зеркало из https://github.com/mozilla/gecko-dev.git
557 строки
16 KiB
C++
557 строки
16 KiB
C++
/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
|
|
/* vim:set ts=2 sw=2 sts=2 et cindent: */
|
|
/* This Source Code Form is subject to the terms of the Mozilla Public
|
|
* License, v. 2.0. If a copy of the MPL was not distributed with this
|
|
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
|
|
|
|
#include <algorithm>
|
|
|
|
#include "nsMemory.h"
|
|
#include "MP3FrameParser.h"
|
|
#include "VideoUtils.h"
|
|
|
|
|
|
#define FROM_BIG_ENDIAN(X) ((uint32_t)((uint8_t)(X)[0] << 24 | (uint8_t)(X)[1] << 16 | \
|
|
(uint8_t)(X)[2] << 8 | (uint8_t)(X)[3]))
|
|
|
|
|
|
namespace mozilla {
|
|
|
|
/*
|
|
* Following code taken from http://www.hydrogenaudio.org/forums/index.php?showtopic=85125
|
|
* with permission from the author, Nick Wallette <sirnickity@gmail.com>.
|
|
*/
|
|
|
|
/* BEGIN shameless copy and paste */
|
|
|
|
// Bitrates - use [version][layer][bitrate]
|
|
const uint16_t mpeg_bitrates[4][4][16] = {
|
|
{ // Version 2.5
|
|
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, // Reserved
|
|
{ 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 0 }, // Layer 3
|
|
{ 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 0 }, // Layer 2
|
|
{ 0, 32, 48, 56, 64, 80, 96, 112, 128, 144, 160, 176, 192, 224, 256, 0 } // Layer 1
|
|
},
|
|
{ // Reserved
|
|
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, // Invalid
|
|
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, // Invalid
|
|
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, // Invalid
|
|
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 } // Invalid
|
|
},
|
|
{ // Version 2
|
|
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, // Reserved
|
|
{ 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 0 }, // Layer 3
|
|
{ 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 0 }, // Layer 2
|
|
{ 0, 32, 48, 56, 64, 80, 96, 112, 128, 144, 160, 176, 192, 224, 256, 0 } // Layer 1
|
|
},
|
|
{ // Version 1
|
|
{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, // Reserved
|
|
{ 0, 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, 320, 0 }, // Layer 3
|
|
{ 0, 32, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, 320, 384, 0 }, // Layer 2
|
|
{ 0, 32, 64, 96, 128, 160, 192, 224, 256, 288, 320, 352, 384, 416, 448, 0 }, // Layer 1
|
|
}
|
|
};
|
|
|
|
// Sample rates - use [version][srate]
|
|
const uint16_t mpeg_srates[4][4] = {
|
|
{ 11025, 12000, 8000, 0 }, // MPEG 2.5
|
|
{ 0, 0, 0, 0 }, // Reserved
|
|
{ 22050, 24000, 16000, 0 }, // MPEG 2
|
|
{ 44100, 48000, 32000, 0 } // MPEG 1
|
|
};
|
|
|
|
// Samples per frame - use [version][layer]
|
|
const uint16_t mpeg_frame_samples[4][4] = {
|
|
// Rsvd 3 2 1 < Layer v Version
|
|
{ 0, 576, 1152, 384 }, // 2.5
|
|
{ 0, 0, 0, 0 }, // Reserved
|
|
{ 0, 576, 1152, 384 }, // 2
|
|
{ 0, 1152, 1152, 384 } // 1
|
|
};
|
|
|
|
// Slot size (MPEG unit of measurement) - use [layer]
|
|
const uint8_t mpeg_slot_size[4] = { 0, 1, 1, 4 }; // Rsvd, 3, 2, 1
|
|
|
|
uint16_t
|
|
MP3Frame::CalculateLength()
|
|
{
|
|
// Lookup real values of these fields
|
|
uint32_t bitrate = mpeg_bitrates[mVersion][mLayer][mBitrate] * 1000;
|
|
uint32_t samprate = mpeg_srates[mVersion][mSampleRate];
|
|
uint16_t samples = mpeg_frame_samples[mVersion][mLayer];
|
|
uint8_t slot_size = mpeg_slot_size[mLayer];
|
|
|
|
// In-between calculations
|
|
float bps = (float)samples / 8.0;
|
|
float fsize = ( (bps * (float)bitrate) / (float)samprate )
|
|
+ ( (mPad) ? slot_size : 0 );
|
|
|
|
// Frame sizes are truncated integers
|
|
return (uint16_t)fsize;
|
|
}
|
|
|
|
/* END shameless copy and paste */
|
|
|
|
|
|
/** MP3Parser methods **/
|
|
|
|
MP3Parser::MP3Parser()
|
|
: mCurrentChar(0)
|
|
{ }
|
|
|
|
void
|
|
MP3Parser::Reset()
|
|
{
|
|
mCurrentChar = 0;
|
|
}
|
|
|
|
uint16_t
|
|
MP3Parser::ParseFrameLength(uint8_t ch)
|
|
{
|
|
mData.mRaw[mCurrentChar] = ch;
|
|
|
|
MP3Frame &frame = mData.mFrame;
|
|
|
|
// Validate MP3 header as we read. We can't mistake the start of an MP3 frame
|
|
// for the middle of another frame due to the sync byte at the beginning
|
|
// of the frame.
|
|
|
|
// The only valid position for an all-high byte is the sync byte at the
|
|
// beginning of the frame.
|
|
if (ch == 0xff) {
|
|
mCurrentChar = 0;
|
|
}
|
|
|
|
// Make sure the current byte is valid in context. If not, reset the parser.
|
|
if (mCurrentChar == 2) {
|
|
if (frame.mBitrate == 0x0f) {
|
|
goto fail;
|
|
}
|
|
} else if (mCurrentChar == 1) {
|
|
if (frame.mSync2 != 0x07
|
|
|| frame.mVersion == 0x01
|
|
|| frame.mLayer == 0x00) {
|
|
goto fail;
|
|
}
|
|
}
|
|
|
|
// The only valid character at the beginning of the header is 0xff. Fail if
|
|
// it's different.
|
|
if (mCurrentChar == 0 && frame.mSync1 != 0xff) {
|
|
// Couldn't find the sync byte. Fail.
|
|
return 0;
|
|
}
|
|
|
|
mCurrentChar++;
|
|
MOZ_ASSERT(mCurrentChar <= sizeof(MP3Frame));
|
|
|
|
// Don't have a full header yet.
|
|
if (mCurrentChar < sizeof(MP3Frame)) {
|
|
return 0;
|
|
}
|
|
|
|
// Woo, valid header. Return the length.
|
|
mCurrentChar = 0;
|
|
return frame.CalculateLength();
|
|
|
|
fail:
|
|
Reset();
|
|
return 0;
|
|
}
|
|
|
|
uint32_t
|
|
MP3Parser::GetSampleRate()
|
|
{
|
|
MP3Frame &frame = mData.mFrame;
|
|
return mpeg_srates[frame.mVersion][frame.mSampleRate];
|
|
}
|
|
|
|
uint32_t
|
|
MP3Parser::GetSamplesPerFrame()
|
|
{
|
|
MP3Frame &frame = mData.mFrame;
|
|
return mpeg_frame_samples[frame.mVersion][frame.mLayer];
|
|
}
|
|
|
|
|
|
/** ID3Parser methods **/
|
|
|
|
const char sID3Head[3] = { 'I', 'D', '3' };
|
|
const uint32_t ID3_HEADER_LENGTH = 10;
|
|
|
|
ID3Parser::ID3Parser()
|
|
: mCurrentChar(0)
|
|
, mHeaderLength(0)
|
|
{ }
|
|
|
|
void
|
|
ID3Parser::Reset()
|
|
{
|
|
mCurrentChar = mHeaderLength = 0;
|
|
}
|
|
|
|
bool
|
|
ID3Parser::ParseChar(char ch)
|
|
{
|
|
// First three bytes of an ID3v2 header must match the string "ID3".
|
|
if (mCurrentChar < sizeof(sID3Head) / sizeof(*sID3Head)
|
|
&& ch != sID3Head[mCurrentChar]) {
|
|
goto fail;
|
|
}
|
|
|
|
// The last four bytes of the header is a 28-bit unsigned integer with the
|
|
// high bit of each byte unset.
|
|
if (mCurrentChar >= 6 && mCurrentChar < ID3_HEADER_LENGTH) {
|
|
if (ch & 0x80) {
|
|
goto fail;
|
|
} else {
|
|
mHeaderLength <<= 7;
|
|
mHeaderLength |= ch;
|
|
}
|
|
}
|
|
|
|
mCurrentChar++;
|
|
|
|
return IsParsed();
|
|
|
|
fail:
|
|
Reset();
|
|
return false;
|
|
}
|
|
|
|
bool
|
|
ID3Parser::IsParsed() const
|
|
{
|
|
return mCurrentChar >= ID3_HEADER_LENGTH;
|
|
}
|
|
|
|
uint32_t
|
|
ID3Parser::GetHeaderLength() const
|
|
{
|
|
MOZ_ASSERT(IsParsed(),
|
|
"Queried length of ID3 header before parsing finished.");
|
|
return mHeaderLength;
|
|
}
|
|
|
|
|
|
/** VBR header helper stuff **/
|
|
|
|
// Helper function to find a VBR header in an MP3 frame.
|
|
// Based on information from
|
|
// http://www.codeproject.com/Articles/8295/MPEG-Audio-Frame-Header
|
|
|
|
const uint32_t VBRI_TAG = FROM_BIG_ENDIAN("VBRI");
|
|
const uint32_t VBRI_OFFSET = 32 - sizeof(MP3Frame);
|
|
const uint32_t VBRI_FRAME_COUNT_OFFSET = VBRI_OFFSET + 14;
|
|
const uint32_t VBRI_MIN_FRAME_SIZE = VBRI_OFFSET + 26;
|
|
|
|
const uint32_t XING_TAG = FROM_BIG_ENDIAN("Xing");
|
|
enum XingFlags {
|
|
XING_HAS_NUM_FRAMES = 0x01,
|
|
XING_HAS_NUM_BYTES = 0x02,
|
|
XING_HAS_TOC = 0x04,
|
|
XING_HAS_VBR_SCALE = 0x08
|
|
};
|
|
|
|
static int64_t
|
|
ParseXing(const char *aBuffer)
|
|
{
|
|
uint32_t flags = FROM_BIG_ENDIAN(aBuffer + 4);
|
|
|
|
if (!(flags & XING_HAS_NUM_FRAMES)) {
|
|
NS_WARNING("VBR file without frame count. Duration estimation likely to "
|
|
"be totally wrong.");
|
|
return -1;
|
|
}
|
|
|
|
int64_t numFrames = -1;
|
|
if (flags & XING_HAS_NUM_FRAMES) {
|
|
numFrames = FROM_BIG_ENDIAN(aBuffer + 8);
|
|
}
|
|
|
|
return numFrames;
|
|
}
|
|
|
|
static int64_t
|
|
FindNumVBRFrames(const nsAutoCString& aFrame)
|
|
{
|
|
const char *buffer = aFrame.get();
|
|
const char *bufferEnd = aFrame.get() + aFrame.Length();
|
|
|
|
// VBRI header is nice and well-defined; let's try to find that first.
|
|
if (aFrame.Length() > VBRI_MIN_FRAME_SIZE &&
|
|
FROM_BIG_ENDIAN(buffer + VBRI_OFFSET) == VBRI_TAG) {
|
|
return FROM_BIG_ENDIAN(buffer + VBRI_FRAME_COUNT_OFFSET);
|
|
}
|
|
|
|
// We have to search for the Xing header as its position can change.
|
|
for (; buffer + sizeof(XING_TAG) < bufferEnd; buffer++) {
|
|
if (FROM_BIG_ENDIAN(buffer) == XING_TAG) {
|
|
return ParseXing(buffer);
|
|
}
|
|
}
|
|
|
|
return -1;
|
|
}
|
|
|
|
|
|
/** MP3FrameParser methods **/
|
|
|
|
// Some MP3's have large ID3v2 tags, up to 150KB, so we allow lots of
|
|
// skipped bytes to be read, just in case, before we give up and assume
|
|
// we're not parsing an MP3 stream.
|
|
static const uint32_t MAX_SKIPPED_BYTES = 4096;
|
|
|
|
enum {
|
|
MP3_HEADER_LENGTH = 4,
|
|
};
|
|
|
|
MP3FrameParser::MP3FrameParser(int64_t aLength)
|
|
: mLock("MP3FrameParser.mLock"),
|
|
mTotalID3Size(0),
|
|
mTotalFrameSize(0),
|
|
mFrameCount(0),
|
|
mOffset(0),
|
|
mLength(aLength),
|
|
mMP3Offset(-1),
|
|
mSamplesPerSecond(0),
|
|
mFirstFrameEnd(-1),
|
|
mIsMP3(MAYBE_MP3)
|
|
{ }
|
|
|
|
nsresult MP3FrameParser::ParseBuffer(const uint8_t* aBuffer,
|
|
uint32_t aLength,
|
|
int64_t aStreamOffset,
|
|
uint32_t* aOutBytesRead)
|
|
{
|
|
// Iterate forwards over the buffer, looking for ID3 tag, or MP3
|
|
// Frame headers.
|
|
const uint8_t *buffer = aBuffer;
|
|
const uint8_t *bufferEnd = aBuffer + aLength;
|
|
|
|
// If we haven't found any MP3 frame data yet, there might be ID3 headers
|
|
// we can skip over.
|
|
if (mMP3Offset < 0) {
|
|
for (const uint8_t *ch = buffer; ch < bufferEnd; ch++) {
|
|
if (mID3Parser.ParseChar(*ch)) {
|
|
// Found an ID3 header. We don't care about the body of the header, so
|
|
// just skip past.
|
|
buffer = ch + mID3Parser.GetHeaderLength() - (ID3_HEADER_LENGTH - 1);
|
|
|
|
if (buffer <= ch) {
|
|
return NS_ERROR_FAILURE;
|
|
}
|
|
|
|
ch = buffer;
|
|
|
|
mTotalID3Size += mID3Parser.GetHeaderLength();
|
|
|
|
// Yes, this is an MP3!
|
|
mIsMP3 = DEFINITELY_MP3;
|
|
|
|
mID3Parser.Reset();
|
|
}
|
|
}
|
|
}
|
|
|
|
// The first MP3 frame in a variable bitrate stream can contain metadata
|
|
// for duration estimation and seeking, so we buffer that first frame here.
|
|
if (aStreamOffset < mFirstFrameEnd) {
|
|
uint64_t copyLen = std::min((int64_t)aLength, mFirstFrameEnd - aStreamOffset);
|
|
mFirstFrame.Append((const char *)buffer, copyLen);
|
|
buffer += copyLen;
|
|
}
|
|
|
|
while (buffer < bufferEnd) {
|
|
uint16_t frameLen = mMP3Parser.ParseFrameLength(*buffer);
|
|
|
|
if (frameLen) {
|
|
// We've found an MP3 frame!
|
|
// This is the first frame (and the only one we'll bother parsing), so:
|
|
// * Mark this stream as MP3;
|
|
// * Store the offset at which the MP3 data started; and
|
|
// * Start buffering the frame, as it might contain handy metadata.
|
|
|
|
// We're now sure this is an MP3 stream.
|
|
mIsMP3 = DEFINITELY_MP3;
|
|
|
|
// We need to know these to convert the number of frames in the stream
|
|
// to the length of the stream in seconds.
|
|
mSamplesPerSecond = mMP3Parser.GetSampleRate();
|
|
mSamplesPerFrame = mMP3Parser.GetSamplesPerFrame();
|
|
|
|
// If the stream has a constant bitrate, we should only need the length
|
|
// of the first frame and the length (in bytes) of the stream to
|
|
// estimate the length (in seconds).
|
|
mTotalFrameSize += frameLen;
|
|
mFrameCount++;
|
|
|
|
// If |mMP3Offset| isn't set then this is the first MP3 frame we have
|
|
// seen in the stream, which is useful for duration estimation.
|
|
if (mMP3Offset > -1) {
|
|
uint16_t skip = frameLen - sizeof(MP3Frame);
|
|
buffer += skip ? skip : 1;
|
|
continue;
|
|
}
|
|
|
|
// Remember the offset of the MP3 stream.
|
|
// We're at the last byte of an MP3Frame, so MP3 data started
|
|
// sizeof(MP3Frame) - 1 bytes ago.
|
|
mMP3Offset = aStreamOffset
|
|
+ (buffer - aBuffer)
|
|
- (sizeof(MP3Frame) - 1);
|
|
|
|
buffer++;
|
|
|
|
// If the stream has a variable bitrate, the first frame has metadata
|
|
// we need for duration estimation and seeking. Start buffering it so we
|
|
// can parse it later.
|
|
mFirstFrameEnd = mMP3Offset + frameLen;
|
|
uint64_t currOffset = buffer - aBuffer + aStreamOffset;
|
|
uint64_t copyLen = std::min(mFirstFrameEnd - currOffset,
|
|
(uint64_t)(bufferEnd - buffer));
|
|
mFirstFrame.Append((const char *)buffer, copyLen);
|
|
|
|
buffer += copyLen;
|
|
|
|
} else {
|
|
// Nothing to see here. Move along.
|
|
buffer++;
|
|
}
|
|
}
|
|
|
|
*aOutBytesRead = buffer - aBuffer;
|
|
|
|
if (mFirstFrameEnd > -1 && mFirstFrameEnd <= aStreamOffset + buffer - aBuffer) {
|
|
// We have our whole first frame. Try to find a VBR header.
|
|
mNumFrames = FindNumVBRFrames(mFirstFrame);
|
|
mFirstFrameEnd = -1;
|
|
}
|
|
|
|
return NS_OK;
|
|
}
|
|
|
|
void MP3FrameParser::Parse(const char* aBuffer, uint32_t aLength, uint64_t aOffset)
|
|
{
|
|
MutexAutoLock mon(mLock);
|
|
|
|
if (HasExactDuration()) {
|
|
// We know the duration; nothing to do here.
|
|
return;
|
|
}
|
|
|
|
const uint8_t* buffer = reinterpret_cast<const uint8_t*>(aBuffer);
|
|
int32_t length = aLength;
|
|
uint64_t offset = aOffset;
|
|
|
|
// Got some data we have seen already. Skip forward to what we need.
|
|
if (aOffset < mOffset) {
|
|
buffer += mOffset - aOffset;
|
|
length -= mOffset - aOffset;
|
|
offset = mOffset;
|
|
|
|
if (length <= 0) {
|
|
return;
|
|
}
|
|
}
|
|
|
|
// If there is a discontinuity in the input stream, reset the state of the
|
|
// parsers so we don't get any partial headers.
|
|
if (mOffset < aOffset) {
|
|
if (!mID3Parser.IsParsed()) {
|
|
// Only reset this if it hasn't finished yet.
|
|
mID3Parser.Reset();
|
|
}
|
|
|
|
if (mFirstFrameEnd > -1) {
|
|
NS_WARNING("Discontinuity in input while buffering first frame.");
|
|
mFirstFrameEnd = -1;
|
|
}
|
|
|
|
mMP3Parser.Reset();
|
|
}
|
|
|
|
uint32_t bytesRead = 0;
|
|
if (NS_FAILED(ParseBuffer(buffer,
|
|
length,
|
|
offset,
|
|
&bytesRead))) {
|
|
return;
|
|
}
|
|
|
|
MOZ_ASSERT(length <= (int)bytesRead, "All bytes should have been consumed");
|
|
|
|
// Update next data offset
|
|
mOffset = offset + bytesRead;
|
|
|
|
// If we've parsed lots of data and we still have nothing, just give up.
|
|
// We don't count ID3 headers towards the skipped bytes count, as MP3 files
|
|
// can have massive ID3 sections.
|
|
if (!mID3Parser.IsParsed() && mMP3Offset < 0 &&
|
|
mOffset - mTotalID3Size > MAX_SKIPPED_BYTES) {
|
|
mIsMP3 = NOT_MP3;
|
|
}
|
|
}
|
|
|
|
int64_t MP3FrameParser::GetDuration()
|
|
{
|
|
MutexAutoLock mon(mLock);
|
|
|
|
if (!ParsedHeaders() || !mSamplesPerSecond) {
|
|
// Not a single frame decoded yet.
|
|
return -1;
|
|
}
|
|
|
|
MOZ_ASSERT(mFrameCount > 0 && mTotalFrameSize > 0,
|
|
"Frame parser should have seen at least one MP3 frame of positive length.");
|
|
|
|
if (!mFrameCount || !mTotalFrameSize) {
|
|
// This should never happen.
|
|
return -1;
|
|
}
|
|
|
|
double frames;
|
|
if (mNumFrames < 0) {
|
|
// Estimate the number of frames in the stream based on the average frame
|
|
// size and the length of the MP3 file.
|
|
double frameSize = (double)mTotalFrameSize / mFrameCount;
|
|
frames = (double)(mLength - mMP3Offset) / frameSize;
|
|
} else {
|
|
// We know the exact number of frames from the VBR header.
|
|
frames = mNumFrames;
|
|
}
|
|
|
|
// The duration of each frame is constant over a given stream.
|
|
double usPerFrame = USECS_PER_S * mSamplesPerFrame / mSamplesPerSecond;
|
|
|
|
return frames * usPerFrame;
|
|
}
|
|
|
|
int64_t MP3FrameParser::GetMP3Offset()
|
|
{
|
|
MutexAutoLock mon(mLock);
|
|
return mMP3Offset;
|
|
}
|
|
|
|
bool MP3FrameParser::ParsedHeaders()
|
|
{
|
|
// We have seen both the beginning and the end of the first MP3 frame in the
|
|
// stream.
|
|
return mMP3Offset > -1 && mFirstFrameEnd < 0;
|
|
}
|
|
|
|
bool MP3FrameParser::HasExactDuration()
|
|
{
|
|
return ParsedHeaders() && mNumFrames > -1;
|
|
}
|
|
|
|
bool MP3FrameParser::NeedsData()
|
|
{
|
|
// If we don't know the duration exactly then either:
|
|
// - we're still waiting for a VBR header; or
|
|
// - we look at all frames to constantly update our duration estimate.
|
|
return IsMP3() && !HasExactDuration();
|
|
}
|
|
|
|
}
|