1 <title>Image Formats</title>
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>
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>
22 <entry><structfield>width</structfield></entry>
23 <entry>Image width in pixels.</entry>
27 <entry><structfield>height</structfield></entry>
28 <entry>Image height in pixels. If <structfield>field</structfield> is
29 one of <constant>V4L2_FIELD_TOP</constant>, <constant>V4L2_FIELD_BOTTOM</constant>
30 or <constant>V4L2_FIELD_ALTERNATE</constant> then height refers to the
31 number of lines in the field, otherwise it refers to the number of
32 lines in the frame (which is twice the field height for interlaced
36 <entry spanname="hspan">Applications set these fields to
37 request an image size, drivers return the closest possible values. In
38 case of planar formats the <structfield>width</structfield> and
39 <structfield>height</structfield> applies to the largest plane. To
40 avoid ambiguities drivers must return values rounded up to a multiple
41 of the scale factor of any smaller planes. For example when the image
42 format is YUV 4:2:0, <structfield>width</structfield> and
43 <structfield>height</structfield> must be multiples of two.</entry>
47 <entry><structfield>pixelformat</structfield></entry>
48 <entry>The pixel format or type of compression, set by the
49 application. This is a little endian <link
50 linkend="v4l2-fourcc">four character code</link>. V4L2 defines
51 standard RGB formats in <xref linkend="rgb-formats" />, YUV formats in <xref
52 linkend="yuv-formats" />, and reserved codes in <xref
53 linkend="reserved-formats" /></entry>
56 <entry>&v4l2-field;</entry>
57 <entry><structfield>field</structfield></entry>
58 <entry>Video images are typically interlaced. Applications
59 can request to capture or output only the top or bottom field, or both
60 fields interlaced or sequentially stored in one buffer or alternating
61 in separate buffers. Drivers return the actual field order selected.
62 For more details on fields see <xref linkend="field-order" />.</entry>
66 <entry><structfield>bytesperline</structfield></entry>
67 <entry>Distance in bytes between the leftmost pixels in two
68 adjacent lines.</entry>
71 <entry spanname="hspan"><para>Both applications and drivers
72 can set this field to request padding bytes at the end of each line.
73 Drivers however may ignore the value requested by the application,
74 returning <structfield>width</structfield> times bytes per pixel or a
75 larger value required by the hardware. That implies applications can
76 just set this field to zero to get a reasonable
77 default.</para><para>Video hardware may access padding bytes,
78 therefore they must reside in accessible memory. Consider cases where
79 padding bytes after the last line of an image cross a system page
80 boundary. Input devices may write padding bytes, the value is
81 undefined. Output devices ignore the contents of padding
82 bytes.</para><para>When the image format is planar the
83 <structfield>bytesperline</structfield> value applies to the first
84 plane and is divided by the same factor as the
85 <structfield>width</structfield> field for the other planes. For
86 example the Cb and Cr planes of a YUV 4:2:0 image have half as many
87 padding bytes following each line as the Y plane. To avoid ambiguities
88 drivers must return a <structfield>bytesperline</structfield> value
89 rounded up to a multiple of the scale factor.</para>
90 <para>For compressed formats the <structfield>bytesperline</structfield>
91 value makes no sense. Applications and drivers must set this to 0 in
92 that case.</para></entry>
96 <entry><structfield>sizeimage</structfield></entry>
97 <entry>Size in bytes of the buffer to hold a complete image,
98 set by the driver. Usually this is
99 <structfield>bytesperline</structfield> times
100 <structfield>height</structfield>. When the image consists of variable
101 length compressed data this is the maximum number of bytes required to
102 hold an image.</entry>
105 <entry>&v4l2-colorspace;</entry>
106 <entry><structfield>colorspace</structfield></entry>
107 <entry>This information supplements the
108 <structfield>pixelformat</structfield> and must be set by the driver for
109 capture streams and by the application for output streams,
110 see <xref linkend="colorspaces" />.</entry>
114 <entry><structfield>priv</structfield></entry>
115 <entry><para>This field indicates whether the remaining fields of the
116 <structname>v4l2_pix_format</structname> structure, also called the extended
117 fields, are valid. When set to <constant>V4L2_PIX_FMT_PRIV_MAGIC</constant>, it
118 indicates that the extended fields have been correctly initialized. When set to
119 any other value it indicates that the extended fields contain undefined values.
121 <para>Applications that wish to use the pixel format extended fields must first
122 ensure that the feature is supported by querying the device for the
123 <link linkend="querycap"><constant>V4L2_CAP_EXT_PIX_FORMAT</constant></link>
124 capability. If the capability isn't set the pixel format extended fields are not
125 supported and using the extended fields will lead to undefined results.</para>
126 <para>To use the extended fields, applications must set the
127 <structfield>priv</structfield> field to
128 <constant>V4L2_PIX_FMT_PRIV_MAGIC</constant>, initialize all the extended fields
129 and zero the unused bytes of the <structname>v4l2_format</structname>
130 <structfield>raw_data</structfield> field.</para>
131 <para>When the <structfield>priv</structfield> field isn't set to
132 <constant>V4L2_PIX_FMT_PRIV_MAGIC</constant> drivers must act as if all the
133 extended fields were set to zero. On return drivers must set the
134 <structfield>priv</structfield> field to
135 <constant>V4L2_PIX_FMT_PRIV_MAGIC</constant> and all the extended fields to
136 applicable values.</para></entry>
140 <entry><structfield>flags</structfield></entry>
141 <entry>Flags set by the application or driver, see <xref
142 linkend="format-flags" />.</entry>
145 <entry>&v4l2-ycbcr-encoding;</entry>
146 <entry><structfield>ycbcr_enc</structfield></entry>
147 <entry>This information supplements the
148 <structfield>colorspace</structfield> and must be set by the driver for
149 capture streams and by the application for output streams,
150 see <xref linkend="colorspaces" />.</entry>
153 <entry>&v4l2-quantization;</entry>
154 <entry><structfield>quantization</structfield></entry>
155 <entry>This information supplements the
156 <structfield>colorspace</structfield> and must be set by the driver for
157 capture streams and by the application for output streams,
158 see <xref linkend="colorspaces" />.</entry>
161 <entry>&v4l2-xfer-func;</entry>
162 <entry><structfield>xfer_func</structfield></entry>
163 <entry>This information supplements the
164 <structfield>colorspace</structfield> and must be set by the driver for
165 capture streams and by the application for output streams,
166 see <xref linkend="colorspaces" />.</entry>
174 <title>Multi-planar format structures</title>
175 <para>The <structname>v4l2_plane_pix_format</structname> structures define
176 size and layout for each of the planes in a multi-planar format.
177 The <structname>v4l2_pix_format_mplane</structname> structure contains
178 information common to all planes (such as image width and height) and
179 an array of <structname>v4l2_plane_pix_format</structname> structures,
180 describing all planes of that format.</para>
181 <table pgwide="1" frame="none" id="v4l2-plane-pix-format">
182 <title>struct <structname>v4l2_plane_pix_format</structname></title>
188 <entry><structfield>sizeimage</structfield></entry>
189 <entry>Maximum size in bytes required for image data in this plane.
194 <entry><structfield>bytesperline</structfield></entry>
195 <entry>Distance in bytes between the leftmost pixels in two adjacent
196 lines. See &v4l2-pix-format;.</entry>
200 <entry><structfield>reserved[6]</structfield></entry>
201 <entry>Reserved for future extensions. Should be zeroed by drivers and
202 applications.</entry>
207 <table pgwide="1" frame="none" id="v4l2-pix-format-mplane">
208 <title>struct <structname>v4l2_pix_format_mplane</structname></title>
214 <entry><structfield>width</structfield></entry>
215 <entry>Image width in pixels. See &v4l2-pix-format;.</entry>
219 <entry><structfield>height</structfield></entry>
220 <entry>Image height in pixels. See &v4l2-pix-format;.</entry>
224 <entry><structfield>pixelformat</structfield></entry>
225 <entry>The pixel format. Both single- and multi-planar four character
226 codes can be used.</entry>
229 <entry>&v4l2-field;</entry>
230 <entry><structfield>field</structfield></entry>
231 <entry>See &v4l2-pix-format;.</entry>
234 <entry>&v4l2-colorspace;</entry>
235 <entry><structfield>colorspace</structfield></entry>
236 <entry>See &v4l2-pix-format;.</entry>
239 <entry>&v4l2-plane-pix-format;</entry>
240 <entry><structfield>plane_fmt[VIDEO_MAX_PLANES]</structfield></entry>
241 <entry>An array of structures describing format of each plane this
242 pixel format consists of. The number of valid entries in this array
243 has to be put in the <structfield>num_planes</structfield>
248 <entry><structfield>num_planes</structfield></entry>
249 <entry>Number of planes (i.e. separate memory buffers) for this format
250 and the number of valid entries in the
251 <structfield>plane_fmt</structfield> array.</entry>
255 <entry><structfield>flags</structfield></entry>
256 <entry>Flags set by the application or driver, see <xref
257 linkend="format-flags" />.</entry>
260 <entry>&v4l2-ycbcr-encoding;</entry>
261 <entry><structfield>ycbcr_enc</structfield></entry>
262 <entry>This information supplements the
263 <structfield>colorspace</structfield> and must be set by the driver for
264 capture streams and by the application for output streams,
265 see <xref linkend="colorspaces" />.</entry>
268 <entry>&v4l2-quantization;</entry>
269 <entry><structfield>quantization</structfield></entry>
270 <entry>This information supplements the
271 <structfield>colorspace</structfield> and must be set by the driver for
272 capture streams and by the application for output streams,
273 see <xref linkend="colorspaces" />.</entry>
276 <entry>&v4l2-xfer-func;</entry>
277 <entry><structfield>xfer_func</structfield></entry>
278 <entry>This information supplements the
279 <structfield>colorspace</structfield> and must be set by the driver for
280 capture streams and by the application for output streams,
281 see <xref linkend="colorspaces" />.</entry>
285 <entry><structfield>reserved[7]</structfield></entry>
286 <entry>Reserved for future extensions. Should be zeroed by drivers
287 and applications.</entry>
295 <title>Standard Image Formats</title>
297 <para>In order to exchange images between drivers and
298 applications, it is necessary to have standard image data formats
299 which both sides will interpret the same way. V4L2 includes several
300 such formats, and this section is intended to be an unambiguous
301 specification of the standard image data formats in V4L2.</para>
303 <para>V4L2 drivers are not limited to these formats, however.
304 Driver-specific formats are possible. In that case the application may
305 depend on a codec to convert images to one of the standard formats
306 when needed. But the data can still be stored and retrieved in the
307 proprietary format. For example, a device may support a proprietary
308 compressed format. Applications can still capture and save the data in
309 the compressed format, saving much disk space, and later use a codec
310 to convert the images to the X Windows screen format when the video is
311 to be displayed.</para>
313 <para>Even so, ultimately, some standard formats are needed, so
314 the V4L2 specification would not be complete without well-defined
315 standard formats.</para>
317 <para>The V4L2 standard formats are mainly uncompressed formats. The
318 pixels are always arranged in memory from left to right, and from top
319 to bottom. The first byte of data in the image buffer is always for
320 the leftmost pixel of the topmost row. Following that is the pixel
321 immediately to its right, and so on until the end of the top row of
322 pixels. Following the rightmost pixel of the row there may be zero or
323 more bytes of padding to guarantee that each row of pixel data has a
324 certain alignment. Following the pad bytes, if any, is data for the
325 leftmost pixel of the second row from the top, and so on. The last row
326 has just as many pad bytes after it as the other rows.</para>
328 <para>In V4L2 each format has an identifier which looks like
329 <constant>PIX_FMT_XXX</constant>, defined in the <link
330 linkend="videodev">videodev2.h</link> header file. These identifiers
331 represent <link linkend="v4l2-fourcc">four character (FourCC) codes</link>
332 which are also listed below, however they are not the same as those
333 used in the Windows world.</para>
335 <para>For some formats, data is stored in separate, discontiguous
336 memory buffers. Those formats are identified by a separate set of FourCC codes
337 and are referred to as "multi-planar formats". For example, a YUV422 frame is
338 normally stored in one memory buffer, but it can also be placed in two or three
339 separate buffers, with Y component in one buffer and CbCr components in another
340 in the 2-planar version or with each component in its own buffer in the
341 3-planar case. Those sub-buffers are referred to as "planes".</para>
344 <section id="colorspaces">
345 <title>Colorspaces</title>
347 <para>'Color' is a very complex concept and depends on physics, chemistry and
348 biology. Just because you have three numbers that describe the 'red', 'green'
349 and 'blue' components of the color of a pixel does not mean that you can accurately
350 display that color. A colorspace defines what it actually <emphasis>means</emphasis>
351 to have an RGB value of e.g. (255, 0, 0). That is, which color should be
352 reproduced on the screen in a perfectly calibrated environment.</para>
354 <para>In order to do that we first need to have a good definition of
355 color, i.e. some way to uniquely and unambiguously define a color so that someone
356 else can reproduce it. Human color vision is trichromatic since the human eye has
357 color receptors that are sensitive to three different wavelengths of light. Hence
358 the need to use three numbers to describe color. Be glad you are not a mantis shrimp
359 as those are sensitive to 12 different wavelengths, so instead of RGB we would be
360 using the ABCDEFGHIJKL colorspace...</para>
362 <para>Color exists only in the eye and brain and is the result of how strongly
363 color receptors are stimulated. This is based on the Spectral
364 Power Distribution (SPD) which is a graph showing the intensity (radiant power)
365 of the light at wavelengths covering the visible spectrum as it enters the eye.
366 The science of colorimetry is about the relationship between the SPD and color as
367 perceived by the human brain.</para>
369 <para>Since the human eye has only three color receptors it is perfectly
370 possible that different SPDs will result in the same stimulation of those receptors
371 and are perceived as the same color, even though the SPD of the light is
374 <para>In the 1920s experiments were devised to determine the relationship
375 between SPDs and the perceived color and that resulted in the CIE 1931 standard
376 that defines spectral weighting functions that model the perception of color.
377 Specifically that standard defines functions that can take an SPD and calculate
378 the stimulus for each color receptor. After some further mathematical transforms
379 these stimuli are known as the <emphasis>CIE XYZ tristimulus</emphasis> values
380 and these X, Y and Z values describe a color as perceived by a human unambiguously.
381 These X, Y and Z values are all in the range [0…1].</para>
383 <para>The Y value in the CIE XYZ colorspace corresponds to luminance. Often
384 the CIE XYZ colorspace is transformed to the normalized CIE xyY colorspace:</para>
386 <para>x = X / (X + Y + Z)</para>
387 <para>y = Y / (X + Y + Z)</para>
389 <para>The x and y values are the chromaticity coordinates and can be used to
390 define a color without the luminance component Y. It is very confusing to
391 have such similar names for these colorspaces. Just be aware that if colors
392 are specified with lower case 'x' and 'y', then the CIE xyY colorspace is
393 used. Upper case 'X' and 'Y' refer to the CIE XYZ colorspace. Also, y has nothing
394 to do with luminance. Together x and y specify a color, and Y the luminance.
395 That is really all you need to remember from a practical point of view. At
396 the end of this section you will find reading resources that go into much more
397 detail if you are interested.
400 <para>A monitor or TV will reproduce colors by emitting light at three
401 different wavelengths, the combination of which will stimulate the color receptors
402 in the eye and thus cause the perception of color. Historically these wavelengths
403 were defined by the red, green and blue phosphors used in the displays. These
404 <emphasis>color primaries</emphasis> are part of what defines a colorspace.</para>
406 <para>Different display devices will have different primaries and some
407 primaries are more suitable for some display technologies than others. This has
408 resulted in a variety of colorspaces that are used for different display
409 technologies or uses. To define a colorspace you need to define the three
410 color primaries (these are typically defined as x, y chromaticity coordinates
411 from the CIE xyY colorspace) but also the white reference: that is the color obtained
412 when all three primaries are at maximum power. This determines the relative power
413 or energy of the primaries. This is usually chosen to be close to daylight which has
414 been defined as the CIE D65 Illuminant.</para>
416 <para>To recapitulate: the CIE XYZ colorspace uniquely identifies colors.
417 Other colorspaces are defined by three chromaticity coordinates defined in the
418 CIE xyY colorspace. Based on those a 3x3 matrix can be constructed that
419 transforms CIE XYZ colors to colors in the new colorspace.
422 <para>Both the CIE XYZ and the RGB colorspace that are derived from the
423 specific chromaticity primaries are linear colorspaces. But neither the eye,
424 nor display technology is linear. Doubling the values of all components in
425 the linear colorspace will not be perceived as twice the intensity of the color.
426 So each colorspace also defines a transfer function that takes a linear color
427 component value and transforms it to the non-linear component value, which is a
428 closer match to the non-linear performance of both the eye and displays. Linear
429 component values are denoted RGB, non-linear are denoted as R'G'B'. In general
430 colors used in graphics are all R'G'B', except in openGL which uses linear RGB.
431 Special care should be taken when dealing with openGL to provide linear RGB colors
432 or to use the built-in openGL support to apply the inverse transfer function.</para>
434 <para>The final piece that defines a colorspace is a function that
435 transforms non-linear R'G'B' to non-linear Y'CbCr. This function is determined
436 by the so-called luma coefficients. There may be multiple possible Y'CbCr
437 encodings allowed for the same colorspace. Many encodings of color
438 prefer to use luma (Y') and chroma (CbCr) instead of R'G'B'. Since the human
439 eye is more sensitive to differences in luminance than in color this encoding
440 allows one to reduce the amount of color information compared to the luma
441 data. Note that the luma (Y') is unrelated to the Y in the CIE XYZ colorspace.
442 Also note that Y'CbCr is often called YCbCr or YUV even though these are
443 strictly speaking wrong.</para>
445 <para>Sometimes people confuse Y'CbCr as being a colorspace. This is not
446 correct, it is just an encoding of an R'G'B' color into luma and chroma
447 values. The underlying colorspace that is associated with the R'G'B' color
448 is also associated with the Y'CbCr color.</para>
450 <para>The final step is how the RGB, R'G'B' or Y'CbCr values are
451 quantized. The CIE XYZ colorspace where X, Y and Z are in the range
452 [0…1] describes all colors that humans can perceive, but the transform to
453 another colorspace will produce colors that are outside the [0…1] range.
454 Once clamped to the [0…1] range those colors can no longer be reproduced
455 in that colorspace. This clamping is what reduces the extent or gamut of the
456 colorspace. How the range of [0…1] is translated to integer values in the
457 range of [0…255] (or higher, depending on the color depth) is called the
458 quantization. This is <emphasis>not</emphasis> part of the colorspace
459 definition. In practice RGB or R'G'B' values are full range, i.e. they
460 use the full [0…255] range. Y'CbCr values on the other hand are limited
461 range with Y' using [16…235] and Cb and Cr using [16…240].</para>
463 <para>Unfortunately, in some cases limited range RGB is also used
464 where the components use the range [16…235]. And full range Y'CbCr also exists
465 using the [0…255] range.</para>
467 <para>In order to correctly interpret a color you need to know the
468 quantization range, whether it is R'G'B' or Y'CbCr, the used Y'CbCr encoding
470 From that information you can calculate the corresponding CIE XYZ color
471 and map that again to whatever colorspace your display device uses.</para>
473 <para>The colorspace definition itself consists of the three
474 chromaticity primaries, the white reference chromaticity, a transfer
475 function and the luma coefficients needed to transform R'G'B' to Y'CbCr. While
476 some colorspace standards correctly define all four, quite often the colorspace
477 standard only defines some, and you have to rely on other standards for
478 the missing pieces. The fact that colorspaces are often a mix of different
479 standards also led to very confusing naming conventions where the name of
480 a standard was used to name a colorspace when in fact that standard was
481 part of various other colorspaces as well.</para>
483 <para>If you want to read more about colors and colorspaces, then the
484 following resources are useful: <xref linkend="poynton" /> is a good practical
485 book for video engineers, <xref linkend="colimg" /> has a much broader scope and
486 describes many more aspects of color (physics, chemistry, biology, etc.).
487 The <ulink url="http://www.brucelindbloom.com">http://www.brucelindbloom.com</ulink>
488 website is an excellent resource, especially with respect to the mathematics behind
489 colorspace conversions. The wikipedia <ulink url="http://en.wikipedia.org/wiki/CIE_1931_color_space#CIE_xy_chromaticity_diagram_and_the_CIE_xyY_color_space">CIE 1931 colorspace</ulink> article
490 is also very useful.</para>
494 <title>Defining Colorspaces in V4L2</title>
495 <para>In V4L2 colorspaces are defined by four values. The first is the colorspace
496 identifier (&v4l2-colorspace;) which defines the chromaticities, the default transfer
497 function, the default Y'CbCr encoding and the default quantization method. The second
498 is the transfer function identifier (&v4l2-xfer-func;) to specify non-standard
499 transfer functions. The third is the Y'CbCr encoding identifier (&v4l2-ycbcr-encoding;)
500 to specify non-standard Y'CbCr encodings and the fourth is the quantization identifier
501 (&v4l2-quantization;) to specify non-standard quantization methods. Most of the time
502 only the colorspace field of &v4l2-pix-format; or &v4l2-pix-format-mplane; needs to
503 be filled in. Note that the default R'G'B' quantization is full range for all
504 colorspaces except for BT.2020 which uses limited range R'G'B' quantization.</para>
506 <table pgwide="1" frame="none" id="v4l2-colorspace">
507 <title>V4L2 Colorspaces</title>
508 <tgroup cols="2" align="left">
512 <entry>Identifier</entry>
513 <entry>Details</entry>
518 <entry><constant>V4L2_COLORSPACE_DEFAULT</constant></entry>
519 <entry>The default colorspace. This can be used by applications to let the
520 driver fill in the colorspace.</entry>
523 <entry><constant>V4L2_COLORSPACE_SMPTE170M</constant></entry>
524 <entry>See <xref linkend="col-smpte-170m" />.</entry>
527 <entry><constant>V4L2_COLORSPACE_REC709</constant></entry>
528 <entry>See <xref linkend="col-rec709" />.</entry>
531 <entry><constant>V4L2_COLORSPACE_SRGB</constant></entry>
532 <entry>See <xref linkend="col-srgb" />.</entry>
535 <entry><constant>V4L2_COLORSPACE_ADOBERGB</constant></entry>
536 <entry>See <xref linkend="col-adobergb" />.</entry>
539 <entry><constant>V4L2_COLORSPACE_BT2020</constant></entry>
540 <entry>See <xref linkend="col-bt2020" />.</entry>
543 <entry><constant>V4L2_COLORSPACE_DCI_P3</constant></entry>
544 <entry>See <xref linkend="col-dcip3" />.</entry>
547 <entry><constant>V4L2_COLORSPACE_SMPTE240M</constant></entry>
548 <entry>See <xref linkend="col-smpte-240m" />.</entry>
551 <entry><constant>V4L2_COLORSPACE_470_SYSTEM_M</constant></entry>
552 <entry>See <xref linkend="col-sysm" />.</entry>
555 <entry><constant>V4L2_COLORSPACE_470_SYSTEM_BG</constant></entry>
556 <entry>See <xref linkend="col-sysbg" />.</entry>
559 <entry><constant>V4L2_COLORSPACE_JPEG</constant></entry>
560 <entry>See <xref linkend="col-jpeg" />.</entry>
563 <entry><constant>V4L2_COLORSPACE_RAW</constant></entry>
564 <entry>The raw colorspace. This is used for raw image capture where
565 the image is minimally processed and is using the internal colorspace
566 of the device. The software that processes an image using this
567 'colorspace' will have to know the internals of the capture device.</entry>
573 <table pgwide="1" frame="none" id="v4l2-xfer-func">
574 <title>V4L2 Transfer Function</title>
575 <tgroup cols="2" align="left">
579 <entry>Identifier</entry>
580 <entry>Details</entry>
585 <entry><constant>V4L2_XFER_FUNC_DEFAULT</constant></entry>
586 <entry>Use the default transfer function as defined by the colorspace.</entry>
589 <entry><constant>V4L2_XFER_FUNC_709</constant></entry>
590 <entry>Use the Rec. 709 transfer function.</entry>
593 <entry><constant>V4L2_XFER_FUNC_SRGB</constant></entry>
594 <entry>Use the sRGB transfer function.</entry>
597 <entry><constant>V4L2_XFER_FUNC_ADOBERGB</constant></entry>
598 <entry>Use the AdobeRGB transfer function.</entry>
601 <entry><constant>V4L2_XFER_FUNC_SMPTE240M</constant></entry>
602 <entry>Use the SMPTE 240M transfer function.</entry>
605 <entry><constant>V4L2_XFER_FUNC_NONE</constant></entry>
606 <entry>Do not use a transfer function (i.e. use linear RGB values).</entry>
609 <entry><constant>V4L2_XFER_FUNC_DCI_P3</constant></entry>
610 <entry>Use the DCI-P3 transfer function.</entry>
613 <entry><constant>V4L2_XFER_FUNC_SMPTE2084</constant></entry>
614 <entry>Use the SMPTE 2084 transfer function.</entry>
620 <table pgwide="1" frame="none" id="v4l2-ycbcr-encoding">
621 <title>V4L2 Y'CbCr Encodings</title>
622 <tgroup cols="2" align="left">
626 <entry>Identifier</entry>
627 <entry>Details</entry>
632 <entry><constant>V4L2_YCBCR_ENC_DEFAULT</constant></entry>
633 <entry>Use the default Y'CbCr encoding as defined by the colorspace.</entry>
636 <entry><constant>V4L2_YCBCR_ENC_601</constant></entry>
637 <entry>Use the BT.601 Y'CbCr encoding.</entry>
640 <entry><constant>V4L2_YCBCR_ENC_709</constant></entry>
641 <entry>Use the Rec. 709 Y'CbCr encoding.</entry>
644 <entry><constant>V4L2_YCBCR_ENC_XV601</constant></entry>
645 <entry>Use the extended gamut xvYCC BT.601 encoding.</entry>
648 <entry><constant>V4L2_YCBCR_ENC_XV709</constant></entry>
649 <entry>Use the extended gamut xvYCC Rec. 709 encoding.</entry>
652 <entry><constant>V4L2_YCBCR_ENC_SYCC</constant></entry>
653 <entry>Use the extended gamut sYCC encoding.</entry>
656 <entry><constant>V4L2_YCBCR_ENC_BT2020</constant></entry>
657 <entry>Use the default non-constant luminance BT.2020 Y'CbCr encoding.</entry>
660 <entry><constant>V4L2_YCBCR_ENC_BT2020_CONST_LUM</constant></entry>
661 <entry>Use the constant luminance BT.2020 Yc'CbcCrc encoding.</entry>
667 <table pgwide="1" frame="none" id="v4l2-quantization">
668 <title>V4L2 Quantization Methods</title>
669 <tgroup cols="2" align="left">
673 <entry>Identifier</entry>
674 <entry>Details</entry>
679 <entry><constant>V4L2_QUANTIZATION_DEFAULT</constant></entry>
680 <entry>Use the default quantization encoding as defined by the colorspace.
681 This is always full range for R'G'B' (except for the BT.2020 colorspace) and usually
682 limited range for Y'CbCr.</entry>
685 <entry><constant>V4L2_QUANTIZATION_FULL_RANGE</constant></entry>
686 <entry>Use the full range quantization encoding. I.e. the range [0…1]
687 is mapped to [0…255] (with possible clipping to [1…254] to avoid the
688 0x00 and 0xff values). Cb and Cr are mapped from [-0.5…0.5] to [0…255]
689 (with possible clipping to [1…254] to avoid the 0x00 and 0xff values).</entry>
692 <entry><constant>V4L2_QUANTIZATION_LIM_RANGE</constant></entry>
693 <entry>Use the limited range quantization encoding. I.e. the range [0…1]
694 is mapped to [16…235]. Cb and Cr are mapped from [-0.5…0.5] to [16…240].
703 <title>Detailed Colorspace Descriptions</title>
704 <section id="col-smpte-170m">
705 <title>Colorspace SMPTE 170M (<constant>V4L2_COLORSPACE_SMPTE170M</constant>)</title>
706 <para>The <xref linkend="smpte170m" /> standard defines the colorspace used by NTSC and PAL and by SDTV
707 in general. The default transfer function is <constant>V4L2_XFER_FUNC_709</constant>.
708 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_601</constant>.
709 The default Y'CbCr quantization is limited range. The chromaticities of the primary colors and
710 the white reference are:</para>
712 <title>SMPTE 170M Chromaticities</title>
713 <tgroup cols="3" align="left">
739 <entry>White Reference (D65)</entry>
740 <entry>0.3127</entry>
741 <entry>0.3290</entry>
746 <para>The red, green and blue chromaticities are also often referred to
747 as the SMPTE C set, so this colorspace is sometimes called SMPTE C as well.</para>
750 <term>The transfer function defined for SMPTE 170M is the same as the
751 one defined in Rec. 709.</term>
753 <para>L' = -1.099(-L)<superscript>0.45</superscript> + 0.099 for L ≤ -0.018</para>
754 <para>L' = 4.5L for -0.018 < L < 0.018</para>
755 <para>L' = 1.099L<superscript>0.45</superscript> - 0.099 for L ≥ 0.018</para>
761 <term>Inverse Transfer function:</term>
763 <para>L = -((L' - 0.099) / -1.099)<superscript>1/0.45</superscript> for L' ≤ -0.081</para>
764 <para>L = L' / 4.5 for -0.081 < L' < 0.081</para>
765 <para>L = ((L' + 0.099) / 1.099)<superscript>1/0.45</superscript> for L' ≥ 0.081</para>
771 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with
772 the following <constant>V4L2_YCBCR_ENC_601</constant> encoding:</term>
774 <para>Y' = 0.299R' + 0.587G' + 0.114B'</para>
775 <para>Cb = -0.169R' - 0.331G' + 0.5B'</para>
776 <para>Cr = 0.5R' - 0.419G' - 0.081B'</para>
780 <para>Y' is clamped to the range [0…1] and Cb and Cr are
781 clamped to the range [-0.5…0.5]. This conversion to Y'CbCr is identical to the one
782 defined in the <xref linkend="itu601" /> standard and this colorspace is sometimes called BT.601 as well, even
783 though BT.601 does not mention any color primaries.</para>
784 <para>The default quantization is limited range, but full range is possible although
788 <section id="col-rec709">
789 <title>Colorspace Rec. 709 (<constant>V4L2_COLORSPACE_REC709</constant>)</title>
790 <para>The <xref linkend="itu709" /> standard defines the colorspace used by HDTV in general.
791 The default transfer function is <constant>V4L2_XFER_FUNC_709</constant>. The default
792 Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_709</constant>. The default Y'CbCr quantization is
793 limited range. The chromaticities of the primary colors and the white reference are:</para>
795 <title>Rec. 709 Chromaticities</title>
796 <tgroup cols="3" align="left">
822 <entry>White Reference (D65)</entry>
823 <entry>0.3127</entry>
824 <entry>0.3290</entry>
829 <para>The full name of this standard is Rec. ITU-R BT.709-5.</para>
832 <term>Transfer function. Normally L is in the range [0…1], but for the extended
833 gamut xvYCC encoding values outside that range are allowed.</term>
835 <para>L' = -1.099(-L)<superscript>0.45</superscript> + 0.099 for L ≤ -0.018</para>
836 <para>L' = 4.5L for -0.018 < L < 0.018</para>
837 <para>L' = 1.099L<superscript>0.45</superscript> - 0.099 for L ≥ 0.018</para>
843 <term>Inverse Transfer function:</term>
845 <para>L = -((L' - 0.099) / -1.099)<superscript>1/0.45</superscript> for L' ≤ -0.081</para>
846 <para>L = L' / 4.5 for -0.081 < L' < 0.081</para>
847 <para>L = ((L' + 0.099) / 1.099)<superscript>1/0.45</superscript> for L' ≥ 0.081</para>
853 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with the following
854 <constant>V4L2_YCBCR_ENC_709</constant> encoding:</term>
856 <para>Y' = 0.2126R' + 0.7152G' + 0.0722B'</para>
857 <para>Cb = -0.1146R' - 0.3854G' + 0.5B'</para>
858 <para>Cr = 0.5R' - 0.4542G' - 0.0458B'</para>
862 <para>Y' is clamped to the range [0…1] and Cb and Cr are
863 clamped to the range [-0.5…0.5].</para>
864 <para>The default quantization is limited range, but full range is possible although
866 <para>The <constant>V4L2_YCBCR_ENC_709</constant> encoding described above is the default
867 for this colorspace, but it can be overridden with <constant>V4L2_YCBCR_ENC_601</constant>, in which
868 case the BT.601 Y'CbCr encoding is used.</para>
869 <para>Two additional extended gamut Y'CbCr encodings are also possible with this colorspace:</para>
872 <term>The xvYCC 709 encoding (<constant>V4L2_YCBCR_ENC_XV709</constant>, <xref linkend="xvycc" />)
873 is similar to the Rec. 709 encoding, but it allows for R', G' and B' values that are outside the range
874 [0…1]. The resulting Y', Cb and Cr values are scaled and offset:</term>
876 <para>Y' = (219 / 256) * (0.2126R' + 0.7152G' + 0.0722B') + (16 / 256)</para>
877 <para>Cb = (224 / 256) * (-0.1146R' - 0.3854G' + 0.5B')</para>
878 <para>Cr = (224 / 256) * (0.5R' - 0.4542G' - 0.0458B')</para>
884 <term>The xvYCC 601 encoding (<constant>V4L2_YCBCR_ENC_XV601</constant>, <xref linkend="xvycc" />) is similar
885 to the BT.601 encoding, but it allows for R', G' and B' values that are outside the range
886 [0…1]. The resulting Y', Cb and Cr values are scaled and offset:</term>
888 <para>Y' = (219 / 256) * (0.299R' + 0.587G' + 0.114B') + (16 / 256)</para>
889 <para>Cb = (224 / 256) * (-0.169R' - 0.331G' + 0.5B')</para>
890 <para>Cr = (224 / 256) * (0.5R' - 0.419G' - 0.081B')</para>
894 <para>Y' is clamped to the range [0…1] and Cb and Cr are clamped
895 to the range [-0.5…0.5]. The non-standard xvYCC 709 or xvYCC 601 encodings can be used by
896 selecting <constant>V4L2_YCBCR_ENC_XV709</constant> or <constant>V4L2_YCBCR_ENC_XV601</constant>.
897 The xvYCC encodings always use full range quantization.</para>
900 <section id="col-srgb">
901 <title>Colorspace sRGB (<constant>V4L2_COLORSPACE_SRGB</constant>)</title>
902 <para>The <xref linkend="srgb" /> standard defines the colorspace used by most webcams
903 and computer graphics. The default transfer function is <constant>V4L2_XFER_FUNC_SRGB</constant>.
904 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_SYCC</constant>. The default Y'CbCr
905 quantization is full range. The chromaticities of the primary colors and the white
906 reference are:</para>
908 <title>sRGB Chromaticities</title>
909 <tgroup cols="3" align="left">
935 <entry>White Reference (D65)</entry>
936 <entry>0.3127</entry>
937 <entry>0.3290</entry>
942 <para>These chromaticities are identical to the Rec. 709 colorspace.</para>
945 <term>Transfer function. Note that negative values for L are only used by the Y'CbCr conversion.</term>
947 <para>L' = -1.055(-L)<superscript>1/2.4</superscript> + 0.055 for L < -0.0031308</para>
948 <para>L' = 12.92L for -0.0031308 ≤ L ≤ 0.0031308</para>
949 <para>L' = 1.055L<superscript>1/2.4</superscript> - 0.055 for 0.0031308 < L ≤ 1</para>
953 <term>Inverse Transfer function:</term>
955 <para>L = -((-L' + 0.055) / 1.055)<superscript>2.4</superscript> for L' < -0.04045</para>
956 <para>L = L' / 12.92 for -0.04045 ≤ L' ≤ 0.04045</para>
957 <para>L = ((L' + 0.055) / 1.055)<superscript>2.4</superscript> for L' > 0.04045</para>
963 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with the following
964 <constant>V4L2_YCBCR_ENC_SYCC</constant> encoding as defined by <xref linkend="sycc" />:</term>
966 <para>Y' = 0.2990R' + 0.5870G' + 0.1140B'</para>
967 <para>Cb = -0.1687R' - 0.3313G' + 0.5B'</para>
968 <para>Cr = 0.5R' - 0.4187G' - 0.0813B'</para>
972 <para>Y' is clamped to the range [0…1] and Cb and Cr are clamped
973 to the range [-0.5…0.5]. The <constant>V4L2_YCBCR_ENC_SYCC</constant> quantization is always
974 full range. Although this Y'CbCr encoding looks very similar to the <constant>V4L2_YCBCR_ENC_XV601</constant>
975 encoding, it is not. The <constant>V4L2_YCBCR_ENC_XV601</constant> scales and offsets the Y'CbCr
976 values before quantization, but this encoding does not do that.</para>
979 <section id="col-adobergb">
980 <title>Colorspace Adobe RGB (<constant>V4L2_COLORSPACE_ADOBERGB</constant>)</title>
981 <para>The <xref linkend="adobergb" /> standard defines the colorspace used by computer graphics
982 that use the AdobeRGB colorspace. This is also known as the <xref linkend="oprgb" /> standard.
983 The default transfer function is <constant>V4L2_XFER_FUNC_ADOBERGB</constant>.
984 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_601</constant>. The default Y'CbCr
985 quantization is limited range. The chromaticities of the primary colors and the white reference
988 <title>Adobe RGB Chromaticities</title>
989 <tgroup cols="3" align="left">
1001 <entry>0.6400</entry>
1002 <entry>0.3300</entry>
1005 <entry>Green</entry>
1006 <entry>0.2100</entry>
1007 <entry>0.7100</entry>
1011 <entry>0.1500</entry>
1012 <entry>0.0600</entry>
1015 <entry>White Reference (D65)</entry>
1016 <entry>0.3127</entry>
1017 <entry>0.3290</entry>
1024 <term>Transfer function:</term>
1026 <para>L' = L<superscript>1/2.19921875</superscript></para>
1030 <term>Inverse Transfer function:</term>
1032 <para>L = L'<superscript>2.19921875</superscript></para>
1038 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with the
1039 following <constant>V4L2_YCBCR_ENC_601</constant> encoding:</term>
1041 <para>Y' = 0.299R' + 0.587G' + 0.114B'</para>
1042 <para>Cb = -0.169R' - 0.331G' + 0.5B'</para>
1043 <para>Cr = 0.5R' - 0.419G' - 0.081B'</para>
1047 <para>Y' is clamped to the range [0…1] and Cb and Cr are
1048 clamped to the range [-0.5…0.5]. This transform is identical to one defined in
1049 SMPTE 170M/BT.601. The Y'CbCr quantization is limited range.</para>
1052 <section id="col-bt2020">
1053 <title>Colorspace BT.2020 (<constant>V4L2_COLORSPACE_BT2020</constant>)</title>
1054 <para>The <xref linkend="itu2020" /> standard defines the colorspace used by Ultra-high definition
1055 television (UHDTV). The default transfer function is <constant>V4L2_XFER_FUNC_709</constant>.
1056 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_BT2020</constant>.
1057 The default R'G'B' quantization is limited range (!), and so is the default Y'CbCr quantization.
1058 The chromaticities of the primary colors and the white reference are:</para>
1059 <table frame="none">
1060 <title>BT.2020 Chromaticities</title>
1061 <tgroup cols="3" align="left">
1065 <entry>Color</entry>
1070 <tbody valign="top">
1073 <entry>0.708</entry>
1074 <entry>0.292</entry>
1077 <entry>Green</entry>
1078 <entry>0.170</entry>
1079 <entry>0.797</entry>
1083 <entry>0.131</entry>
1084 <entry>0.046</entry>
1087 <entry>White Reference (D65)</entry>
1088 <entry>0.3127</entry>
1089 <entry>0.3290</entry>
1096 <term>Transfer function (same as Rec. 709):</term>
1098 <para>L' = 4.5L for 0 ≤ L < 0.018</para>
1099 <para>L' = 1.099L<superscript>0.45</superscript> - 0.099 for 0.018 ≤ L ≤ 1</para>
1103 <term>Inverse Transfer function:</term>
1105 <para>L = L' / 4.5 for L' < 0.081</para>
1106 <para>L = ((L' + 0.099) / 1.099)<superscript>1/0.45</superscript> for L' ≥ 0.081</para>
1112 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with the
1113 following <constant>V4L2_YCBCR_ENC_BT2020</constant> encoding:</term>
1115 <para>Y' = 0.2627R' + 0.6780G' + 0.0593B'</para>
1116 <para>Cb = -0.1396R' - 0.3604G' + 0.5B'</para>
1117 <para>Cr = 0.5R' - 0.4598G' - 0.0402B'</para>
1121 <para>Y' is clamped to the range [0…1] and Cb and Cr are
1122 clamped to the range [-0.5…0.5]. The Y'CbCr quantization is limited range.</para>
1123 <para>There is also an alternate constant luminance R'G'B' to Yc'CbcCrc
1124 (<constant>V4L2_YCBCR_ENC_BT2020_CONST_LUM</constant>) encoding:</para>
1129 <para>Yc' = (0.2627R + 0.6780G + 0.0593B)'</para>
1135 <term>B' - Yc' ≤ 0:</term>
1137 <para>Cbc = (B' - Yc') / 1.9404</para>
1143 <term>B' - Yc' > 0:</term>
1145 <para>Cbc = (B' - Yc') / 1.5816</para>
1151 <term>R' - Yc' ≤ 0:</term>
1153 <para>Crc = (R' - Y') / 1.7184</para>
1159 <term>R' - Yc' > 0:</term>
1161 <para>Crc = (R' - Y') / 0.9936</para>
1165 <para>Yc' is clamped to the range [0…1] and Cbc and Crc are
1166 clamped to the range [-0.5…0.5]. The Yc'CbcCrc quantization is limited range.</para>
1169 <section id="col-dcip3">
1170 <title>Colorspace DCI-P3 (<constant>V4L2_COLORSPACE_DCI_P3</constant>)</title>
1171 <para>The <xref linkend="smpte431" /> standard defines the colorspace used by cinema
1172 projectors that use the DCI-P3 colorspace.
1173 The default transfer function is <constant>V4L2_XFER_FUNC_DCI_P3</constant>.
1174 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_709</constant>. Note that this
1175 colorspace does not specify a Y'CbCr encoding since it is not meant to be encoded
1176 to Y'CbCr. So this default Y'CbCr encoding was picked because it is the HDTV
1177 encoding. The default Y'CbCr quantization is limited range. The chromaticities of
1178 the primary colors and the white reference are:</para>
1179 <table frame="none">
1180 <title>DCI-P3 Chromaticities</title>
1181 <tgroup cols="3" align="left">
1185 <entry>Color</entry>
1190 <tbody valign="top">
1193 <entry>0.6800</entry>
1194 <entry>0.3200</entry>
1197 <entry>Green</entry>
1198 <entry>0.2650</entry>
1199 <entry>0.6900</entry>
1203 <entry>0.1500</entry>
1204 <entry>0.0600</entry>
1207 <entry>White Reference</entry>
1208 <entry>0.3140</entry>
1209 <entry>0.3510</entry>
1216 <term>Transfer function:</term>
1218 <para>L' = L<superscript>1/2.6</superscript></para>
1222 <term>Inverse Transfer function:</term>
1224 <para>L = L'<superscript>2.6</superscript></para>
1228 <para>Y'CbCr encoding is not specified. V4L2 defaults to Rec. 709.</para>
1231 <section id="col-smpte-240m">
1232 <title>Colorspace SMPTE 240M (<constant>V4L2_COLORSPACE_SMPTE240M</constant>)</title>
1233 <para>The <xref linkend="smpte240m" /> standard was an interim standard used during
1234 the early days of HDTV (1988-1998). It has been superseded by Rec. 709.
1235 The default transfer function is <constant>V4L2_XFER_FUNC_SMPTE240M</constant>.
1236 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_SMPTE240M</constant>.
1237 The default Y'CbCr quantization is limited range. The chromaticities of the primary colors and the
1238 white reference are:</para>
1239 <table frame="none">
1240 <title>SMPTE 240M Chromaticities</title>
1241 <tgroup cols="3" align="left">
1245 <entry>Color</entry>
1250 <tbody valign="top">
1253 <entry>0.630</entry>
1254 <entry>0.340</entry>
1257 <entry>Green</entry>
1258 <entry>0.310</entry>
1259 <entry>0.595</entry>
1263 <entry>0.155</entry>
1264 <entry>0.070</entry>
1267 <entry>White Reference (D65)</entry>
1268 <entry>0.3127</entry>
1269 <entry>0.3290</entry>
1274 <para>These chromaticities are identical to the SMPTE 170M colorspace.</para>
1277 <term>Transfer function:</term>
1279 <para>L' = 4L for 0 ≤ L < 0.0228</para>
1280 <para>L' = 1.1115L<superscript>0.45</superscript> - 0.1115 for 0.0228 ≤ L ≤ 1</para>
1284 <term>Inverse Transfer function:</term>
1286 <para>L = L' / 4 for 0 ≤ L' < 0.0913</para>
1287 <para>L = ((L' + 0.1115) / 1.1115)<superscript>1/0.45</superscript> for L' ≥ 0.0913</para>
1293 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with the
1294 following <constant>V4L2_YCBCR_ENC_SMPTE240M</constant> encoding:</term>
1296 <para>Y' = 0.2122R' + 0.7013G' + 0.0865B'</para>
1297 <para>Cb = -0.1161R' - 0.3839G' + 0.5B'</para>
1298 <para>Cr = 0.5R' - 0.4451G' - 0.0549B'</para>
1302 <para>Yc' is clamped to the range [0…1] and Cbc and Crc are
1303 clamped to the range [-0.5…0.5]. The Y'CbCr quantization is limited range.</para>
1306 <section id="col-sysm">
1307 <title>Colorspace NTSC 1953 (<constant>V4L2_COLORSPACE_470_SYSTEM_M</constant>)</title>
1308 <para>This standard defines the colorspace used by NTSC in 1953. In practice this
1309 colorspace is obsolete and SMPTE 170M should be used instead.
1310 The default transfer function is <constant>V4L2_XFER_FUNC_709</constant>.
1311 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_601</constant>.
1312 The default Y'CbCr quantization is limited range.
1313 The chromaticities of the primary colors and the white reference are:</para>
1314 <table frame="none">
1315 <title>NTSC 1953 Chromaticities</title>
1316 <tgroup cols="3" align="left">
1320 <entry>Color</entry>
1325 <tbody valign="top">
1332 <entry>Green</entry>
1342 <entry>White Reference (C)</entry>
1343 <entry>0.310</entry>
1344 <entry>0.316</entry>
1349 <para>Note that this colorspace uses Illuminant C instead of D65 as the
1350 white reference. To correctly convert an image in this colorspace to another
1351 that uses D65 you need to apply a chromatic adaptation algorithm such as the
1352 Bradford method.</para>
1355 <term>The transfer function was never properly defined for NTSC 1953. The
1356 Rec. 709 transfer function is recommended in the literature:</term>
1358 <para>L' = 4.5L for 0 ≤ L < 0.018</para>
1359 <para>L' = 1.099L<superscript>0.45</superscript> - 0.099 for 0.018 ≤ L ≤ 1</para>
1363 <term>Inverse Transfer function:</term>
1365 <para>L = L' / 4.5 for L' < 0.081</para>
1366 <para>L = ((L' + 0.099) / 1.099)<superscript>1/0.45</superscript> for L' ≥ 0.081</para>
1372 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with the
1373 following <constant>V4L2_YCBCR_ENC_601</constant> encoding:</term>
1375 <para>Y' = 0.299R' + 0.587G' + 0.114B'</para>
1376 <para>Cb = -0.169R' - 0.331G' + 0.5B'</para>
1377 <para>Cr = 0.5R' - 0.419G' - 0.081B'</para>
1381 <para>Y' is clamped to the range [0…1] and Cb and Cr are
1382 clamped to the range [-0.5…0.5]. The Y'CbCr quantization is limited range.
1383 This transform is identical to one defined in SMPTE 170M/BT.601.</para>
1386 <section id="col-sysbg">
1387 <title>Colorspace EBU Tech. 3213 (<constant>V4L2_COLORSPACE_470_SYSTEM_BG</constant>)</title>
1388 <para>The <xref linkend="tech3213" /> standard defines the colorspace used by PAL/SECAM in 1975. In practice this
1389 colorspace is obsolete and SMPTE 170M should be used instead.
1390 The default transfer function is <constant>V4L2_XFER_FUNC_709</constant>.
1391 The default Y'CbCr encoding is <constant>V4L2_YCBCR_ENC_601</constant>.
1392 The default Y'CbCr quantization is limited range.
1393 The chromaticities of the primary colors and the white reference are:</para>
1394 <table frame="none">
1395 <title>EBU Tech. 3213 Chromaticities</title>
1396 <tgroup cols="3" align="left">
1400 <entry>Color</entry>
1405 <tbody valign="top">
1412 <entry>Green</entry>
1422 <entry>White Reference (D65)</entry>
1423 <entry>0.3127</entry>
1424 <entry>0.3290</entry>
1431 <term>The transfer function was never properly defined for this colorspace.
1432 The Rec. 709 transfer function is recommended in the literature:</term>
1434 <para>L' = 4.5L for 0 ≤ L < 0.018</para>
1435 <para>L' = 1.099L<superscript>0.45</superscript> - 0.099 for 0.018 ≤ L ≤ 1</para>
1439 <term>Inverse Transfer function:</term>
1441 <para>L = L' / 4.5 for L' < 0.081</para>
1442 <para>L = ((L' + 0.099) / 1.099)<superscript>1/0.45</superscript> for L' ≥ 0.081</para>
1448 <term>The luminance (Y') and color difference (Cb and Cr) are obtained with the
1449 following <constant>V4L2_YCBCR_ENC_601</constant> encoding:</term>
1451 <para>Y' = 0.299R' + 0.587G' + 0.114B'</para>
1452 <para>Cb = -0.169R' - 0.331G' + 0.5B'</para>
1453 <para>Cr = 0.5R' - 0.419G' - 0.081B'</para>
1457 <para>Y' is clamped to the range [0…1] and Cb and Cr are
1458 clamped to the range [-0.5…0.5]. The Y'CbCr quantization is limited range.
1459 This transform is identical to one defined in SMPTE 170M/BT.601.</para>
1462 <section id="col-jpeg">
1463 <title>Colorspace JPEG (<constant>V4L2_COLORSPACE_JPEG</constant>)</title>
1464 <para>This colorspace defines the colorspace used by most (Motion-)JPEG formats. The chromaticities
1465 of the primary colors and the white reference are identical to sRGB. The transfer
1466 function use is <constant>V4L2_XFER_FUNC_SRGB</constant>. The Y'CbCr encoding is
1467 <constant>V4L2_YCBCR_ENC_601</constant> with full range quantization where
1468 Y' is scaled to [0…255] and Cb/Cr are scaled to [-128…128] and
1469 then clipped to [-128…127].</para>
1470 <para>Note that the JPEG standard does not actually store colorspace information.
1471 So if something other than sRGB is used, then the driver will have to set that information
1472 explicitly. Effectively <constant>V4L2_COLORSPACE_JPEG</constant> can be considered to be
1473 an abbreviation for <constant>V4L2_COLORSPACE_SRGB</constant>, <constant>V4L2_YCBCR_ENC_601</constant>
1474 and <constant>V4L2_QUANTIZATION_FULL_RANGE</constant>.</para>
1480 <title>Detailed Transfer Function Descriptions</title>
1481 <section id="xf-smpte-2084">
1482 <title>Transfer Function SMPTE 2084 (<constant>V4L2_XFER_FUNC_SMPTE2084</constant>)</title>
1483 <para>The <xref linkend="smpte2084" /> standard defines the transfer function used by
1484 High Dynamic Range content.</para>
1487 <term>Constants:</term>
1489 <para>m1 = (2610 / 4096) / 4</para>
1490 <para>m2 = (2523 / 4096) * 128</para>
1491 <para>c1 = 3424 / 4096</para>
1492 <para>c2 = (2413 / 4096) * 32</para>
1493 <para>c3 = (2392 / 4096) * 32</para>
1497 <term>Transfer function:</term>
1499 <para>L' = ((c1 + c2 * L<superscript>m1</superscript>) / (1 + c3 * L<superscript>m1</superscript>))<superscript>m2</superscript></para>
1505 <term>Inverse Transfer function:</term>
1507 <para>L = (max(L'<superscript>1/m2</superscript> - c1, 0) / (c2 - c3 * L'<superscript>1/m2</superscript>))<superscript>1/m1</superscript></para>
1514 <section id="pixfmt-indexed">
1515 <title>Indexed Format</title>
1517 <para>In this format each pixel is represented by an 8 bit index
1518 into a 256 entry ARGB palette. It is intended for <link
1519 linkend="osd">Video Output Overlays</link> only. There are no ioctls to
1520 access the palette, this must be done with ioctls of the Linux framebuffer API.</para>
1522 <table pgwide="0" frame="none">
1523 <title>Indexed Image Format</title>
1524 <tgroup cols="37" align="center">
1525 <colspec colname="id" align="left" />
1526 <colspec colname="fourcc" />
1527 <colspec colname="bit" />
1529 <colspec colnum="4" colname="b07" align="center" />
1530 <colspec colnum="5" colname="b06" align="center" />
1531 <colspec colnum="6" colname="b05" align="center" />
1532 <colspec colnum="7" colname="b04" align="center" />
1533 <colspec colnum="8" colname="b03" align="center" />
1534 <colspec colnum="9" colname="b02" align="center" />
1535 <colspec colnum="10" colname="b01" align="center" />
1536 <colspec colnum="11" colname="b00" align="center" />
1538 <spanspec namest="b07" nameend="b00" spanname="b0" />
1539 <spanspec namest="b17" nameend="b10" spanname="b1" />
1540 <spanspec namest="b27" nameend="b20" spanname="b2" />
1541 <spanspec namest="b37" nameend="b30" spanname="b3" />
1544 <entry>Identifier</entry>
1546 <entry> </entry>
1547 <entry spanname="b0">Byte 0</entry>
1550 <entry> </entry>
1551 <entry> </entry>
1563 <tbody valign="top">
1564 <row id="V4L2-PIX-FMT-PAL8">
1565 <entry><constant>V4L2_PIX_FMT_PAL8</constant></entry>
1566 <entry>'PAL8'</entry>
1568 <entry>i<subscript>7</subscript></entry>
1569 <entry>i<subscript>6</subscript></entry>
1570 <entry>i<subscript>5</subscript></entry>
1571 <entry>i<subscript>4</subscript></entry>
1572 <entry>i<subscript>3</subscript></entry>
1573 <entry>i<subscript>2</subscript></entry>
1574 <entry>i<subscript>1</subscript></entry>
1575 <entry>i<subscript>0</subscript></entry>
1582 <section id="pixfmt-rgb">
1583 <title>RGB Formats</title>
1598 <section id="yuv-formats">
1599 <title>YUV Formats</title>
1601 <para>YUV is the format native to TV broadcast and composite video
1602 signals. It separates the brightness information (Y) from the color
1603 information (U and V or Cb and Cr). The color information consists of
1604 red and blue <emphasis>color difference</emphasis> signals, this way
1605 the green component can be reconstructed by subtracting from the
1606 brightness component. See <xref linkend="colorspaces" /> for conversion
1607 examples. YUV was chosen because early television would only transmit
1608 brightness information. To add color in a way compatible with existing
1609 receivers a new signal carrier was added to transmit the color
1610 difference signals. Secondary in the YUV format the U and V components
1611 usually have lower resolution than the Y component. This is an analog
1612 video compression technique taking advantage of a property of the
1613 human visual system, being more sensitive to brightness
1645 <title>Compressed Formats</title>
1647 <table pgwide="1" frame="none" id="compressed-formats">
1648 <title>Compressed Image Formats</title>
1649 <tgroup cols="3" align="left">
1653 <entry>Identifier</entry>
1655 <entry>Details</entry>
1658 <tbody valign="top">
1659 <row id="V4L2-PIX-FMT-JPEG">
1660 <entry><constant>V4L2_PIX_FMT_JPEG</constant></entry>
1661 <entry>'JPEG'</entry>
1662 <entry>TBD. See also &VIDIOC-G-JPEGCOMP;,
1663 &VIDIOC-S-JPEGCOMP;.</entry>
1665 <row id="V4L2-PIX-FMT-MPEG">
1666 <entry><constant>V4L2_PIX_FMT_MPEG</constant></entry>
1667 <entry>'MPEG'</entry>
1668 <entry>MPEG multiplexed stream. The actual format is determined by
1669 extended control <constant>V4L2_CID_MPEG_STREAM_TYPE</constant>, see
1670 <xref linkend="mpeg-control-id" />.</entry>
1672 <row id="V4L2-PIX-FMT-H264">
1673 <entry><constant>V4L2_PIX_FMT_H264</constant></entry>
1674 <entry>'H264'</entry>
1675 <entry>H264 video elementary stream with start codes.</entry>
1677 <row id="V4L2-PIX-FMT-H264-NO-SC">
1678 <entry><constant>V4L2_PIX_FMT_H264_NO_SC</constant></entry>
1679 <entry>'AVC1'</entry>
1680 <entry>H264 video elementary stream without start codes.</entry>
1682 <row id="V4L2-PIX-FMT-H264-MVC">
1683 <entry><constant>V4L2_PIX_FMT_H264_MVC</constant></entry>
1684 <entry>'M264'</entry>
1685 <entry>H264 MVC video elementary stream.</entry>
1687 <row id="V4L2-PIX-FMT-H263">
1688 <entry><constant>V4L2_PIX_FMT_H263</constant></entry>
1689 <entry>'H263'</entry>
1690 <entry>H263 video elementary stream.</entry>
1692 <row id="V4L2-PIX-FMT-MPEG1">
1693 <entry><constant>V4L2_PIX_FMT_MPEG1</constant></entry>
1694 <entry>'MPG1'</entry>
1695 <entry>MPEG1 video elementary stream.</entry>
1697 <row id="V4L2-PIX-FMT-MPEG2">
1698 <entry><constant>V4L2_PIX_FMT_MPEG2</constant></entry>
1699 <entry>'MPG2'</entry>
1700 <entry>MPEG2 video elementary stream.</entry>
1702 <row id="V4L2-PIX-FMT-MPEG4">
1703 <entry><constant>V4L2_PIX_FMT_MPEG4</constant></entry>
1704 <entry>'MPG4'</entry>
1705 <entry>MPEG4 video elementary stream.</entry>
1707 <row id="V4L2-PIX-FMT-XVID">
1708 <entry><constant>V4L2_PIX_FMT_XVID</constant></entry>
1709 <entry>'XVID'</entry>
1710 <entry>Xvid video elementary stream.</entry>
1712 <row id="V4L2-PIX-FMT-VC1-ANNEX-G">
1713 <entry><constant>V4L2_PIX_FMT_VC1_ANNEX_G</constant></entry>
1714 <entry>'VC1G'</entry>
1715 <entry>VC1, SMPTE 421M Annex G compliant stream.</entry>
1717 <row id="V4L2-PIX-FMT-VC1-ANNEX-L">
1718 <entry><constant>V4L2_PIX_FMT_VC1_ANNEX_L</constant></entry>
1719 <entry>'VC1L'</entry>
1720 <entry>VC1, SMPTE 421M Annex L compliant stream.</entry>
1722 <row id="V4L2-PIX-FMT-VP8">
1723 <entry><constant>V4L2_PIX_FMT_VP8</constant></entry>
1724 <entry>'VP80'</entry>
1725 <entry>VP8 video elementary stream.</entry>
1732 <section id="sdr-formats">
1733 <title>SDR Formats</title>
1735 <para>These formats are used for <link linkend="sdr">SDR</link>
1736 interface only.</para>
1746 <section id="pixfmt-reserved">
1747 <title>Reserved Format Identifiers</title>
1749 <para>These formats are not defined by this specification, they
1750 are just listed for reference and to avoid naming conflicts. If you
1751 want to register your own format, send an e-mail to the linux-media mailing
1752 list &v4l-ml; for inclusion in the <filename>videodev2.h</filename>
1753 file. If you want to share your format with other developers add a
1754 link to your documentation and send a copy to the linux-media mailing list
1755 for inclusion in this section. If you think your format should be listed
1756 in a standard format section please make a proposal on the linux-media mailing
1759 <table pgwide="1" frame="none" id="reserved-formats">
1760 <title>Reserved Image Formats</title>
1761 <tgroup cols="3" align="left">
1765 <entry>Identifier</entry>
1767 <entry>Details</entry>
1770 <tbody valign="top">
1771 <row id="V4L2-PIX-FMT-DV">
1772 <entry><constant>V4L2_PIX_FMT_DV</constant></entry>
1773 <entry>'dvsd'</entry>
1774 <entry>unknown</entry>
1776 <row id="V4L2-PIX-FMT-ET61X251">
1777 <entry><constant>V4L2_PIX_FMT_ET61X251</constant></entry>
1778 <entry>'E625'</entry>
1779 <entry>Compressed format of the ET61X251 driver.</entry>
1781 <row id="V4L2-PIX-FMT-HI240">
1782 <entry><constant>V4L2_PIX_FMT_HI240</constant></entry>
1783 <entry>'HI24'</entry>
1784 <entry><para>8 bit RGB format used by the BTTV driver.</para></entry>
1786 <row id="V4L2-PIX-FMT-HM12">
1787 <entry><constant>V4L2_PIX_FMT_HM12</constant></entry>
1788 <entry>'HM12'</entry>
1789 <entry><para>YUV 4:2:0 format used by the
1790 IVTV driver, <ulink url="http://www.ivtvdriver.org/">
1791 http://www.ivtvdriver.org/</ulink></para><para>The format is documented in the
1792 kernel sources in the file <filename>Documentation/video4linux/cx2341x/README.hm12</filename>
1795 <row id="V4L2-PIX-FMT-CPIA1">
1796 <entry><constant>V4L2_PIX_FMT_CPIA1</constant></entry>
1797 <entry>'CPIA'</entry>
1798 <entry>YUV format used by the gspca cpia1 driver.</entry>
1800 <row id="V4L2-PIX-FMT-JPGL">
1801 <entry><constant>V4L2_PIX_FMT_JPGL</constant></entry>
1802 <entry>'JPGL'</entry>
1803 <entry>JPEG-Light format (Pegasus Lossless JPEG)
1804 used in Divio webcams NW 80x.</entry>
1806 <row id="V4L2-PIX-FMT-SPCA501">
1807 <entry><constant>V4L2_PIX_FMT_SPCA501</constant></entry>
1808 <entry>'S501'</entry>
1809 <entry>YUYV per line used by the gspca driver.</entry>
1811 <row id="V4L2-PIX-FMT-SPCA505">
1812 <entry><constant>V4L2_PIX_FMT_SPCA505</constant></entry>
1813 <entry>'S505'</entry>
1814 <entry>YYUV per line used by the gspca driver.</entry>
1816 <row id="V4L2-PIX-FMT-SPCA508">
1817 <entry><constant>V4L2_PIX_FMT_SPCA508</constant></entry>
1818 <entry>'S508'</entry>
1819 <entry>YUVY per line used by the gspca driver.</entry>
1821 <row id="V4L2-PIX-FMT-SPCA561">
1822 <entry><constant>V4L2_PIX_FMT_SPCA561</constant></entry>
1823 <entry>'S561'</entry>
1824 <entry>Compressed GBRG Bayer format used by the gspca driver.</entry>
1826 <row id="V4L2-PIX-FMT-PAC207">
1827 <entry><constant>V4L2_PIX_FMT_PAC207</constant></entry>
1828 <entry>'P207'</entry>
1829 <entry>Compressed BGGR Bayer format used by the gspca driver.</entry>
1831 <row id="V4L2-PIX-FMT-MR97310A">
1832 <entry><constant>V4L2_PIX_FMT_MR97310A</constant></entry>
1833 <entry>'M310'</entry>
1834 <entry>Compressed BGGR Bayer format used by the gspca driver.</entry>
1836 <row id="V4L2-PIX-FMT-JL2005BCD">
1837 <entry><constant>V4L2_PIX_FMT_JL2005BCD</constant></entry>
1838 <entry>'JL20'</entry>
1839 <entry>JPEG compressed RGGB Bayer format used by the gspca driver.</entry>
1841 <row id="V4L2-PIX-FMT-OV511">
1842 <entry><constant>V4L2_PIX_FMT_OV511</constant></entry>
1843 <entry>'O511'</entry>
1844 <entry>OV511 JPEG format used by the gspca driver.</entry>
1846 <row id="V4L2-PIX-FMT-OV518">
1847 <entry><constant>V4L2_PIX_FMT_OV518</constant></entry>
1848 <entry>'O518'</entry>
1849 <entry>OV518 JPEG format used by the gspca driver.</entry>
1851 <row id="V4L2-PIX-FMT-PJPG">
1852 <entry><constant>V4L2_PIX_FMT_PJPG</constant></entry>
1853 <entry>'PJPG'</entry>
1854 <entry>Pixart 73xx JPEG format used by the gspca driver.</entry>
1856 <row id="V4L2-PIX-FMT-SE401">
1857 <entry><constant>V4L2_PIX_FMT_SE401</constant></entry>
1858 <entry>'S401'</entry>
1859 <entry>Compressed RGB format used by the gspca se401 driver</entry>
1861 <row id="V4L2-PIX-FMT-SQ905C">
1862 <entry><constant>V4L2_PIX_FMT_SQ905C</constant></entry>
1863 <entry>'905C'</entry>
1864 <entry>Compressed RGGB bayer format used by the gspca driver.</entry>
1866 <row id="V4L2-PIX-FMT-MJPEG">
1867 <entry><constant>V4L2_PIX_FMT_MJPEG</constant></entry>
1868 <entry>'MJPG'</entry>
1869 <entry>Compressed format used by the Zoran driver</entry>
1871 <row id="V4L2-PIX-FMT-PWC1">
1872 <entry><constant>V4L2_PIX_FMT_PWC1</constant></entry>
1873 <entry>'PWC1'</entry>
1874 <entry>Compressed format of the PWC driver.</entry>
1876 <row id="V4L2-PIX-FMT-PWC2">
1877 <entry><constant>V4L2_PIX_FMT_PWC2</constant></entry>
1878 <entry>'PWC2'</entry>
1879 <entry>Compressed format of the PWC driver.</entry>
1881 <row id="V4L2-PIX-FMT-SN9C10X">
1882 <entry><constant>V4L2_PIX_FMT_SN9C10X</constant></entry>
1883 <entry>'S910'</entry>
1884 <entry>Compressed format of the SN9C102 driver.</entry>
1886 <row id="V4L2-PIX-FMT-SN9C20X-I420">
1887 <entry><constant>V4L2_PIX_FMT_SN9C20X_I420</constant></entry>
1888 <entry>'S920'</entry>
1889 <entry>YUV 4:2:0 format of the gspca sn9c20x driver.</entry>
1891 <row id="V4L2-PIX-FMT-SN9C2028">
1892 <entry><constant>V4L2_PIX_FMT_SN9C2028</constant></entry>
1893 <entry>'SONX'</entry>
1894 <entry>Compressed GBRG bayer format of the gspca sn9c2028 driver.</entry>
1896 <row id="V4L2-PIX-FMT-STV0680">
1897 <entry><constant>V4L2_PIX_FMT_STV0680</constant></entry>
1898 <entry>'S680'</entry>
1899 <entry>Bayer format of the gspca stv0680 driver.</entry>
1901 <row id="V4L2-PIX-FMT-WNVA">
1902 <entry><constant>V4L2_PIX_FMT_WNVA</constant></entry>
1903 <entry>'WNVA'</entry>
1904 <entry><para>Used by the Winnov Videum driver, <ulink
1905 url="http://www.thedirks.org/winnov/">
1906 http://www.thedirks.org/winnov/</ulink></para></entry>
1908 <row id="V4L2-PIX-FMT-TM6000">
1909 <entry><constant>V4L2_PIX_FMT_TM6000</constant></entry>
1910 <entry>'TM60'</entry>
1911 <entry><para>Used by Trident tm6000</para></entry>
1913 <row id="V4L2-PIX-FMT-CIT-YYVYUY">
1914 <entry><constant>V4L2_PIX_FMT_CIT_YYVYUY</constant></entry>
1915 <entry>'CITV'</entry>
1916 <entry><para>Used by xirlink CIT, found at IBM webcams.</para>
1917 <para>Uses one line of Y then 1 line of VYUY</para>
1920 <row id="V4L2-PIX-FMT-KONICA420">
1921 <entry><constant>V4L2_PIX_FMT_KONICA420</constant></entry>
1922 <entry>'KONI'</entry>
1923 <entry><para>Used by Konica webcams.</para>
1924 <para>YUV420 planar in blocks of 256 pixels.</para>
1927 <row id="V4L2-PIX-FMT-YYUV">
1928 <entry><constant>V4L2_PIX_FMT_YYUV</constant></entry>
1929 <entry>'YYUV'</entry>
1930 <entry>unknown</entry>
1932 <row id="V4L2-PIX-FMT-Y4">
1933 <entry><constant>V4L2_PIX_FMT_Y4</constant></entry>
1934 <entry>'Y04 '</entry>
1935 <entry>Old 4-bit greyscale format. Only the most significant 4 bits of each byte are used,
1936 the other bits are set to 0.</entry>
1938 <row id="V4L2-PIX-FMT-Y6">
1939 <entry><constant>V4L2_PIX_FMT_Y6</constant></entry>
1940 <entry>'Y06 '</entry>
1941 <entry>Old 6-bit greyscale format. Only the most significant 6 bits of each byte are used,
1942 the other bits are set to 0.</entry>
1944 <row id="V4L2-PIX-FMT-S5C-UYVY-JPG">
1945 <entry><constant>V4L2_PIX_FMT_S5C_UYVY_JPG</constant></entry>
1946 <entry>'S5CI'</entry>
1947 <entry>Two-planar format used by Samsung S5C73MX cameras. The
1948 first plane contains interleaved JPEG and UYVY image data, followed by meta data
1949 in form of an array of offsets to the UYVY data blocks. The actual pointer array
1950 follows immediately the interleaved JPEG/UYVY data, the number of entries in
1951 this array equals the height of the UYVY image. Each entry is a 4-byte unsigned
1952 integer in big endian order and it's an offset to a single pixel line of the
1953 UYVY image. The first plane can start either with JPEG or UYVY data chunk. The
1954 size of a single UYVY block equals the UYVY image's width multiplied by 2. The
1955 size of a JPEG chunk depends on the image and can vary with each line.
1956 <para>The second plane, at an offset of 4084 bytes, contains a 4-byte offset to
1957 the pointer array in the first plane. This offset is followed by a 4-byte value
1958 indicating size of the pointer array. All numbers in the second plane are also
1959 in big endian order. Remaining data in the second plane is undefined. The
1960 information in the second plane allows to easily find location of the pointer
1961 array, which can be different for each frame. The size of the pointer array is
1962 constant for given UYVY image height.</para>
1963 <para>In order to extract UYVY and JPEG frames an application can initially set
1964 a data pointer to the start of first plane and then add an offset from the first
1965 entry of the pointers table. Such a pointer indicates start of an UYVY image
1966 pixel line. Whole UYVY line can be copied to a separate buffer. These steps
1967 should be repeated for each line, i.e. the number of entries in the pointer
1968 array. Anything what's in between the UYVY lines is JPEG data and should be
1969 concatenated to form the JPEG stream. </para>
1976 <table frame="none" pgwide="1" id="format-flags">
1977 <title>Format Flags</title>
1980 <tbody valign="top">
1982 <entry><constant>V4L2_PIX_FMT_FLAG_PREMUL_ALPHA</constant></entry>
1983 <entry>0x00000001</entry>
1984 <entry>The color values are premultiplied by the alpha channel
1985 value. For example, if a light blue pixel with 50% transparency was described by
1986 RGBA values (128, 192, 255, 128), the same pixel described with premultiplied
1987 colors would be described by RGBA values (64, 96, 128, 128) </entry>
projects / karo-tx-linux.git / blob