gecko/content/media/MediaData.cpp
Chris Pearce b26d467d58 Bug 778077 - Implement HTMLMediaElement.fastSeek(time). r=cajbir
Implement HTMLMediaElement.fastSeek(), basically by changing all the
MediaDecoderReader::Seek() overrides to not call
MediaDecoderReader::DecodeToTarget(), and have MediaDecoderReader::DecodeSeek()
call DecodeToTarget() if we're doing an accurate (non-fast) seek.

Update gizmo.mp4 to have a keyframe every second, instead of only 1 keyframe at
the start of stream. This makes the unit test I added more useful for mp4...

I pushed most of the seek target clamping logic in MediaDecoder up into
HTMLMediaElement, so that we're clamping in fewer places. Note
MediaDecoderStateMachine::Seek() still sanity checks the seek target.

We have to update the currentTime/MediaDecoder playback position after a seek
completes now, rather than assuming the seek always got it exactly right.

Removed those pesky assertions about seek target lying in the first frame after
seek, since actually sometimes the media doesn't have samples for all streams
after a seek (either due to the media being encoded like that, or because of a
bug in the platform's decoder, not entirely sure).

Green: https://tbpl.mozilla.org/?tree=Try&rev=b028258565e2
2014-03-28 15:50:28 +13:00

409 lines
14 KiB
C++

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* 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 "MediaData.h"
#include "MediaInfo.h"
#ifdef MOZ_OMX_DECODER
#include "GrallocImages.h"
#endif
#include "VideoUtils.h"
#include "ImageContainer.h"
namespace mozilla {
using namespace mozilla::gfx;
using layers::ImageContainer;
using layers::PlanarYCbCrImage;
using layers::PlanarYCbCrData;
void
AudioData::EnsureAudioBuffer()
{
if (mAudioBuffer)
return;
mAudioBuffer = SharedBuffer::Create(mFrames*mChannels*sizeof(AudioDataValue));
AudioDataValue* data = static_cast<AudioDataValue*>(mAudioBuffer->Data());
for (uint32_t i = 0; i < mFrames; ++i) {
for (uint32_t j = 0; j < mChannels; ++j) {
data[j*mFrames + i] = mAudioData[i*mChannels + j];
}
}
}
static bool
ValidatePlane(const VideoData::YCbCrBuffer::Plane& aPlane)
{
return aPlane.mWidth <= PlanarYCbCrImage::MAX_DIMENSION &&
aPlane.mHeight <= PlanarYCbCrImage::MAX_DIMENSION &&
aPlane.mWidth * aPlane.mHeight < MAX_VIDEO_WIDTH * MAX_VIDEO_HEIGHT &&
aPlane.mStride > 0;
}
#ifdef MOZ_WIDGET_GONK
static bool
IsYV12Format(const VideoData::YCbCrBuffer::Plane& aYPlane,
const VideoData::YCbCrBuffer::Plane& aCbPlane,
const VideoData::YCbCrBuffer::Plane& aCrPlane)
{
return
aYPlane.mWidth % 2 == 0 &&
aYPlane.mHeight % 2 == 0 &&
aYPlane.mWidth / 2 == aCbPlane.mWidth &&
aYPlane.mHeight / 2 == aCbPlane.mHeight &&
aCbPlane.mWidth == aCrPlane.mWidth &&
aCbPlane.mHeight == aCrPlane.mHeight;
}
#endif
VideoData::VideoData(int64_t aOffset, int64_t aTime, int64_t aDuration, int64_t aTimecode)
: MediaData(VIDEO_FRAME, aOffset, aTime, aDuration),
mTimecode(aTimecode),
mDuplicate(true),
mKeyframe(false)
{
MOZ_COUNT_CTOR(VideoData);
NS_ASSERTION(mDuration >= 0, "Frame must have non-negative duration.");
}
VideoData::VideoData(int64_t aOffset,
int64_t aTime,
int64_t aDuration,
bool aKeyframe,
int64_t aTimecode,
IntSize aDisplay)
: MediaData(VIDEO_FRAME, aOffset, aTime, aDuration),
mDisplay(aDisplay),
mTimecode(aTimecode),
mDuplicate(false),
mKeyframe(aKeyframe)
{
MOZ_COUNT_CTOR(VideoData);
NS_ASSERTION(mDuration >= 0, "Frame must have non-negative duration.");
}
VideoData::~VideoData()
{
MOZ_COUNT_DTOR(VideoData);
}
size_t
VideoData::SizeOfIncludingThis(MallocSizeOf aMallocSizeOf) const
{
size_t size = aMallocSizeOf(this);
// Currently only PLANAR_YCBCR has a well defined function for determining
// it's size, so reporting is limited to that type.
if (mImage && mImage->GetFormat() == ImageFormat::PLANAR_YCBCR) {
const mozilla::layers::PlanarYCbCrImage* img =
static_cast<const mozilla::layers::PlanarYCbCrImage*>(mImage.get());
size += img->SizeOfIncludingThis(aMallocSizeOf);
}
return size;
}
/* static */
VideoData* VideoData::ShallowCopyUpdateDuration(VideoData* aOther,
int64_t aDuration)
{
VideoData* v = new VideoData(aOther->mOffset,
aOther->mTime,
aDuration,
aOther->mKeyframe,
aOther->mTimecode,
aOther->mDisplay);
v->mImage = aOther->mImage;
return v;
}
/* static */
VideoData* VideoData::ShallowCopyUpdateTimestamp(VideoData* aOther,
int64_t aTimestamp)
{
NS_ENSURE_TRUE(aOther, nullptr);
VideoData* v = new VideoData(aOther->mOffset,
aTimestamp,
aOther->GetEndTime() - aTimestamp,
aOther->mKeyframe,
aOther->mTimecode,
aOther->mDisplay);
v->mImage = aOther->mImage;
return v;
}
/* static */
void VideoData::SetVideoDataToImage(PlanarYCbCrImage* aVideoImage,
VideoInfo& aInfo,
const YCbCrBuffer &aBuffer,
const IntRect& aPicture,
bool aCopyData)
{
if (!aVideoImage) {
return;
}
const YCbCrBuffer::Plane &Y = aBuffer.mPlanes[0];
const YCbCrBuffer::Plane &Cb = aBuffer.mPlanes[1];
const YCbCrBuffer::Plane &Cr = aBuffer.mPlanes[2];
PlanarYCbCrData data;
data.mYChannel = Y.mData + Y.mOffset;
data.mYSize = IntSize(Y.mWidth, Y.mHeight);
data.mYStride = Y.mStride;
data.mYSkip = Y.mSkip;
data.mCbChannel = Cb.mData + Cb.mOffset;
data.mCrChannel = Cr.mData + Cr.mOffset;
data.mCbCrSize = IntSize(Cb.mWidth, Cb.mHeight);
data.mCbCrStride = Cb.mStride;
data.mCbSkip = Cb.mSkip;
data.mCrSkip = Cr.mSkip;
data.mPicX = aPicture.x;
data.mPicY = aPicture.y;
data.mPicSize = aPicture.Size();
data.mStereoMode = aInfo.mStereoMode;
aVideoImage->SetDelayedConversion(true);
if (aCopyData) {
aVideoImage->SetData(data);
} else {
aVideoImage->SetDataNoCopy(data);
}
}
/* static */
VideoData* VideoData::Create(VideoInfo& aInfo,
ImageContainer* aContainer,
Image* aImage,
int64_t aOffset,
int64_t aTime,
int64_t aDuration,
const YCbCrBuffer& aBuffer,
bool aKeyframe,
int64_t aTimecode,
const IntRect& aPicture)
{
if (!aImage && !aContainer) {
// Create a dummy VideoData with no image. This gives us something to
// send to media streams if necessary.
nsAutoPtr<VideoData> v(new VideoData(aOffset,
aTime,
aDuration,
aKeyframe,
aTimecode,
aInfo.mDisplay.ToIntSize()));
return v.forget();
}
// The following situation should never happen unless there is a bug
// in the decoder
if (aBuffer.mPlanes[1].mWidth != aBuffer.mPlanes[2].mWidth ||
aBuffer.mPlanes[1].mHeight != aBuffer.mPlanes[2].mHeight) {
NS_ERROR("C planes with different sizes");
return nullptr;
}
// The following situations could be triggered by invalid input
if (aPicture.width <= 0 || aPicture.height <= 0) {
NS_WARNING("Empty picture rect");
return nullptr;
}
if (!ValidatePlane(aBuffer.mPlanes[0]) || !ValidatePlane(aBuffer.mPlanes[1]) ||
!ValidatePlane(aBuffer.mPlanes[2])) {
NS_WARNING("Invalid plane size");
return nullptr;
}
// Ensure the picture size specified in the headers can be extracted out of
// the frame we've been supplied without indexing out of bounds.
CheckedUint32 xLimit = aPicture.x + CheckedUint32(aPicture.width);
CheckedUint32 yLimit = aPicture.y + CheckedUint32(aPicture.height);
if (!xLimit.isValid() || xLimit.value() > aBuffer.mPlanes[0].mStride ||
!yLimit.isValid() || yLimit.value() > aBuffer.mPlanes[0].mHeight)
{
// The specified picture dimensions can't be contained inside the video
// frame, we'll stomp memory if we try to copy it. Fail.
NS_WARNING("Overflowing picture rect");
return nullptr;
}
nsAutoPtr<VideoData> v(new VideoData(aOffset,
aTime,
aDuration,
aKeyframe,
aTimecode,
aInfo.mDisplay.ToIntSize()));
#ifdef MOZ_WIDGET_GONK
const YCbCrBuffer::Plane &Y = aBuffer.mPlanes[0];
const YCbCrBuffer::Plane &Cb = aBuffer.mPlanes[1];
const YCbCrBuffer::Plane &Cr = aBuffer.mPlanes[2];
#endif
if (!aImage) {
// Currently our decoder only knows how to output to ImageFormat::PLANAR_YCBCR
// format.
#ifdef MOZ_WIDGET_GONK
if (IsYV12Format(Y, Cb, Cr)) {
v->mImage = aContainer->CreateImage(ImageFormat::GRALLOC_PLANAR_YCBCR);
}
#endif
if (!v->mImage) {
v->mImage = aContainer->CreateImage(ImageFormat::PLANAR_YCBCR);
}
} else {
v->mImage = aImage;
}
if (!v->mImage) {
return nullptr;
}
NS_ASSERTION(v->mImage->GetFormat() == ImageFormat::PLANAR_YCBCR ||
v->mImage->GetFormat() == ImageFormat::GRALLOC_PLANAR_YCBCR,
"Wrong format?");
PlanarYCbCrImage* videoImage = static_cast<PlanarYCbCrImage*>(v->mImage.get());
if (!aImage) {
VideoData::SetVideoDataToImage(videoImage, aInfo, aBuffer, aPicture,
true /* aCopyData */);
} else {
VideoData::SetVideoDataToImage(videoImage, aInfo, aBuffer, aPicture,
false /* aCopyData */);
}
#ifdef MOZ_WIDGET_GONK
if (!videoImage->IsValid() && !aImage && IsYV12Format(Y, Cb, Cr)) {
// Failed to allocate gralloc. Try fallback.
v->mImage = aContainer->CreateImage(ImageFormat::PLANAR_YCBCR);
if (!v->mImage) {
return nullptr;
}
videoImage = static_cast<PlanarYCbCrImage*>(v->mImage.get());
VideoData::SetVideoDataToImage(videoImage, aInfo, aBuffer, aPicture,
true /* aCopyData */);
}
#endif
return v.forget();
}
/* static */
VideoData* VideoData::Create(VideoInfo& aInfo,
ImageContainer* aContainer,
int64_t aOffset,
int64_t aTime,
int64_t aDuration,
const YCbCrBuffer& aBuffer,
bool aKeyframe,
int64_t aTimecode,
const IntRect& aPicture)
{
return Create(aInfo, aContainer, nullptr, aOffset, aTime, aDuration, aBuffer,
aKeyframe, aTimecode, aPicture);
}
/* static */
VideoData* VideoData::Create(VideoInfo& aInfo,
Image* aImage,
int64_t aOffset,
int64_t aTime,
int64_t aDuration,
const YCbCrBuffer& aBuffer,
bool aKeyframe,
int64_t aTimecode,
const IntRect& aPicture)
{
return Create(aInfo, nullptr, aImage, aOffset, aTime, aDuration, aBuffer,
aKeyframe, aTimecode, aPicture);
}
/* static */
VideoData* VideoData::CreateFromImage(VideoInfo& aInfo,
ImageContainer* aContainer,
int64_t aOffset,
int64_t aTime,
int64_t aDuration,
const nsRefPtr<Image>& aImage,
bool aKeyframe,
int64_t aTimecode,
const IntRect& aPicture)
{
nsAutoPtr<VideoData> v(new VideoData(aOffset,
aTime,
aDuration,
aKeyframe,
aTimecode,
aInfo.mDisplay.ToIntSize()));
v->mImage = aImage;
return v.forget();
}
#ifdef MOZ_OMX_DECODER
/* static */
VideoData* VideoData::Create(VideoInfo& aInfo,
ImageContainer* aContainer,
int64_t aOffset,
int64_t aTime,
int64_t aDuration,
mozilla::layers::GraphicBufferLocked* aBuffer,
bool aKeyframe,
int64_t aTimecode,
const IntRect& aPicture)
{
if (!aContainer) {
// Create a dummy VideoData with no image. This gives us something to
// send to media streams if necessary.
nsAutoPtr<VideoData> v(new VideoData(aOffset,
aTime,
aDuration,
aKeyframe,
aTimecode,
aInfo.mDisplay.ToIntSize()));
return v.forget();
}
// The following situations could be triggered by invalid input
if (aPicture.width <= 0 || aPicture.height <= 0) {
NS_WARNING("Empty picture rect");
return nullptr;
}
// Ensure the picture size specified in the headers can be extracted out of
// the frame we've been supplied without indexing out of bounds.
CheckedUint32 xLimit = aPicture.x + CheckedUint32(aPicture.width);
CheckedUint32 yLimit = aPicture.y + CheckedUint32(aPicture.height);
if (!xLimit.isValid() || !yLimit.isValid())
{
// The specified picture dimensions can't be contained inside the video
// frame, we'll stomp memory if we try to copy it. Fail.
NS_WARNING("Overflowing picture rect");
return nullptr;
}
nsAutoPtr<VideoData> v(new VideoData(aOffset,
aTime,
aDuration,
aKeyframe,
aTimecode,
aInfo.mDisplay.ToIntSize()));
v->mImage = aContainer->CreateImage(ImageFormat::GRALLOC_PLANAR_YCBCR);
if (!v->mImage) {
return nullptr;
}
NS_ASSERTION(v->mImage->GetFormat() == ImageFormat::GRALLOC_PLANAR_YCBCR,
"Wrong format?");
typedef mozilla::layers::GrallocImage GrallocImage;
GrallocImage* videoImage = static_cast<GrallocImage*>(v->mImage.get());
GrallocImage::GrallocData data;
data.mPicSize = aPicture.Size();
data.mGraphicBuffer = aBuffer;
videoImage->SetData(data);
return v.forget();
}
#endif // MOZ_OMX_DECODER
} // namespace mozilla