1 /* 2 * Copyright (C) 2010 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 * See the License for the specific language governing permissions and 14 * limitations under the License. 15 */ 16 17 //#define LOG_NDEBUG 0 18 #define LOG_TAG "ESQueue" 19 #include <media/stagefright/foundation/ADebug.h> 20 21 #include "ESQueue.h" 22 23 #include <media/stagefright/foundation/hexdump.h> 24 #include <media/stagefright/foundation/ABitReader.h> 25 #include <media/stagefright/foundation/ABuffer.h> 26 #include <media/stagefright/foundation/AMessage.h> 27 #include <media/stagefright/MediaErrors.h> 28 #include <media/stagefright/MediaDefs.h> 29 #include <media/stagefright/MetaData.h> 30 #include <media/stagefright/Utils.h> 31 #include <media/cas/DescramblerAPI.h> 32 #include <media/hardware/CryptoAPI.h> 33 34 #include "include/avc_utils.h" 35 36 #include <inttypes.h> 37 #include <netinet/in.h> 38 39 namespace android { 40 41 ElementaryStreamQueue::ElementaryStreamQueue(Mode mode, uint32_t flags) 42 : mMode(mode), 43 mFlags(flags), 44 mEOSReached(false), 45 mCASystemId(0), 46 mAUIndex(0) { 47 48 ALOGV("ElementaryStreamQueue(%p) mode %x flags %x isScrambled %d isSampleEncrypted %d", 49 this, mode, flags, isScrambled(), isSampleEncrypted()); 50 51 // Create the decryptor anyway since we don't know the use-case unless key is provided 52 // Won't decrypt if key info not available (e.g., scanner/extractor just parsing ts files) 53 mSampleDecryptor = isSampleEncrypted() ? new HlsSampleDecryptor : NULL; 54 } 55 56 sp<MetaData> ElementaryStreamQueue::getFormat() { 57 return mFormat; 58 } 59 60 void ElementaryStreamQueue::clear(bool clearFormat) { 61 if (mBuffer != NULL) { 62 mBuffer->setRange(0, 0); 63 } 64 65 mRangeInfos.clear(); 66 67 if (mScrambledBuffer != NULL) { 68 mScrambledBuffer->setRange(0, 0); 69 } 70 mScrambledRangeInfos.clear(); 71 72 if (clearFormat) { 73 mFormat.clear(); 74 } 75 76 mEOSReached = false; 77 } 78 79 bool ElementaryStreamQueue::isScrambled() const { 80 return (mFlags & kFlag_ScrambledData) != 0; 81 } 82 83 void ElementaryStreamQueue::setCasInfo( 84 int32_t systemId, const std::vector<uint8_t> &sessionId) { 85 mCASystemId = systemId; 86 mCasSessionId = sessionId; 87 } 88 89 // Parse AC3 header assuming the current ptr is start position of syncframe, 90 // update metadata only applicable, and return the payload size 91 static unsigned parseAC3SyncFrame( 92 const uint8_t *ptr, size_t size, sp<MetaData> *metaData) { 93 static const unsigned channelCountTable[] = {2, 1, 2, 3, 3, 4, 4, 5}; 94 static const unsigned samplingRateTable[] = {48000, 44100, 32000}; 95 96 static const unsigned frameSizeTable[19][3] = { 97 { 64, 69, 96 }, 98 { 80, 87, 120 }, 99 { 96, 104, 144 }, 100 { 112, 121, 168 }, 101 { 128, 139, 192 }, 102 { 160, 174, 240 }, 103 { 192, 208, 288 }, 104 { 224, 243, 336 }, 105 { 256, 278, 384 }, 106 { 320, 348, 480 }, 107 { 384, 417, 576 }, 108 { 448, 487, 672 }, 109 { 512, 557, 768 }, 110 { 640, 696, 960 }, 111 { 768, 835, 1152 }, 112 { 896, 975, 1344 }, 113 { 1024, 1114, 1536 }, 114 { 1152, 1253, 1728 }, 115 { 1280, 1393, 1920 }, 116 }; 117 118 ABitReader bits(ptr, size); 119 if (bits.numBitsLeft() < 16) { 120 return 0; 121 } 122 if (bits.getBits(16) != 0x0B77) { 123 return 0; 124 } 125 126 if (bits.numBitsLeft() < 16 + 2 + 6 + 5 + 3 + 3) { 127 ALOGV("Not enough bits left for further parsing"); 128 return 0; 129 } 130 bits.skipBits(16); // crc1 131 132 unsigned fscod = bits.getBits(2); 133 if (fscod == 3) { 134 ALOGW("Incorrect fscod in AC3 header"); 135 return 0; 136 } 137 138 unsigned frmsizecod = bits.getBits(6); 139 if (frmsizecod > 37) { 140 ALOGW("Incorrect frmsizecod in AC3 header"); 141 return 0; 142 } 143 144 unsigned bsid = bits.getBits(5); 145 if (bsid > 8) { 146 ALOGW("Incorrect bsid in AC3 header. Possibly E-AC-3?"); 147 return 0; 148 } 149 150 unsigned bsmod __unused = bits.getBits(3); 151 unsigned acmod = bits.getBits(3); 152 unsigned cmixlev __unused = 0; 153 unsigned surmixlev __unused = 0; 154 unsigned dsurmod __unused = 0; 155 156 if ((acmod & 1) > 0 && acmod != 1) { 157 if (bits.numBitsLeft() < 2) { 158 return 0; 159 } 160 cmixlev = bits.getBits(2); 161 } 162 if ((acmod & 4) > 0) { 163 if (bits.numBitsLeft() < 2) { 164 return 0; 165 } 166 surmixlev = bits.getBits(2); 167 } 168 if (acmod == 2) { 169 if (bits.numBitsLeft() < 2) { 170 return 0; 171 } 172 dsurmod = bits.getBits(2); 173 } 174 175 if (bits.numBitsLeft() < 1) { 176 return 0; 177 } 178 unsigned lfeon = bits.getBits(1); 179 180 unsigned samplingRate = samplingRateTable[fscod]; 181 unsigned payloadSize = frameSizeTable[frmsizecod >> 1][fscod]; 182 if (fscod == 1) { 183 payloadSize += frmsizecod & 1; 184 } 185 payloadSize <<= 1; // convert from 16-bit words to bytes 186 187 unsigned channelCount = channelCountTable[acmod] + lfeon; 188 189 if (metaData != NULL) { 190 (*metaData)->setCString(kKeyMIMEType, MEDIA_MIMETYPE_AUDIO_AC3); 191 (*metaData)->setInt32(kKeyChannelCount, channelCount); 192 (*metaData)->setInt32(kKeySampleRate, samplingRate); 193 } 194 195 return payloadSize; 196 } 197 198 static bool IsSeeminglyValidAC3Header(const uint8_t *ptr, size_t size) { 199 return parseAC3SyncFrame(ptr, size, NULL) > 0; 200 } 201 202 static bool IsSeeminglyValidADTSHeader( 203 const uint8_t *ptr, size_t size, size_t *frameLength) { 204 if (size < 7) { 205 // Not enough data to verify header. 206 return false; 207 } 208 209 if (ptr[0] != 0xff || (ptr[1] >> 4) != 0x0f) { 210 return false; 211 } 212 213 unsigned layer = (ptr[1] >> 1) & 3; 214 215 if (layer != 0) { 216 return false; 217 } 218 219 unsigned ID = (ptr[1] >> 3) & 1; 220 unsigned profile_ObjectType = ptr[2] >> 6; 221 222 if (ID == 1 && profile_ObjectType == 3) { 223 // MPEG-2 profile 3 is reserved. 224 return false; 225 } 226 227 size_t frameLengthInHeader = 228 ((ptr[3] & 3) << 11) + (ptr[4] << 3) + ((ptr[5] >> 5) & 7); 229 if (frameLengthInHeader > size) { 230 return false; 231 } 232 233 *frameLength = frameLengthInHeader; 234 return true; 235 } 236 237 static bool IsSeeminglyValidMPEGAudioHeader(const uint8_t *ptr, size_t size) { 238 if (size < 3) { 239 // Not enough data to verify header. 240 return false; 241 } 242 243 if (ptr[0] != 0xff || (ptr[1] >> 5) != 0x07) { 244 return false; 245 } 246 247 unsigned ID = (ptr[1] >> 3) & 3; 248 249 if (ID == 1) { 250 return false; // reserved 251 } 252 253 unsigned layer = (ptr[1] >> 1) & 3; 254 255 if (layer == 0) { 256 return false; // reserved 257 } 258 259 unsigned bitrateIndex = (ptr[2] >> 4); 260 261 if (bitrateIndex == 0x0f) { 262 return false; // reserved 263 } 264 265 unsigned samplingRateIndex = (ptr[2] >> 2) & 3; 266 267 if (samplingRateIndex == 3) { 268 return false; // reserved 269 } 270 271 return true; 272 } 273 274 status_t ElementaryStreamQueue::appendData( 275 const void *data, size_t size, int64_t timeUs, 276 int32_t payloadOffset, uint32_t pesScramblingControl) { 277 278 if (mEOSReached) { 279 ALOGE("appending data after EOS"); 280 return ERROR_MALFORMED; 281 } 282 if (mBuffer == NULL || mBuffer->size() == 0) { 283 switch (mMode) { 284 case H264: 285 case MPEG_VIDEO: 286 { 287 #if 0 288 if (size < 4 || memcmp("\x00\x00\x00\x01", data, 4)) { 289 return ERROR_MALFORMED; 290 } 291 #else 292 uint8_t *ptr = (uint8_t *)data; 293 294 ssize_t startOffset = -1; 295 for (size_t i = 0; i + 2 < size; ++i) { 296 if (!memcmp("\x00\x00\x01", &ptr[i], 3)) { 297 startOffset = i; 298 break; 299 } 300 } 301 302 if (startOffset < 0) { 303 return ERROR_MALFORMED; 304 } 305 306 if (mFormat == NULL && startOffset > 0) { 307 ALOGI("found something resembling an H.264/MPEG syncword " 308 "at offset %zd", 309 startOffset); 310 } 311 312 data = &ptr[startOffset]; 313 size -= startOffset; 314 #endif 315 break; 316 } 317 318 case MPEG4_VIDEO: 319 { 320 #if 0 321 if (size < 3 || memcmp("\x00\x00\x01", data, 3)) { 322 return ERROR_MALFORMED; 323 } 324 #else 325 uint8_t *ptr = (uint8_t *)data; 326 327 ssize_t startOffset = -1; 328 for (size_t i = 0; i + 2 < size; ++i) { 329 if (!memcmp("\x00\x00\x01", &ptr[i], 3)) { 330 startOffset = i; 331 break; 332 } 333 } 334 335 if (startOffset < 0) { 336 return ERROR_MALFORMED; 337 } 338 339 if (startOffset > 0) { 340 ALOGI("found something resembling an H.264/MPEG syncword " 341 "at offset %zd", 342 startOffset); 343 } 344 345 data = &ptr[startOffset]; 346 size -= startOffset; 347 #endif 348 break; 349 } 350 351 case AAC: 352 { 353 uint8_t *ptr = (uint8_t *)data; 354 355 #if 0 356 if (size < 2 || ptr[0] != 0xff || (ptr[1] >> 4) != 0x0f) { 357 return ERROR_MALFORMED; 358 } 359 #else 360 ssize_t startOffset = -1; 361 size_t frameLength; 362 for (size_t i = 0; i < size; ++i) { 363 if (IsSeeminglyValidADTSHeader( 364 &ptr[i], size - i, &frameLength)) { 365 startOffset = i; 366 break; 367 } 368 } 369 370 if (startOffset < 0) { 371 return ERROR_MALFORMED; 372 } 373 374 if (startOffset > 0) { 375 ALOGI("found something resembling an AAC syncword at " 376 "offset %zd", 377 startOffset); 378 } 379 380 if (frameLength != size - startOffset) { 381 ALOGV("First ADTS AAC frame length is %zd bytes, " 382 "while the buffer size is %zd bytes.", 383 frameLength, size - startOffset); 384 } 385 386 data = &ptr[startOffset]; 387 size -= startOffset; 388 #endif 389 break; 390 } 391 392 case AC3: 393 { 394 uint8_t *ptr = (uint8_t *)data; 395 396 ssize_t startOffset = -1; 397 for (size_t i = 0; i < size; ++i) { 398 if (IsSeeminglyValidAC3Header(&ptr[i], size - i)) { 399 startOffset = i; 400 break; 401 } 402 } 403 404 if (startOffset < 0) { 405 return ERROR_MALFORMED; 406 } 407 408 if (startOffset > 0) { 409 ALOGI("found something resembling an AC3 syncword at " 410 "offset %zd", 411 startOffset); 412 } 413 414 data = &ptr[startOffset]; 415 size -= startOffset; 416 break; 417 } 418 419 case MPEG_AUDIO: 420 { 421 uint8_t *ptr = (uint8_t *)data; 422 423 ssize_t startOffset = -1; 424 for (size_t i = 0; i < size; ++i) { 425 if (IsSeeminglyValidMPEGAudioHeader(&ptr[i], size - i)) { 426 startOffset = i; 427 break; 428 } 429 } 430 431 if (startOffset < 0) { 432 return ERROR_MALFORMED; 433 } 434 435 if (startOffset > 0) { 436 ALOGI("found something resembling an MPEG audio " 437 "syncword at offset %zd", 438 startOffset); 439 } 440 441 data = &ptr[startOffset]; 442 size -= startOffset; 443 break; 444 } 445 446 case PCM_AUDIO: 447 case METADATA: 448 { 449 break; 450 } 451 452 default: 453 ALOGE("Unknown mode: %d", mMode); 454 return ERROR_MALFORMED; 455 } 456 } 457 458 size_t neededSize = (mBuffer == NULL ? 0 : mBuffer->size()) + size; 459 if (mBuffer == NULL || neededSize > mBuffer->capacity()) { 460 neededSize = (neededSize + 65535) & ~65535; 461 462 ALOGV("resizing buffer to size %zu", neededSize); 463 464 sp<ABuffer> buffer = new ABuffer(neededSize); 465 if (mBuffer != NULL) { 466 memcpy(buffer->data(), mBuffer->data(), mBuffer->size()); 467 buffer->setRange(0, mBuffer->size()); 468 } else { 469 buffer->setRange(0, 0); 470 } 471 472 mBuffer = buffer; 473 } 474 475 memcpy(mBuffer->data() + mBuffer->size(), data, size); 476 mBuffer->setRange(0, mBuffer->size() + size); 477 478 RangeInfo info; 479 info.mLength = size; 480 info.mTimestampUs = timeUs; 481 info.mPesOffset = payloadOffset; 482 info.mPesScramblingControl = pesScramblingControl; 483 mRangeInfos.push_back(info); 484 485 #if 0 486 if (mMode == AAC) { 487 ALOGI("size = %zu, timeUs = %.2f secs", size, timeUs / 1E6); 488 hexdump(data, size); 489 } 490 #endif 491 492 return OK; 493 } 494 495 void ElementaryStreamQueue::appendScrambledData( 496 const void *data, size_t size, 497 int32_t keyId, bool isSync, 498 sp<ABuffer> clearSizes, sp<ABuffer> encSizes) { 499 if (!isScrambled()) { 500 return; 501 } 502 503 size_t neededSize = (mScrambledBuffer == NULL ? 0 : mScrambledBuffer->size()) + size; 504 if (mScrambledBuffer == NULL || neededSize > mScrambledBuffer->capacity()) { 505 neededSize = (neededSize + 65535) & ~65535; 506 507 ALOGI("resizing scrambled buffer to size %zu", neededSize); 508 509 sp<ABuffer> buffer = new ABuffer(neededSize); 510 if (mScrambledBuffer != NULL) { 511 memcpy(buffer->data(), mScrambledBuffer->data(), mScrambledBuffer->size()); 512 buffer->setRange(0, mScrambledBuffer->size()); 513 } else { 514 buffer->setRange(0, 0); 515 } 516 517 mScrambledBuffer = buffer; 518 } 519 memcpy(mScrambledBuffer->data() + mScrambledBuffer->size(), data, size); 520 mScrambledBuffer->setRange(0, mScrambledBuffer->size() + size); 521 522 ScrambledRangeInfo scrambledInfo; 523 scrambledInfo.mLength = size; 524 scrambledInfo.mKeyId = keyId; 525 scrambledInfo.mIsSync = isSync; 526 scrambledInfo.mClearSizes = clearSizes; 527 scrambledInfo.mEncSizes = encSizes; 528 529 ALOGV("[stream %d] appending scrambled range: size=%zu", mMode, size); 530 531 mScrambledRangeInfos.push_back(scrambledInfo); 532 } 533 534 sp<ABuffer> ElementaryStreamQueue::dequeueScrambledAccessUnit() { 535 size_t nextScan = mBuffer->size(); 536 mBuffer->setRange(0, 0); 537 int32_t pesOffset = 0, pesScramblingControl = 0; 538 int64_t timeUs = fetchTimestamp(nextScan, &pesOffset, &pesScramblingControl); 539 if (timeUs < 0ll) { 540 ALOGE("Negative timeUs"); 541 return NULL; 542 } 543 544 // return scrambled unit 545 int32_t keyId = pesScramblingControl, isSync = 0, scrambledLength = 0; 546 sp<ABuffer> clearSizes, encSizes; 547 while (mScrambledRangeInfos.size() > mRangeInfos.size()) { 548 auto it = mScrambledRangeInfos.begin(); 549 ALOGV("[stream %d] fetching scrambled range: size=%zu", mMode, it->mLength); 550 551 if (scrambledLength > 0) { 552 // This shouldn't happen since we always dequeue the entire PES. 553 ALOGW("Discarding srambled length %d", scrambledLength); 554 } 555 scrambledLength = it->mLength; 556 557 // TODO: handle key id change, use first non-zero keyId for now 558 if (keyId == 0) { 559 keyId = it->mKeyId; 560 } 561 clearSizes = it->mClearSizes; 562 encSizes = it->mEncSizes; 563 isSync = it->mIsSync; 564 mScrambledRangeInfos.erase(it); 565 } 566 if (scrambledLength == 0) { 567 ALOGE("[stream %d] empty scrambled unit!", mMode); 568 return NULL; 569 } 570 571 // skip the PES header, and copy the rest into scrambled access unit 572 sp<ABuffer> scrambledAccessUnit = ABuffer::CreateAsCopy( 573 mScrambledBuffer->data() + pesOffset, 574 scrambledLength - pesOffset); 575 576 // fix up first sample size after skipping the PES header 577 if (pesOffset > 0) { 578 int32_t &firstClearSize = *(int32_t*)clearSizes->data(); 579 int32_t &firstEncSize = *(int32_t*)encSizes->data(); 580 // Cut away the PES header 581 if (firstClearSize >= pesOffset) { 582 // This is for TS-level scrambling, we descrambled the first 583 // (or it was clear to begin with) 584 firstClearSize -= pesOffset; 585 } else if (firstEncSize >= pesOffset) { 586 // This can only be PES-level scrambling 587 firstEncSize -= pesOffset; 588 } 589 } 590 591 scrambledAccessUnit->meta()->setInt64("timeUs", timeUs); 592 if (isSync) { 593 scrambledAccessUnit->meta()->setInt32("isSync", 1); 594 } 595 596 // fill in CryptoInfo fields for AnotherPacketSource::read() 597 // MediaCas doesn't use cryptoMode, but set to non-zero value here. 598 scrambledAccessUnit->meta()->setInt32( 599 "cryptoMode", CryptoPlugin::kMode_AES_CTR); 600 scrambledAccessUnit->meta()->setInt32("cryptoKey", keyId); 601 scrambledAccessUnit->meta()->setBuffer("clearBytes", clearSizes); 602 scrambledAccessUnit->meta()->setBuffer("encBytes", encSizes); 603 604 memmove(mScrambledBuffer->data(), 605 mScrambledBuffer->data() + scrambledLength, 606 mScrambledBuffer->size() - scrambledLength); 607 608 mScrambledBuffer->setRange(0, mScrambledBuffer->size() - scrambledLength); 609 610 ALOGV("[stream %d] dequeued scrambled AU: timeUs=%lld, size=%zu", 611 mMode, (long long)timeUs, scrambledAccessUnit->size()); 612 613 return scrambledAccessUnit; 614 } 615 616 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnit() { 617 if ((mFlags & kFlag_AlignedData) && mMode == H264 && !isScrambled()) { 618 if (mRangeInfos.empty()) { 619 return NULL; 620 } 621 622 RangeInfo info = *mRangeInfos.begin(); 623 mRangeInfos.erase(mRangeInfos.begin()); 624 625 sp<ABuffer> accessUnit = new ABuffer(info.mLength); 626 memcpy(accessUnit->data(), mBuffer->data(), info.mLength); 627 accessUnit->meta()->setInt64("timeUs", info.mTimestampUs); 628 629 memmove(mBuffer->data(), 630 mBuffer->data() + info.mLength, 631 mBuffer->size() - info.mLength); 632 633 mBuffer->setRange(0, mBuffer->size() - info.mLength); 634 635 if (mFormat == NULL) { 636 mFormat = MakeAVCCodecSpecificData(accessUnit); 637 } 638 639 return accessUnit; 640 } 641 642 switch (mMode) { 643 case H264: 644 return dequeueAccessUnitH264(); 645 case AAC: 646 return dequeueAccessUnitAAC(); 647 case AC3: 648 return dequeueAccessUnitAC3(); 649 case MPEG_VIDEO: 650 return dequeueAccessUnitMPEGVideo(); 651 case MPEG4_VIDEO: 652 return dequeueAccessUnitMPEG4Video(); 653 case PCM_AUDIO: 654 return dequeueAccessUnitPCMAudio(); 655 case METADATA: 656 return dequeueAccessUnitMetadata(); 657 default: 658 if (mMode != MPEG_AUDIO) { 659 ALOGE("Unknown mode"); 660 return NULL; 661 } 662 return dequeueAccessUnitMPEGAudio(); 663 } 664 } 665 666 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitAC3() { 667 unsigned syncStartPos = 0; // in bytes 668 unsigned payloadSize = 0; 669 sp<MetaData> format = new MetaData; 670 671 ALOGV("dequeueAccessUnit_AC3[%d]: mBuffer %p(%zu)", mAUIndex, mBuffer->data(), mBuffer->size()); 672 673 while (true) { 674 if (syncStartPos + 2 >= mBuffer->size()) { 675 return NULL; 676 } 677 678 payloadSize = parseAC3SyncFrame( 679 mBuffer->data() + syncStartPos, 680 mBuffer->size() - syncStartPos, 681 &format); 682 if (payloadSize > 0) { 683 break; 684 } 685 686 ALOGV("dequeueAccessUnit_AC3[%d]: syncStartPos %u payloadSize %u", 687 mAUIndex, syncStartPos, payloadSize); 688 689 ++syncStartPos; 690 } 691 692 if (mBuffer->size() < syncStartPos + payloadSize) { 693 ALOGV("Not enough buffer size for AC3"); 694 return NULL; 695 } 696 697 if (mFormat == NULL) { 698 mFormat = format; 699 } 700 701 702 int64_t timeUs = fetchTimestamp(syncStartPos + payloadSize); 703 if (timeUs < 0ll) { 704 ALOGE("negative timeUs"); 705 return NULL; 706 } 707 708 // Not decrypting if key info not available (e.g., scanner/extractor parsing ts files) 709 if (mSampleDecryptor != NULL) { 710 mSampleDecryptor->processAC3(mBuffer->data() + syncStartPos, payloadSize); 711 } 712 mAUIndex++; 713 714 sp<ABuffer> accessUnit = new ABuffer(syncStartPos + payloadSize); 715 memcpy(accessUnit->data(), mBuffer->data(), syncStartPos + payloadSize); 716 717 accessUnit->meta()->setInt64("timeUs", timeUs); 718 accessUnit->meta()->setInt32("isSync", 1); 719 720 memmove( 721 mBuffer->data(), 722 mBuffer->data() + syncStartPos + payloadSize, 723 mBuffer->size() - syncStartPos - payloadSize); 724 725 mBuffer->setRange(0, mBuffer->size() - syncStartPos - payloadSize); 726 727 return accessUnit; 728 } 729 730 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitPCMAudio() { 731 if (mBuffer->size() < 4) { 732 return NULL; 733 } 734 735 ABitReader bits(mBuffer->data(), 4); 736 if (bits.getBits(8) != 0xa0) { 737 ALOGE("Unexpected bit values"); 738 return NULL; 739 } 740 unsigned numAUs = bits.getBits(8); 741 bits.skipBits(8); 742 unsigned quantization_word_length __unused = bits.getBits(2); 743 unsigned audio_sampling_frequency = bits.getBits(3); 744 unsigned num_channels = bits.getBits(3); 745 746 if (audio_sampling_frequency != 2) { 747 ALOGE("Wrong sampling freq"); 748 return NULL; 749 } 750 if (num_channels != 1u) { 751 ALOGE("Wrong channel #"); 752 return NULL; 753 } 754 755 if (mFormat == NULL) { 756 mFormat = new MetaData; 757 mFormat->setCString(kKeyMIMEType, MEDIA_MIMETYPE_AUDIO_RAW); 758 mFormat->setInt32(kKeyChannelCount, 2); 759 mFormat->setInt32(kKeySampleRate, 48000); 760 mFormat->setInt32(kKeyPcmEncoding, kAudioEncodingPcm16bit); 761 } 762 763 static const size_t kFramesPerAU = 80; 764 size_t frameSize = 2 /* numChannels */ * sizeof(int16_t); 765 766 size_t payloadSize = numAUs * frameSize * kFramesPerAU; 767 768 if (mBuffer->size() < 4 + payloadSize) { 769 return NULL; 770 } 771 772 sp<ABuffer> accessUnit = new ABuffer(payloadSize); 773 memcpy(accessUnit->data(), mBuffer->data() + 4, payloadSize); 774 775 int64_t timeUs = fetchTimestamp(payloadSize + 4); 776 if (timeUs < 0ll) { 777 ALOGE("Negative timeUs"); 778 return NULL; 779 } 780 accessUnit->meta()->setInt64("timeUs", timeUs); 781 accessUnit->meta()->setInt32("isSync", 1); 782 783 int16_t *ptr = (int16_t *)accessUnit->data(); 784 for (size_t i = 0; i < payloadSize / sizeof(int16_t); ++i) { 785 ptr[i] = ntohs(ptr[i]); 786 } 787 788 memmove( 789 mBuffer->data(), 790 mBuffer->data() + 4 + payloadSize, 791 mBuffer->size() - 4 - payloadSize); 792 793 mBuffer->setRange(0, mBuffer->size() - 4 - payloadSize); 794 795 return accessUnit; 796 } 797 798 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitAAC() { 799 if (mBuffer->size() == 0) { 800 return NULL; 801 } 802 803 if (mRangeInfos.empty()) { 804 return NULL; 805 } 806 807 const RangeInfo &info = *mRangeInfos.begin(); 808 if (mBuffer->size() < info.mLength) { 809 return NULL; 810 } 811 812 if (info.mTimestampUs < 0ll) { 813 ALOGE("Negative info.mTimestampUs"); 814 return NULL; 815 } 816 817 ALOGV("dequeueAccessUnit_AAC[%d]: mBuffer %zu info.mLength %zu", 818 mAUIndex, mBuffer->size(), info.mLength); 819 820 struct ADTSPosition { 821 size_t offset; 822 size_t headerSize; 823 size_t length; 824 }; 825 826 Vector<ADTSPosition> frames; 827 828 // The idea here is consume all AAC frames starting at offsets before 829 // info.mLength so we can assign a meaningful timestamp without 830 // having to interpolate. 831 // The final AAC frame may well extend into the next RangeInfo but 832 // that's ok. 833 size_t offset = 0; 834 while (offset < info.mLength) { 835 if (offset + 7 > mBuffer->size()) { 836 return NULL; 837 } 838 839 ABitReader bits(mBuffer->data() + offset, mBuffer->size() - offset); 840 841 // adts_fixed_header 842 843 if (bits.getBits(12) != 0xfffu) { 844 ALOGE("Wrong atds_fixed_header"); 845 return NULL; 846 } 847 bits.skipBits(3); // ID, layer 848 bool protection_absent = bits.getBits(1) != 0; 849 850 if (mFormat == NULL) { 851 unsigned profile = bits.getBits(2); 852 if (profile == 3u) { 853 ALOGE("profile should not be 3"); 854 return NULL; 855 } 856 unsigned sampling_freq_index = bits.getBits(4); 857 bits.getBits(1); // private_bit 858 unsigned channel_configuration = bits.getBits(3); 859 if (channel_configuration == 0u) { 860 ALOGE("channel_config should not be 0"); 861 return NULL; 862 } 863 bits.skipBits(2); // original_copy, home 864 865 mFormat = MakeAACCodecSpecificData( 866 profile, sampling_freq_index, channel_configuration); 867 868 mFormat->setInt32(kKeyIsADTS, true); 869 870 int32_t sampleRate; 871 int32_t numChannels; 872 if (!mFormat->findInt32(kKeySampleRate, &sampleRate)) { 873 ALOGE("SampleRate not found"); 874 return NULL; 875 } 876 if (!mFormat->findInt32(kKeyChannelCount, &numChannels)) { 877 ALOGE("ChannelCount not found"); 878 return NULL; 879 } 880 881 ALOGI("found AAC codec config (%d Hz, %d channels)", 882 sampleRate, numChannels); 883 } else { 884 // profile_ObjectType, sampling_frequency_index, private_bits, 885 // channel_configuration, original_copy, home 886 bits.skipBits(12); 887 } 888 889 // adts_variable_header 890 891 // copyright_identification_bit, copyright_identification_start 892 bits.skipBits(2); 893 894 unsigned aac_frame_length = bits.getBits(13); 895 if (aac_frame_length == 0){ 896 ALOGE("b/62673179, Invalid AAC frame length!"); 897 android_errorWriteLog(0x534e4554, "62673179"); 898 return NULL; 899 } 900 901 bits.skipBits(11); // adts_buffer_fullness 902 903 unsigned number_of_raw_data_blocks_in_frame = bits.getBits(2); 904 905 if (number_of_raw_data_blocks_in_frame != 0) { 906 // To be implemented. 907 ALOGE("Should not reach here."); 908 return NULL; 909 } 910 911 if (offset + aac_frame_length > mBuffer->size()) { 912 return NULL; 913 } 914 915 size_t headerSize = protection_absent ? 7 : 9; 916 917 // tracking the frame positions first then decrypt only if an accessUnit to be generated 918 if (mSampleDecryptor != NULL) { 919 ADTSPosition frame = { 920 .offset = offset, 921 .headerSize = headerSize, 922 .length = aac_frame_length 923 }; 924 925 frames.push(frame); 926 } 927 928 offset += aac_frame_length; 929 } 930 931 // Decrypting only if the loop didn't exit early and an accessUnit is about to be generated 932 // Not decrypting if key info not available (e.g., scanner/extractor parsing ts files) 933 if (mSampleDecryptor != NULL) { 934 for (size_t frameId = 0; frameId < frames.size(); frameId++) { 935 const ADTSPosition &frame = frames.itemAt(frameId); 936 937 mSampleDecryptor->processAAC(frame.headerSize, 938 mBuffer->data() + frame.offset, frame.length); 939 // ALOGV("dequeueAccessUnitAAC[%zu]: while offset %zu headerSize %zu frame_len %zu", 940 // frameId, frame.offset, frame.headerSize, frame.length); 941 } 942 } 943 mAUIndex++; 944 945 int64_t timeUs = fetchTimestamp(offset); 946 947 sp<ABuffer> accessUnit = new ABuffer(offset); 948 memcpy(accessUnit->data(), mBuffer->data(), offset); 949 950 memmove(mBuffer->data(), mBuffer->data() + offset, 951 mBuffer->size() - offset); 952 mBuffer->setRange(0, mBuffer->size() - offset); 953 954 accessUnit->meta()->setInt64("timeUs", timeUs); 955 accessUnit->meta()->setInt32("isSync", 1); 956 957 return accessUnit; 958 } 959 960 int64_t ElementaryStreamQueue::fetchTimestamp( 961 size_t size, int32_t *pesOffset, int32_t *pesScramblingControl) { 962 int64_t timeUs = -1; 963 bool first = true; 964 965 while (size > 0) { 966 if (mRangeInfos.empty()) { 967 return timeUs; 968 } 969 970 RangeInfo *info = &*mRangeInfos.begin(); 971 972 if (first) { 973 timeUs = info->mTimestampUs; 974 if (pesOffset != NULL) { 975 *pesOffset = info->mPesOffset; 976 } 977 if (pesScramblingControl != NULL) { 978 *pesScramblingControl = info->mPesScramblingControl; 979 } 980 first = false; 981 } 982 983 if (info->mLength > size) { 984 info->mLength -= size; 985 size = 0; 986 } else { 987 size -= info->mLength; 988 989 mRangeInfos.erase(mRangeInfos.begin()); 990 info = NULL; 991 } 992 993 } 994 995 if (timeUs == 0ll) { 996 ALOGV("Returning 0 timestamp"); 997 } 998 999 return timeUs; 1000 } 1001 1002 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitH264() { 1003 if (isScrambled()) { 1004 if (mBuffer == NULL || mBuffer->size() == 0) { 1005 return NULL; 1006 } 1007 if (mFormat == NULL) { 1008 mFormat = MakeAVCCodecSpecificData(mBuffer); 1009 if (mFormat == NULL) { 1010 ALOGI("Creating dummy AVC format for scrambled content"); 1011 mFormat = new MetaData; 1012 mFormat->setCString(kKeyMIMEType, MEDIA_MIMETYPE_VIDEO_AVC); 1013 mFormat->setInt32(kKeyWidth, 1280); 1014 mFormat->setInt32(kKeyHeight, 720); 1015 } 1016 // for MediaExtractor.CasInfo 1017 mFormat->setInt32(kKeyCASystemID, mCASystemId); 1018 mFormat->setData(kKeyCASessionID, 0, 1019 mCasSessionId.data(), mCasSessionId.size()); 1020 } 1021 return dequeueScrambledAccessUnit(); 1022 } 1023 1024 const uint8_t *data = mBuffer->data(); 1025 1026 size_t size = mBuffer->size(); 1027 Vector<NALPosition> nals; 1028 1029 size_t totalSize = 0; 1030 size_t seiCount = 0; 1031 1032 status_t err; 1033 const uint8_t *nalStart; 1034 size_t nalSize; 1035 bool foundSlice = false; 1036 bool foundIDR = false; 1037 1038 ALOGV("dequeueAccessUnit_H264[%d] %p/%zu", mAUIndex, data, size); 1039 1040 while ((err = getNextNALUnit(&data, &size, &nalStart, &nalSize)) == OK) { 1041 if (nalSize == 0) continue; 1042 1043 unsigned nalType = nalStart[0] & 0x1f; 1044 bool flush = false; 1045 1046 if (nalType == 1 || nalType == 5) { 1047 if (nalType == 5) { 1048 foundIDR = true; 1049 } 1050 if (foundSlice) { 1051 //TODO: Shouldn't this have been called with nalSize-1? 1052 ABitReader br(nalStart + 1, nalSize); 1053 unsigned first_mb_in_slice = parseUE(&br); 1054 1055 if (first_mb_in_slice == 0) { 1056 // This slice starts a new frame. 1057 1058 flush = true; 1059 } 1060 } 1061 1062 foundSlice = true; 1063 } else if ((nalType == 9 || nalType == 7) && foundSlice) { 1064 // Access unit delimiter and SPS will be associated with the 1065 // next frame. 1066 1067 flush = true; 1068 } else if (nalType == 6 && nalSize > 0) { 1069 // found non-zero sized SEI 1070 ++seiCount; 1071 } 1072 1073 if (flush) { 1074 // The access unit will contain all nal units up to, but excluding 1075 // the current one, separated by 0x00 0x00 0x00 0x01 startcodes. 1076 1077 size_t auSize = 4 * nals.size() + totalSize; 1078 sp<ABuffer> accessUnit = new ABuffer(auSize); 1079 sp<ABuffer> sei; 1080 1081 if (seiCount > 0) { 1082 sei = new ABuffer(seiCount * sizeof(NALPosition)); 1083 accessUnit->meta()->setBuffer("sei", sei); 1084 } 1085 1086 #if !LOG_NDEBUG 1087 AString out; 1088 #endif 1089 1090 size_t dstOffset = 0; 1091 size_t seiIndex = 0; 1092 size_t shrunkBytes = 0; 1093 for (size_t i = 0; i < nals.size(); ++i) { 1094 const NALPosition &pos = nals.itemAt(i); 1095 1096 unsigned nalType = mBuffer->data()[pos.nalOffset] & 0x1f; 1097 1098 if (nalType == 6 && pos.nalSize > 0) { 1099 if (seiIndex >= sei->size() / sizeof(NALPosition)) { 1100 ALOGE("Wrong seiIndex"); 1101 return NULL; 1102 } 1103 NALPosition &seiPos = ((NALPosition *)sei->data())[seiIndex++]; 1104 seiPos.nalOffset = dstOffset + 4; 1105 seiPos.nalSize = pos.nalSize; 1106 } 1107 1108 #if !LOG_NDEBUG 1109 char tmp[128]; 1110 sprintf(tmp, "0x%02x", nalType); 1111 if (i > 0) { 1112 out.append(", "); 1113 } 1114 out.append(tmp); 1115 #endif 1116 1117 memcpy(accessUnit->data() + dstOffset, "\x00\x00\x00\x01", 4); 1118 1119 if (mSampleDecryptor != NULL && (nalType == 1 || nalType == 5)) { 1120 uint8_t *nalData = mBuffer->data() + pos.nalOffset; 1121 size_t newSize = mSampleDecryptor->processNal(nalData, pos.nalSize); 1122 // Note: the data can shrink due to unescaping 1123 memcpy(accessUnit->data() + dstOffset + 4, 1124 nalData, 1125 newSize); 1126 dstOffset += newSize + 4; 1127 1128 size_t thisShrunkBytes = pos.nalSize - newSize; 1129 //ALOGV("dequeueAccessUnitH264[%d]: nalType: %d -> %zu (%zu)", 1130 // nalType, (int)pos.nalSize, newSize, thisShrunkBytes); 1131 1132 shrunkBytes += thisShrunkBytes; 1133 } 1134 else { 1135 memcpy(accessUnit->data() + dstOffset + 4, 1136 mBuffer->data() + pos.nalOffset, 1137 pos.nalSize); 1138 1139 dstOffset += pos.nalSize + 4; 1140 //ALOGV("dequeueAccessUnitH264 [%d] %d @%d", 1141 // nalType, (int)pos.nalSize, (int)pos.nalOffset); 1142 } 1143 } 1144 1145 #if !LOG_NDEBUG 1146 ALOGV("accessUnit contains nal types %s", out.c_str()); 1147 #endif 1148 1149 const NALPosition &pos = nals.itemAt(nals.size() - 1); 1150 size_t nextScan = pos.nalOffset + pos.nalSize; 1151 1152 memmove(mBuffer->data(), 1153 mBuffer->data() + nextScan, 1154 mBuffer->size() - nextScan); 1155 1156 mBuffer->setRange(0, mBuffer->size() - nextScan); 1157 1158 int64_t timeUs = fetchTimestamp(nextScan); 1159 if (timeUs < 0ll) { 1160 ALOGE("Negative timeUs"); 1161 return NULL; 1162 } 1163 1164 accessUnit->meta()->setInt64("timeUs", timeUs); 1165 if (foundIDR) { 1166 accessUnit->meta()->setInt32("isSync", 1); 1167 } 1168 1169 if (mFormat == NULL) { 1170 mFormat = MakeAVCCodecSpecificData(accessUnit); 1171 } 1172 1173 if (mSampleDecryptor != NULL && shrunkBytes > 0) { 1174 size_t adjustedSize = accessUnit->size() - shrunkBytes; 1175 ALOGV("dequeueAccessUnitH264[%d]: AU size adjusted %zu -> %zu", 1176 mAUIndex, accessUnit->size(), adjustedSize); 1177 accessUnit->setRange(0, adjustedSize); 1178 } 1179 1180 ALOGV("dequeueAccessUnitH264[%d]: AU %p(%zu) dstOffset:%zu, nals:%zu, totalSize:%zu ", 1181 mAUIndex, accessUnit->data(), accessUnit->size(), 1182 dstOffset, nals.size(), totalSize); 1183 mAUIndex++; 1184 1185 return accessUnit; 1186 } 1187 1188 NALPosition pos; 1189 pos.nalOffset = nalStart - mBuffer->data(); 1190 pos.nalSize = nalSize; 1191 1192 nals.push(pos); 1193 1194 totalSize += nalSize; 1195 } 1196 if (err != (status_t)-EAGAIN) { 1197 ALOGE("Unexpeted err"); 1198 return NULL; 1199 } 1200 1201 return NULL; 1202 } 1203 1204 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitMPEGAudio() { 1205 const uint8_t *data = mBuffer->data(); 1206 size_t size = mBuffer->size(); 1207 1208 if (size < 4) { 1209 return NULL; 1210 } 1211 1212 uint32_t header = U32_AT(data); 1213 1214 size_t frameSize; 1215 int samplingRate, numChannels, bitrate, numSamples; 1216 if (!GetMPEGAudioFrameSize( 1217 header, &frameSize, &samplingRate, &numChannels, 1218 &bitrate, &numSamples)) { 1219 ALOGE("Failed to get audio frame size"); 1220 return NULL; 1221 } 1222 1223 if (size < frameSize) { 1224 return NULL; 1225 } 1226 1227 unsigned layer = 4 - ((header >> 17) & 3); 1228 1229 sp<ABuffer> accessUnit = new ABuffer(frameSize); 1230 memcpy(accessUnit->data(), data, frameSize); 1231 1232 memmove(mBuffer->data(), 1233 mBuffer->data() + frameSize, 1234 mBuffer->size() - frameSize); 1235 1236 mBuffer->setRange(0, mBuffer->size() - frameSize); 1237 1238 int64_t timeUs = fetchTimestamp(frameSize); 1239 if (timeUs < 0ll) { 1240 ALOGE("Negative timeUs"); 1241 return NULL; 1242 } 1243 1244 accessUnit->meta()->setInt64("timeUs", timeUs); 1245 accessUnit->meta()->setInt32("isSync", 1); 1246 1247 if (mFormat == NULL) { 1248 mFormat = new MetaData; 1249 1250 switch (layer) { 1251 case 1: 1252 mFormat->setCString( 1253 kKeyMIMEType, MEDIA_MIMETYPE_AUDIO_MPEG_LAYER_I); 1254 break; 1255 case 2: 1256 mFormat->setCString( 1257 kKeyMIMEType, MEDIA_MIMETYPE_AUDIO_MPEG_LAYER_II); 1258 break; 1259 case 3: 1260 mFormat->setCString( 1261 kKeyMIMEType, MEDIA_MIMETYPE_AUDIO_MPEG); 1262 break; 1263 default: 1264 return NULL; 1265 } 1266 1267 mFormat->setInt32(kKeySampleRate, samplingRate); 1268 mFormat->setInt32(kKeyChannelCount, numChannels); 1269 } 1270 1271 return accessUnit; 1272 } 1273 1274 static void EncodeSize14(uint8_t **_ptr, size_t size) { 1275 if (size > 0x3fff) { 1276 ALOGE("Wrong size"); 1277 return; 1278 } 1279 1280 uint8_t *ptr = *_ptr; 1281 1282 *ptr++ = 0x80 | (size >> 7); 1283 *ptr++ = size & 0x7f; 1284 1285 *_ptr = ptr; 1286 } 1287 1288 static sp<ABuffer> MakeMPEGVideoESDS(const sp<ABuffer> &csd) { 1289 sp<ABuffer> esds = new ABuffer(csd->size() + 25); 1290 1291 uint8_t *ptr = esds->data(); 1292 *ptr++ = 0x03; 1293 EncodeSize14(&ptr, 22 + csd->size()); 1294 1295 *ptr++ = 0x00; // ES_ID 1296 *ptr++ = 0x00; 1297 1298 *ptr++ = 0x00; // streamDependenceFlag, URL_Flag, OCRstreamFlag 1299 1300 *ptr++ = 0x04; 1301 EncodeSize14(&ptr, 16 + csd->size()); 1302 1303 *ptr++ = 0x40; // Audio ISO/IEC 14496-3 1304 1305 for (size_t i = 0; i < 12; ++i) { 1306 *ptr++ = 0x00; 1307 } 1308 1309 *ptr++ = 0x05; 1310 EncodeSize14(&ptr, csd->size()); 1311 1312 memcpy(ptr, csd->data(), csd->size()); 1313 1314 return esds; 1315 } 1316 1317 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitMPEGVideo() { 1318 if (isScrambled()) { 1319 if (mBuffer == NULL || mBuffer->size() == 0) { 1320 return NULL; 1321 } 1322 if (mFormat == NULL) { 1323 ALOGI("Creating dummy MPEG format for scrambled content"); 1324 mFormat = new MetaData; 1325 mFormat->setCString(kKeyMIMEType, MEDIA_MIMETYPE_VIDEO_MPEG2); 1326 mFormat->setInt32(kKeyWidth, 1280); 1327 mFormat->setInt32(kKeyHeight, 720); 1328 1329 // for MediaExtractor.CasInfo 1330 mFormat->setInt32(kKeyCASystemID, mCASystemId); 1331 mFormat->setData(kKeyCASessionID, 0, 1332 mCasSessionId.data(), mCasSessionId.size()); 1333 } 1334 return dequeueScrambledAccessUnit(); 1335 } 1336 1337 const uint8_t *data = mBuffer->data(); 1338 size_t size = mBuffer->size(); 1339 1340 Vector<size_t> userDataPositions; 1341 1342 bool sawPictureStart = false; 1343 int pprevStartCode = -1; 1344 int prevStartCode = -1; 1345 int currentStartCode = -1; 1346 bool gopFound = false; 1347 bool isClosedGop = false; 1348 bool brokenLink = false; 1349 1350 size_t offset = 0; 1351 while (offset + 3 < size) { 1352 if (memcmp(&data[offset], "\x00\x00\x01", 3)) { 1353 ++offset; 1354 continue; 1355 } 1356 1357 pprevStartCode = prevStartCode; 1358 prevStartCode = currentStartCode; 1359 currentStartCode = data[offset + 3]; 1360 1361 if (currentStartCode == 0xb3 && mFormat == NULL) { 1362 memmove(mBuffer->data(), mBuffer->data() + offset, size - offset); 1363 size -= offset; 1364 (void)fetchTimestamp(offset); 1365 offset = 0; 1366 mBuffer->setRange(0, size); 1367 } 1368 1369 if ((prevStartCode == 0xb3 && currentStartCode != 0xb5) 1370 || (pprevStartCode == 0xb3 && prevStartCode == 0xb5)) { 1371 // seqHeader without/with extension 1372 1373 if (mFormat == NULL) { 1374 if (size < 7u) { 1375 ALOGE("Size too small"); 1376 return NULL; 1377 } 1378 1379 unsigned width = 1380 (data[4] << 4) | data[5] >> 4; 1381 1382 unsigned height = 1383 ((data[5] & 0x0f) << 8) | data[6]; 1384 1385 mFormat = new MetaData; 1386 mFormat->setCString(kKeyMIMEType, MEDIA_MIMETYPE_VIDEO_MPEG2); 1387 mFormat->setInt32(kKeyWidth, width); 1388 mFormat->setInt32(kKeyHeight, height); 1389 1390 ALOGI("found MPEG2 video codec config (%d x %d)", width, height); 1391 1392 sp<ABuffer> csd = new ABuffer(offset); 1393 memcpy(csd->data(), data, offset); 1394 1395 memmove(mBuffer->data(), 1396 mBuffer->data() + offset, 1397 mBuffer->size() - offset); 1398 1399 mBuffer->setRange(0, mBuffer->size() - offset); 1400 size -= offset; 1401 (void)fetchTimestamp(offset); 1402 offset = 0; 1403 1404 // hexdump(csd->data(), csd->size()); 1405 1406 sp<ABuffer> esds = MakeMPEGVideoESDS(csd); 1407 mFormat->setData( 1408 kKeyESDS, kTypeESDS, esds->data(), esds->size()); 1409 1410 return NULL; 1411 } 1412 } 1413 1414 if (mFormat != NULL && currentStartCode == 0xb8) { 1415 // GOP layer 1416 if (offset + 7 >= size) { 1417 ALOGE("Size too small"); 1418 return NULL; 1419 } 1420 gopFound = true; 1421 isClosedGop = (data[offset + 7] & 0x40) != 0; 1422 brokenLink = (data[offset + 7] & 0x20) != 0; 1423 } 1424 1425 if (mFormat != NULL && currentStartCode == 0xb2) { 1426 userDataPositions.add(offset); 1427 } 1428 1429 if (mFormat != NULL && currentStartCode == 0x00) { 1430 // Picture start 1431 1432 if (!sawPictureStart) { 1433 sawPictureStart = true; 1434 } else { 1435 sp<ABuffer> accessUnit = new ABuffer(offset); 1436 memcpy(accessUnit->data(), data, offset); 1437 1438 memmove(mBuffer->data(), 1439 mBuffer->data() + offset, 1440 mBuffer->size() - offset); 1441 1442 mBuffer->setRange(0, mBuffer->size() - offset); 1443 1444 int64_t timeUs = fetchTimestamp(offset); 1445 if (timeUs < 0ll) { 1446 ALOGE("Negative timeUs"); 1447 return NULL; 1448 } 1449 1450 offset = 0; 1451 1452 accessUnit->meta()->setInt64("timeUs", timeUs); 1453 if (gopFound && (!brokenLink || isClosedGop)) { 1454 accessUnit->meta()->setInt32("isSync", 1); 1455 } 1456 1457 ALOGV("returning MPEG video access unit at time %" PRId64 " us", 1458 timeUs); 1459 1460 // hexdump(accessUnit->data(), accessUnit->size()); 1461 1462 if (userDataPositions.size() > 0) { 1463 sp<ABuffer> mpegUserData = 1464 new ABuffer(userDataPositions.size() * sizeof(size_t)); 1465 if (mpegUserData != NULL && mpegUserData->data() != NULL) { 1466 for (size_t i = 0; i < userDataPositions.size(); ++i) { 1467 memcpy( 1468 mpegUserData->data() + i * sizeof(size_t), 1469 &userDataPositions[i], sizeof(size_t)); 1470 } 1471 accessUnit->meta()->setBuffer("mpegUserData", mpegUserData); 1472 } 1473 } 1474 1475 return accessUnit; 1476 } 1477 } 1478 1479 ++offset; 1480 } 1481 1482 return NULL; 1483 } 1484 1485 static ssize_t getNextChunkSize( 1486 const uint8_t *data, size_t size) { 1487 static const char kStartCode[] = "\x00\x00\x01"; 1488 1489 if (size < 3) { 1490 return -EAGAIN; 1491 } 1492 1493 if (memcmp(kStartCode, data, 3)) { 1494 return -EAGAIN; 1495 } 1496 1497 size_t offset = 3; 1498 while (offset + 2 < size) { 1499 if (!memcmp(&data[offset], kStartCode, 3)) { 1500 return offset; 1501 } 1502 1503 ++offset; 1504 } 1505 1506 return -EAGAIN; 1507 } 1508 1509 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitMPEG4Video() { 1510 uint8_t *data = mBuffer->data(); 1511 size_t size = mBuffer->size(); 1512 1513 enum { 1514 SKIP_TO_VISUAL_OBJECT_SEQ_START, 1515 EXPECT_VISUAL_OBJECT_START, 1516 EXPECT_VO_START, 1517 EXPECT_VOL_START, 1518 WAIT_FOR_VOP_START, 1519 SKIP_TO_VOP_START, 1520 1521 } state; 1522 1523 if (mFormat == NULL) { 1524 state = SKIP_TO_VISUAL_OBJECT_SEQ_START; 1525 } else { 1526 state = SKIP_TO_VOP_START; 1527 } 1528 1529 int32_t width = -1, height = -1; 1530 1531 size_t offset = 0; 1532 ssize_t chunkSize; 1533 while ((chunkSize = getNextChunkSize( 1534 &data[offset], size - offset)) > 0) { 1535 bool discard = false; 1536 1537 unsigned chunkType = data[offset + 3]; 1538 1539 switch (state) { 1540 case SKIP_TO_VISUAL_OBJECT_SEQ_START: 1541 { 1542 if (chunkType == 0xb0) { 1543 // Discard anything before this marker. 1544 1545 state = EXPECT_VISUAL_OBJECT_START; 1546 } else { 1547 discard = true; 1548 } 1549 break; 1550 } 1551 1552 case EXPECT_VISUAL_OBJECT_START: 1553 { 1554 if (chunkType != 0xb5) { 1555 ALOGE("Unexpected chunkType"); 1556 return NULL; 1557 } 1558 state = EXPECT_VO_START; 1559 break; 1560 } 1561 1562 case EXPECT_VO_START: 1563 { 1564 if (chunkType > 0x1f) { 1565 ALOGE("Unexpected chunkType"); 1566 return NULL; 1567 } 1568 state = EXPECT_VOL_START; 1569 break; 1570 } 1571 1572 case EXPECT_VOL_START: 1573 { 1574 if ((chunkType & 0xf0) != 0x20) { 1575 ALOGE("Wrong chunkType"); 1576 return NULL; 1577 } 1578 1579 if (!ExtractDimensionsFromVOLHeader( 1580 &data[offset], chunkSize, 1581 &width, &height)) { 1582 ALOGE("Failed to get dimension"); 1583 return NULL; 1584 } 1585 1586 state = WAIT_FOR_VOP_START; 1587 break; 1588 } 1589 1590 case WAIT_FOR_VOP_START: 1591 { 1592 if (chunkType == 0xb3 || chunkType == 0xb6) { 1593 // group of VOP or VOP start. 1594 1595 mFormat = new MetaData; 1596 mFormat->setCString( 1597 kKeyMIMEType, MEDIA_MIMETYPE_VIDEO_MPEG4); 1598 1599 mFormat->setInt32(kKeyWidth, width); 1600 mFormat->setInt32(kKeyHeight, height); 1601 1602 ALOGI("found MPEG4 video codec config (%d x %d)", 1603 width, height); 1604 1605 sp<ABuffer> csd = new ABuffer(offset); 1606 memcpy(csd->data(), data, offset); 1607 1608 // hexdump(csd->data(), csd->size()); 1609 1610 sp<ABuffer> esds = MakeMPEGVideoESDS(csd); 1611 mFormat->setData( 1612 kKeyESDS, kTypeESDS, 1613 esds->data(), esds->size()); 1614 1615 discard = true; 1616 state = SKIP_TO_VOP_START; 1617 } 1618 1619 break; 1620 } 1621 1622 case SKIP_TO_VOP_START: 1623 { 1624 if (chunkType == 0xb6) { 1625 int vopCodingType = (data[offset + 4] & 0xc0) >> 6; 1626 1627 offset += chunkSize; 1628 1629 sp<ABuffer> accessUnit = new ABuffer(offset); 1630 memcpy(accessUnit->data(), data, offset); 1631 1632 memmove(data, &data[offset], size - offset); 1633 size -= offset; 1634 mBuffer->setRange(0, size); 1635 1636 int64_t timeUs = fetchTimestamp(offset); 1637 if (timeUs < 0ll) { 1638 ALOGE("Negative timeus"); 1639 return NULL; 1640 } 1641 1642 offset = 0; 1643 1644 accessUnit->meta()->setInt64("timeUs", timeUs); 1645 if (vopCodingType == 0) { // intra-coded VOP 1646 accessUnit->meta()->setInt32("isSync", 1); 1647 } 1648 1649 ALOGV("returning MPEG4 video access unit at time %" PRId64 " us", 1650 timeUs); 1651 1652 // hexdump(accessUnit->data(), accessUnit->size()); 1653 1654 return accessUnit; 1655 } else if (chunkType != 0xb3) { 1656 offset += chunkSize; 1657 discard = true; 1658 } 1659 1660 break; 1661 } 1662 1663 default: 1664 ALOGE("Unknown state: %d", state); 1665 return NULL; 1666 } 1667 1668 if (discard) { 1669 (void)fetchTimestamp(offset); 1670 memmove(data, &data[offset], size - offset); 1671 size -= offset; 1672 offset = 0; 1673 mBuffer->setRange(0, size); 1674 } else { 1675 offset += chunkSize; 1676 } 1677 } 1678 1679 return NULL; 1680 } 1681 1682 void ElementaryStreamQueue::signalEOS() { 1683 if (!mEOSReached) { 1684 if (mMode == MPEG_VIDEO) { 1685 const char *theEnd = "\x00\x00\x01\x00"; 1686 appendData(theEnd, 4, 0); 1687 } 1688 mEOSReached = true; 1689 } else { 1690 ALOGW("EOS already signaled"); 1691 } 1692 } 1693 1694 sp<ABuffer> ElementaryStreamQueue::dequeueAccessUnitMetadata() { 1695 size_t size = mBuffer->size(); 1696 if (!size) { 1697 return NULL; 1698 } 1699 1700 sp<ABuffer> accessUnit = new ABuffer(size); 1701 int64_t timeUs = fetchTimestamp(size); 1702 accessUnit->meta()->setInt64("timeUs", timeUs); 1703 1704 memcpy(accessUnit->data(), mBuffer->data(), size); 1705 mBuffer->setRange(0, 0); 1706 1707 if (mFormat == NULL) { 1708 mFormat = new MetaData; 1709 mFormat->setCString(kKeyMIMEType, MEDIA_MIMETYPE_DATA_TIMED_ID3); 1710 } 1711 1712 return accessUnit; 1713 } 1714 1715 void ElementaryStreamQueue::signalNewSampleAesKey(const sp<AMessage> &keyItem) { 1716 if (mSampleDecryptor == NULL) { 1717 ALOGE("signalNewSampleAesKey: Stream %x is not encrypted; keyItem: %p", 1718 mMode, keyItem.get()); 1719 return; 1720 } 1721 1722 mSampleDecryptor->signalNewSampleAesKey(keyItem); 1723 } 1724 1725 1726 } // namespace android 1727
