1 /*
2 * libjingle
3 * Copyright 2004 Google Inc.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions are met:
7 *
8 * 1. Redistributions of source code must retain the above copyright notice,
9 * this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright notice,
11 * this list of conditions and the following disclaimer in the documentation
12 * and/or other materials provided with the distribution.
13 * 3. The name of the author may not be used to endorse or promote products
14 * derived from this software without specific prior written permission.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED
17 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
18 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
19 * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
20 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
21 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
22 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
23 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
24 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
25 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26 */
27
28 // Common definition for video, including fourcc and VideoFormat.
29
30 #ifndef TALK_MEDIA_BASE_VIDEOCOMMON_H_ // NOLINT
31 #define TALK_MEDIA_BASE_VIDEOCOMMON_H_
32
33 #include <string>
34
35 #include "webrtc/base/basictypes.h"
36 #include "webrtc/base/timeutils.h"
37
38 namespace cricket {
39
40 // TODO(janahan): For now, a hard-coded ssrc is used as the video ssrc.
41 // This is because when the video frame is passed to the mediaprocessor for
42 // processing, it doesn't have the correct ssrc. Since currently only Tx
43 // Video processing is supported, this is ok. When we switch over to trigger
44 // from capturer, this should be fixed and this const removed.
45 const uint32_t kDummyVideoSsrc = 0xFFFFFFFF;
46
47 // Minimum interval is 10k fps.
48 #define FPS_TO_INTERVAL(fps) \
49 (fps ? rtc::kNumNanosecsPerSec / fps : \
50 rtc::kNumNanosecsPerSec / 10000)
51
52 //////////////////////////////////////////////////////////////////////////////
53 // Definition of FourCC codes
54 //////////////////////////////////////////////////////////////////////////////
55 // Convert four characters to a FourCC code.
56 // Needs to be a macro otherwise the OS X compiler complains when the kFormat*
57 // constants are used in a switch.
58 #define FOURCC(a, b, c, d) \
59 ((static_cast<uint32_t>(a)) | (static_cast<uint32_t>(b) << 8) | \
60 (static_cast<uint32_t>(c) << 16) | (static_cast<uint32_t>(d) << 24))
61 // Some pages discussing FourCC codes:
62 // http://www.fourcc.org/yuv.php
63 // http://v4l2spec.bytesex.org/spec/book1.htm
64 // http://developer.apple.com/quicktime/icefloe/dispatch020.html
65 // http://msdn.microsoft.com/library/windows/desktop/dd206750.aspx#nv12
66 // http://people.xiph.org/~xiphmont/containers/nut/nut4cc.txt
67
68 // FourCC codes grouped according to implementation efficiency.
69 // Primary formats should convert in 1 efficient step.
70 // Secondary formats are converted in 2 steps.
71 // Auxilliary formats call primary converters.
72 enum FourCC {
73 // 9 Primary YUV formats: 5 planar, 2 biplanar, 2 packed.
74 FOURCC_I420 = FOURCC('I', '4', '2', '0'),
75 FOURCC_I422 = FOURCC('I', '4', '2', '2'),
76 FOURCC_I444 = FOURCC('I', '4', '4', '4'),
77 FOURCC_I411 = FOURCC('I', '4', '1', '1'),
78 FOURCC_I400 = FOURCC('I', '4', '0', '0'),
79 FOURCC_NV21 = FOURCC('N', 'V', '2', '1'),
80 FOURCC_NV12 = FOURCC('N', 'V', '1', '2'),
81 FOURCC_YUY2 = FOURCC('Y', 'U', 'Y', '2'),
82 FOURCC_UYVY = FOURCC('U', 'Y', 'V', 'Y'),
83
84 // 2 Secondary YUV formats: row biplanar.
85 FOURCC_M420 = FOURCC('M', '4', '2', '0'),
86
87 // 9 Primary RGB formats: 4 32 bpp, 2 24 bpp, 3 16 bpp.
88 FOURCC_ARGB = FOURCC('A', 'R', 'G', 'B'),
89 FOURCC_BGRA = FOURCC('B', 'G', 'R', 'A'),
90 FOURCC_ABGR = FOURCC('A', 'B', 'G', 'R'),
91 FOURCC_24BG = FOURCC('2', '4', 'B', 'G'),
92 FOURCC_RAW = FOURCC('r', 'a', 'w', ' '),
93 FOURCC_RGBA = FOURCC('R', 'G', 'B', 'A'),
94 FOURCC_RGBP = FOURCC('R', 'G', 'B', 'P'), // bgr565.
95 FOURCC_RGBO = FOURCC('R', 'G', 'B', 'O'), // abgr1555.
96 FOURCC_R444 = FOURCC('R', '4', '4', '4'), // argb4444.
97
98 // 4 Secondary RGB formats: 4 Bayer Patterns.
99 FOURCC_RGGB = FOURCC('R', 'G', 'G', 'B'),
100 FOURCC_BGGR = FOURCC('B', 'G', 'G', 'R'),
101 FOURCC_GRBG = FOURCC('G', 'R', 'B', 'G'),
102 FOURCC_GBRG = FOURCC('G', 'B', 'R', 'G'),
103
104 // 1 Primary Compressed YUV format.
105 FOURCC_MJPG = FOURCC('M', 'J', 'P', 'G'),
106
107 // 5 Auxiliary YUV variations: 3 with U and V planes are swapped, 1 Alias.
108 FOURCC_YV12 = FOURCC('Y', 'V', '1', '2'),
109 FOURCC_YV16 = FOURCC('Y', 'V', '1', '6'),
110 FOURCC_YV24 = FOURCC('Y', 'V', '2', '4'),
111 FOURCC_YU12 = FOURCC('Y', 'U', '1', '2'), // Linux version of I420.
112 FOURCC_J420 = FOURCC('J', '4', '2', '0'),
113 FOURCC_J400 = FOURCC('J', '4', '0', '0'),
114
115 // 14 Auxiliary aliases. CanonicalFourCC() maps these to canonical fourcc.
116 FOURCC_IYUV = FOURCC('I', 'Y', 'U', 'V'), // Alias for I420.
117 FOURCC_YU16 = FOURCC('Y', 'U', '1', '6'), // Alias for I422.
118 FOURCC_YU24 = FOURCC('Y', 'U', '2', '4'), // Alias for I444.
119 FOURCC_YUYV = FOURCC('Y', 'U', 'Y', 'V'), // Alias for YUY2.
120 FOURCC_YUVS = FOURCC('y', 'u', 'v', 's'), // Alias for YUY2 on Mac.
121 FOURCC_HDYC = FOURCC('H', 'D', 'Y', 'C'), // Alias for UYVY.
122 FOURCC_2VUY = FOURCC('2', 'v', 'u', 'y'), // Alias for UYVY on Mac.
123 FOURCC_JPEG = FOURCC('J', 'P', 'E', 'G'), // Alias for MJPG.
124 FOURCC_DMB1 = FOURCC('d', 'm', 'b', '1'), // Alias for MJPG on Mac.
125 FOURCC_BA81 = FOURCC('B', 'A', '8', '1'), // Alias for BGGR.
126 FOURCC_RGB3 = FOURCC('R', 'G', 'B', '3'), // Alias for RAW.
127 FOURCC_BGR3 = FOURCC('B', 'G', 'R', '3'), // Alias for 24BG.
128 FOURCC_CM32 = FOURCC(0, 0, 0, 32), // Alias for BGRA kCMPixelFormat_32ARGB
129 FOURCC_CM24 = FOURCC(0, 0, 0, 24), // Alias for RAW kCMPixelFormat_24RGB
130
131 // 1 Auxiliary compressed YUV format set aside for capturer.
132 FOURCC_H264 = FOURCC('H', '2', '6', '4'),
133 };
134
135 // Match any fourcc.
136
137 // We move this out of the enum because using it in many places caused
138 // the compiler to get grumpy, presumably since the above enum is
139 // backed by an int.
140 static const uint32_t FOURCC_ANY = 0xFFFFFFFF;
141
142 // Converts fourcc aliases into canonical ones.
143 uint32_t CanonicalFourCC(uint32_t fourcc);
144
145 // Get FourCC code as a string.
GetFourccName(uint32_t fourcc)146 inline std::string GetFourccName(uint32_t fourcc) {
147 std::string name;
148 name.push_back(static_cast<char>(fourcc & 0xFF));
149 name.push_back(static_cast<char>((fourcc >> 8) & 0xFF));
150 name.push_back(static_cast<char>((fourcc >> 16) & 0xFF));
151 name.push_back(static_cast<char>((fourcc >> 24) & 0xFF));
152 return name;
153 }
154
155 // Computes a scale less to fit in max_pixels while maintaining aspect ratio.
156 void ComputeScaleMaxPixels(int frame_width, int frame_height, int max_pixels,
157 int* scaled_width, int* scaled_height);
158
159 // For low fps, max pixels limit is set to Retina MacBookPro 15" resolution of
160 // 2880 x 1800 as of 4/18/2013.
161 // For high fps, maximum pixels limit is set based on common 24" monitor
162 // resolution of 2048 x 1280 as of 6/13/2013. The Retina resolution is
163 // therefore reduced to 1440 x 900.
164 void ComputeScale(int frame_width, int frame_height, int fps,
165 int* scaled_width, int* scaled_height);
166
167 // Compute the frame size that conversion should crop to based on aspect ratio.
168 // Ensures size is multiple of 2 due to I420 and conversion limitations.
169 void ComputeCrop(int cropped_format_width, int cropped_format_height,
170 int frame_width, int frame_height,
171 int pixel_width, int pixel_height,
172 int rotation,
173 int* cropped_width, int* cropped_height);
174
175 // Compute the frame size that makes pixels square pixel aspect ratio.
176 void ComputeScaleToSquarePixels(int in_width, int in_height,
177 int pixel_width, int pixel_height,
178 int* scaled_width, int* scaled_height);
179
180 //////////////////////////////////////////////////////////////////////////////
181 // Definition of VideoFormat.
182 //////////////////////////////////////////////////////////////////////////////
183
184 // VideoFormat with Plain Old Data for global variables.
185 struct VideoFormatPod {
186 int width; // Number of pixels.
187 int height; // Number of pixels.
188 int64_t interval; // Nanoseconds.
189 uint32_t fourcc; // Color space. FOURCC_ANY means that any color space is OK.
190 };
191
192 struct VideoFormat : VideoFormatPod {
193 static const int64_t kMinimumInterval =
194 rtc::kNumNanosecsPerSec / 10000; // 10k fps.
195
VideoFormatVideoFormat196 VideoFormat() {
197 Construct(0, 0, 0, 0);
198 }
199
VideoFormatVideoFormat200 VideoFormat(int w, int h, int64_t interval_ns, uint32_t cc) {
201 Construct(w, h, interval_ns, cc);
202 }
203
VideoFormatVideoFormat204 explicit VideoFormat(const VideoFormatPod& format) {
205 Construct(format.width, format.height, format.interval, format.fourcc);
206 }
207
ConstructVideoFormat208 void Construct(int w, int h, int64_t interval_ns, uint32_t cc) {
209 width = w;
210 height = h;
211 interval = interval_ns;
212 fourcc = cc;
213 }
214
FpsToIntervalVideoFormat215 static int64_t FpsToInterval(int fps) {
216 return fps ? rtc::kNumNanosecsPerSec / fps : kMinimumInterval;
217 }
218
IntervalToFpsVideoFormat219 static int IntervalToFps(int64_t interval) {
220 if (!interval) {
221 return 0;
222 }
223 return static_cast<int>(rtc::kNumNanosecsPerSec / interval);
224 }
225
IntervalToFpsFloatVideoFormat226 static float IntervalToFpsFloat(int64_t interval) {
227 if (!interval) {
228 return 0.f;
229 }
230 return static_cast<float>(rtc::kNumNanosecsPerSec) /
231 static_cast<float>(interval);
232 }
233
234 bool operator==(const VideoFormat& format) const {
235 return width == format.width && height == format.height &&
236 interval == format.interval && fourcc == format.fourcc;
237 }
238
239 bool operator!=(const VideoFormat& format) const {
240 return !(*this == format);
241 }
242
243 bool operator<(const VideoFormat& format) const {
244 return (fourcc < format.fourcc) ||
245 (fourcc == format.fourcc && width < format.width) ||
246 (fourcc == format.fourcc && width == format.width &&
247 height < format.height) ||
248 (fourcc == format.fourcc && width == format.width &&
249 height == format.height && interval > format.interval);
250 }
251
framerateVideoFormat252 int framerate() const { return IntervalToFps(interval); }
253
254 // Check if both width and height are 0.
IsSize0x0VideoFormat255 bool IsSize0x0() const { return 0 == width && 0 == height; }
256
257 // Check if this format is less than another one by comparing the resolution
258 // and frame rate.
IsPixelRateLessVideoFormat259 bool IsPixelRateLess(const VideoFormat& format) const {
260 return width * height * framerate() <
261 format.width * format.height * format.framerate();
262 }
263
264 // Get a string presentation in the form of "fourcc width x height x fps"
265 std::string ToString() const;
266 };
267
268 } // namespace cricket
269
270 #endif // TALK_MEDIA_BASE_VIDEOCOMMON_H_ // NOLINT
271