Age | Commit message (Collapse) | Author |
|
|
|
|
|
frame display in PreviewRenderer surface after importing video"
|
|
|
|
|
|
|
|
|
|
Add destructor to call fclose.
Change-Id: I765c56236f8df58af998439492b0b955ab57606f
|
|
See I2c0a87b0fc049e855e21f455b806287b578091de
Change-Id: Iec8566bce0a4763a9a17b4a46f15aa39b5bcd141
|
|
|
|
|
|
Remove checks that causes crash for rtsp streamed h.263 content
with certain values in the RTP payload header:
Remove zero check for the five reserved bits in the payload header.
According to RFC 4629 these bits MUST be ignored by receivers.
Remove zero-check for the VRC (Video Redundancy Coding) bit,
skip packet instead.
Remove zero-check for the PLEN bits (extra picture header),
skip packet instead.
Remove zero-check for the PEBIT bits (extra picture header),
skip packet instead.
Remove corresponding zero check for the four resreved bits in the
AMR payload header. According to RFC 4867 these bits MUST be
ignored by receivers.
Change-Id: I7fc21d69a19d23da24f9267623c338d415ef1387
|
|
support stereo/multichannel 24-bit format
kMaxFrameSize is fixed to 32768. When converting 24-bit to 16-bit,
number of samlpes is maxBytesToRead /3.
In this case, if the maxBytesToRead is not multiple of 3,
pcm data is messed when converting.
Bug:7630939
Change-Id: I0ea1b53eb1272a8d83b63815fc0a05b73cef75f1
|
|
|
|
When thumbnails were generated they could be generated at random
angles as the mRotationAngle variable was not initialized to any
value. This variable would have to be explicitly overwritten to not
cause random rotation. Changed the implementation to initialize the
value to 0 (no rotation). mRotationAngle was also missing in the
copy constructor.
Change-Id: I67a5340fdd807c6ab3a3da5eecb09b5b9d5f4666
|
|
|
|
|
|
If the set up of the RTSP stream contains an incorrect or otherwise
problematic URL, some servers will send an unsolicited server response
that contains a negative number in the sequence number (CSeq) field.
This negative value is not returned from the function findPendingRequest(),
so the check in notifyResponseListener() will not work. Instead there will
be a crash when 0 is used as the index to find a matching request/response
pair that doesn’t exist.
The fix is to return the received sequence number also when it is an
unsolicited server-client message.
Change-Id: Iedaba8a63dece7b43bce007069baefbfd10970b8
|
|
For an MP3 source, within the prepare command, ID3 tags are checked in search of
gapless playback info. This causes problems for streamed sources. If ID3v2 tags
aren't present, then a check is done for ID3v1 tags. This results in a read
command that asks the cache to jump to the end of the file, and subsequently
make an extra http call to request those bytes. For a streamed source, this
causes the file to not be downloaded automatically when MediaPlayer::prepare()
is called, and causes stuttering and extra buffering time to be needed when
start() is finally called.
The solution is to ignore the ID3v1 tags as the gapless info would never exist
there, and only check for ID3v2 tags.
Change-Id: I7d1b94cffbfe7c38ca094834dedbc92a58855e20
|
|
|
|
|
|
This fix handles problems with several asynchronous calls
within streaming. This case is when the phone has sent a
request to the server and while the response is being sent
back by the server the request is aborted by the user.
The fix is an if case that checks if we have aborted while
waiting for a response from the server. If we have aborted
we should ignore the late response instead of continuing.
Change-Id: I1264bb992f6abcaee1f10a89479e08b54a95e3c2
|
|
The mIsDrm is a bool that isn't initialized.
This causes it to be true in most default cases.
Change-Id: I41b534514bf6a3ca88a9f0994b814d55fcd7453b
|
|
|
|
|
|
This fix is protecting AudioTrack::dump from SIGSEGEV
when calling dumpsys shell command.
Change-Id: I30d136e195a12d6fdad41e14f557b0ad9e59b8a2
|
|
|
|
Don't duplicate strings when retrieveing metadata from media
files. As any requests for metadata strings would pass through
the binder, this would cause the reference to the duplicate string
to be lost, causing a memory leak as the duplicate would not be
freed.
Change-Id: I2593733472b1bb589bc502b2c11080f581015bb5
|
|
|
|
|
|
If the codec was chosen based on mime type, the caller does
not know what component actually was chosen. This allows
getting essential information (such as supported color formats,
for a video encoder) for this component.
Change-Id: Ie471f40f8104b37d27ced3dba5a54facc6504b1b
|
|
The MPEG4Extractor was parsing 8 bits instead of 16 bits when parsing
'trkn' and 'disk'. Also added support for 16 bytes size 'disk'.
Change-Id: I22b4de2ac800881884d5759776cb380917522a87
|
|
If enabling seq_scaling_matrix_present_flag = 1 the AVCC chunk can
be larger than the original buffer size. Changed to using ABuffer
instead.
Change-Id: Idacc14b45ea2634c5e608919f3ce567f23363135
|
|
Repeated leaks detected in mediaserver context. Most chunks small but
some bigger. Adding up it was not uncommon to see 200Mb footprint.
libc memdebug allocation call stack showed the root cause in "covr"
parsing - that explains why content often was empty but sometimes big.
Change-Id: I0de6c1eeef174d2529973ff9c6a020ec3dd44c75
|
|
MPEG4 audio packets may be multiplexed using the so called
LATM (Low Overhead Audio Transport Multiplex) scheme.
LATM parsing was recently introduced in Stagefright and it
has caused issues in cases when the LATM config element
cannot be parsed correctly. The main problem occurrs when
the AudioSpecificConfig part of the config element contains
more information than what is expected, causing the
frameLengthType parameter to get the wrong value. This fix
introduces default values of some config parameters that are
used in case config parsing fails.
Change-Id: I3cb35df76826f95ca0831dc08c2a1e7c6c2c586d
|
|
|
|
The *.mpga files are playable, but are not correctly scanned.
This is because they are prevented from being scanned in
StagefrightMediaScanner.cpp.
What this fix does is to add the extension to the list of valid
file extensions so that the scanner handles the filetype properly.
We have previously added the .mpga extension to the framework to
make it playable, but not added it so that the scanner scans it.
Change-Id: I02a44d770adb80d38e8bed77d0d76efa1b28a861
|
|
FLACExtractor had artificial limits which did not allow multichannel
audio, 88.2 or 96 kHz samplerate, these were removed since it is up
to the audio subsystem to handle.
Change-Id: Iadc20cdf4724814b5c980827a528a1e1b4bd6a82
|
|
display in PreviewRenderer surface after importing video
1. this is caused mFrameIntervalMs isn't assigned for use-software-codec case
2. without the assingment, some decoded frames will not be copied to the buffer queue of VideoEditorVideoDecoder
Change-Id: I578954e38958ebac8f258300e60b8ddf60e50521
Signed-off-by: beanz <beanz@marvell.com>
|
|
Change-Id: Ic8aedfa821d79698d31cf7ee3bf382aae85981f6
Signed-off-by: beanz <beanz@marvell.com>
|
|
* commit '78b012f5bc05706d1ce412811ccfaa9385dae0ab':
|
|
Change-Id: Idf4c25115d89397ba668fc290344b2e7c1ea1993
|
|
* commit '1c6833afd6cfdbc6c696ed0719ceedc37ec0b534':
Ignore SIGPIPE during write() to broken pipe
|
|
This happens occasionally when taking a bugreport.
Bug: 6447319
Change-Id: Ia6531a4a3658461f8fd3f7106e7996da7cc5933a
|
|
Change-Id: Icad86b92debd9605c745283aeffc3f459a3549ce
|
|
|
|
configurable
* commit '96076964863ee6887f0bed9d0f11f424b48ab9b9':
Make video encoders' bitrate mode (constant, variable, ...) configurable
|
|
through the desired output format. Configure the video encoder to use
constant bitrate mode for WFD.
Change-Id: Id7bd619598153c13448a9c5acd69d80f8a01f333
related-to-bug: 7459597
|
|
Change-Id: I0fd73e8e0b948c2d01fdb9f4c6e005ee9cd2ac44
|
|
|