry>'TM60'</entry>
 904             <entry><para>Used by Trident tm6000</para></entry>
 905           </row>
 906           <row id="V4L2-PIX-FMT-CIT-YYVYUY">
 907             <entry><constant>V4L2_PIX_FMT_CIT_YYVYUY</constant></entry>
 908             <entry>'CITV'</entry>
 909             <entry><para>Used by xirlink CIT, found at IBM webcams.</para>
 910                    <para>Uses one line of Y then 1 line of VYUY</para>
 911             </entry>
 912           </row>
 913           <row id="V4L2-PIX-FMT-KONICA420">
 914             <entry><constant>V4L2_PIX_FMT_KONICA420</constant></entry>
 915             <entry>'KONI'</entry>
 916             <entry><para>Used by Konica webcams.</para>
 917                    <para>YUV420 planar in blocks of 256 pixels.</para>
 918             </entry>
 919           </row>
 920           <row id="V4L2-PIX-FMT-YYUV">
 921             <entry><constant>V4L2_PIX_FMT_YYUV</constant></entry>
 922             <entry>'YYUV'</entry>
 923             <entry>unknown</entry>
 924           </row>
 925           <row id="V4L2-PIX-FMT-Y4">
 926             <entry><constant>V4L2_PIX_FMT_Y4</constant></entry>
 927             <entry>'Y04 '</entry>
 928             <entry>Old 4-bit greyscale format. Only the least significant 4 bits of each byte are used,
 929 the other bits are set to 0.</entry>
 930           </row>
 931           <row id="V4L2-PIX-FMT-Y6">
 932             <entry><constant>V4L2_PIX_FMT_Y6</constant></entry>
 933             <entry>'Y06 '</entry>
 934             <entry>Old 6-bit greyscale format. Only the least significant 6 bits of each byte are used,
 935 the other bits are set to 0.</entry>
 936           </row>
 937         </tbody>
 938       </tgroup>
 939     </table>
 940   </section>
 941
 942   <!--
 943 Local Variables:
 944 mode: sgml
 945 sgml-parent-document: "v4l2.sgml"
 946 indent-tabs-mode: nil
 947 End:
 948   -->