| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
'addNativeJarLibsJoglCfg(..)' to GlueGen, commit c0ead6fa10280f8076704726d59f482b183fd77e
|
|
|
|
| |
purposes .. (ATI fglrx driver issues)
|
|
|
|
|
|
|
|
|
| |
may lead to a JVM freeze ..
.. on ATI fglrx driver 32bit on 64bit w/ a frozen shared GL context involved.
Hence we have to rely on the driver cleanup when JVM hits 'exit',
equal to the Windows implementation.
|
|
|
|
| |
(cosmetic change only); Typo in comment; TestSharedContextListNEWT2: Stop animator.
|
|
|
|
|
|
|
|
| |
Relying on junit 4.11 + hamcrest 1.2 (deployed in Gluegen junit.jar)
See gluegen commits:
- 10a28c860bd3f4ccecef156c9bbbc88087bf568d
- 70c9ba258c5ba3d78f0f52faf37d88b4c62d0c96
|
|
|
|
|
|
|
| |
compiled for it -> pass
Scenario ffmpeg-0.10, where we are not prepared (compiled-in) for sw-resample support.
Don't use if compiled in version (CC) is < 0 (n/a), and allow to pass at load time.
|
| |
|
|
|
|
| |
next frame after play() will provide new frame. Added API doc.
|
| |
|
|
|
|
| |
-> GLES2 (in profile set)
|
|
|
|
| |
don't use tchar.h; Fix compiler warning: Add missing (intptr_t) cast.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
>= 4.8.1
- The previous MingW64 would still work, however we should align 32bit and 64bit versions
- Mingw 32bit does not support dshow/com properly!
- Use MingW64 32bit builds solve the problem
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
orientation change (flipped), API-doc,
- State
- Fix state transition (initGL() error)
- Camera options
- options uses ';' as query separator
- don't use 'default' options, driver should know
- Detect and act on orientation change (flipped)
- ffmpeg impl detects if flipped changes and triggers a SIZE update event.
This allows application to react, i.e. re-init GL and use new TextureCoord's.
Test: Works well on Windows w/ rawvideo dshow camera driver/codec.
- API-doc
- TexSeqEventListener/GLMediaEventListener usage / constraints (GL, ..)
- State transition fix
|
|
|
|
| |
characteristics.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
'camera ID'
If linesize is < 0, it is not invalid as assumed in commit eca6a5cb1e2beda84dfbafc31ed225e272f4f3fb,
but vertically flipped (bottom-up).
We have to adjust the data pointers, which are moved to the upper end of memory as well
and can proceed as usual.
TODO:
- Update texture 'mustFlipVertically' to 'false' in this case.
- Later:
- Allow updating texture size ..
- Whole pixel-fmt/texture-lookup-shader association must scale better,
i.e. extract the 'knowledge' into one class, use a static shader code
using uniforms instead of hard-coded values .. etc.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
windows, 2 more pixel formats, fail-safe data handling
- add support for ffmpeg 2 / libav 10 -> lavc55_lavf55_lavu52_lavr01
- add support for ffmpeg libswresample (similar to libavresample)
- handle BGRA (GL type) and BGR24 (texture shader)
- Change Camera URI semantics, drop 'host' and use 'path' for camera ID
and use 'query' for options.
- add support for Window's DShow camera selection
- our camera id -> index of list of video-input devices,
this gives us same behavior as w/ Linux
- requires windows libs: strmiids, uuid, ole32, oleaut32
- Compiles w/ MingW64, works w/ libav/ffmpeg
- TODO: test compilation w/ MingW 32bit !
- don't push data to texture if (linesize <= 0)
this may happen due to buggy decoder / setup ..
Tested manually on GNU/Linux x64 and Windows x64:
- GNU/Linux libav 0.8, libav 9, libav 10, ffmpeg 1.2, ffmpeg 2.0
- Windows libav 0.8, libav 9, ffmpeg 2.0
- videos and camera
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
def. high camera options, cleanup symbols)
- Fix libav/ffmpeg compilation
- Split native GLContext code from JoglCommon
- JoglCommon is required for ffmpeg_* c-compile/link
- Supported versions now:
- 0.8 53.53.51
- 9.0 54.54.52
- FFMPEGMediaPlayer
- Update API doc, add compatibility .. etc
- Pixel format conversions (via shader texture lookup func):
- YUV420P, YUVJ420P
- YUV422P, YUVJ422P
- YUYV422
- Properly handle aid/vid
- In camera mode: set high default values
- TODO: Make it configurable via camera URI:
- video_size
- framerate
- ?
- FFMPEGDynamicLibraryBundleInfo
- Cleanup symbols / remove unused (pre 53)
- Add av_dict_* methods
|
|
|
|
| |
for float values)
|
|
|
|
|
|
|
|
|
|
|
| |
version dependent c-files individually and inject object files. ; ffmpeg *register_all() at setStream0(..)
- Use 'dot less' dir/file names
- Compile ffmpeg version dependent c-files individually and inject object files.
- ffmpeg *register_all() at setStream0(..)
- Only register devices if available _and_ camera is requested.
|
| |
|
|
|
|
| |
'stub_includes'
|
|
|
|
| |
GL_UNSIGNED_SHORT_8_8_APPLE, GL_UNSIGNED_SHORT_8_8_REV_APPLE.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
missing symbol 'av_realloc'.
- Add camera input
- Use URI w/ scheme 'camera' to determine camera input is desired,
use URI host as camera id.
E.g. 'camera://0' for 1st camera.
- AndroidGLMediaPlayerAPI14: Via 'Camera'
- FFMPEG*: Via libavdevice, device name and input format
- TODO: Add controls to manipulate camera if available
- FFMPEG*
- Add symbols
- avcodec_register_all
- av_realloc (was missing)
- avdevice_register_all
- Load libavdevice (opt)
- Camera:
- Use <ID> (windows) and /dev/video<ID> other OS
- simply find the input format in native code
- Support YUYV422 (used in video4linux2, etc.)
- Stuff 2x 16bpp (YUYV) into one RGBA pixel!
- Add texture format for 16bpp
- Add texture lookup shader
- Fix av_packet leak in readNextImpl(..)
- Restore orig pointer and size values,
we may have moved along within packet.
Then call av_free_packet().
- Use null AudioSink if audio-id is NONE
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
FFMPEGNatives -> FFMPEGv08Natives + FFMPEGv09Natives
Enables FFMPEGMediaPlayer to work w/ either ffmpeg/libav version 8 or 9 w/ same JOGL binary
Same C source code is compiled against
1: version 0.8 FFMPEGv08Natives lavc53.lavf53.lavu51
2: version 0.9 FFMPEGv09Natives lavc54.lavf54.lavu52.lavr01
FFMPEGv08Natives and FFMPEGv09Natives implements FFMPEGNatives,
native C code uses CPP '##' macro concatenation to produce unique function names.
To enable 'cpp' to find the libav* header files matching the desired version,
we have placed them in the c-file's folder, issued '#include "path/file.h"
and added symbolic links to allow finding same module and 'sister modules':
ls -l libavformat/
..
lrwxrwxrwx 1 sven sven 13 Aug 26 12:56 libavcodec -> ../libavcodec
lrwxrwxrwx 1 sven sven 14 Aug 26 12:56 libavformat -> ../libavformat
lrwxrwxrwx 1 sven sven 12 Aug 26 12:57 libavutil -> ../libavutil
..
At static init FFMPEGDynamicLibraryBundleInfo, determines the runtime version
and instantiates the matching FFMPEGNatives, or null if non matches.
FFMPEGMediaPlayer still compares the compile-time and runtime versions.
FFMPEGMediaPlayer passes it's own instance to FFMPEGNatives for callbacks.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
negotiation w/ AudioSink; Misc
- Add libavresample support
- Resample if avail && (!AV_SAMPLE_FMT_S16 || !prefSampleRate || !sinkSupported)
- Resample to: prefSampleRate (if set), AV_SAMPLE_FMT_S16 and min(channelCount, maxChannelCount)
- Proper AudioFormat negotiation w/ AudioSink;
- Utilize AudioSink's 'isSupported(AudioFormat)'
- Misc
- use 'av_get_bytes_per_sample(fmt)' always, don't assume 2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
adds AL_SOFT_buffer_samples support w/ full AL caps
- Fixe type names:
- Remove AudioDataType, we only support PCM here anyways
- AudioDataFormat -> AudioFormat / Add 'planar' attribute to distingush packed/planar data type
- Validate float types
- Enhance AudioFormat negotiation
- Add 'isSupported(AudioFormat format)' which _shall_ be used before 'init(..)'
to test/negotiate format
- Add getMaxSupportedChannels(), which may be used w/ getPreferredFormat() if orig requested format fails
via 'isSupported(..)'
- 'init(..)' returns boolean only.
- ALAudioSink adds AL_SOFT_buffer_samples support w/ full AL caps
- Determine whether AL_SOFT_buffer_samples is supported
- Use new JOAL ALHelper to convert AudioFormat -> AL-types,
which also answers the 'isSupported(..)' query.
- Now allows multiple: channles, sample-types, etc.
|
| |
|
|
|
|
| |
libavresample of version 9
|
|
|
|
|
|
|
|
|
|
|
| |
- Add compile-time/runtime version check, fail if major versions do not match
assuming binary incompatibility
- Add: 'av_find_input_format' for future video input support
- Manually map '/dev/video<NUM>' to video input - not working yet.
- WINDOWS: Set file to '<NUM>'
- Set input format string depending on OS
|
| |
|
|
|
|
|
|
|
|
|
|
| |
their sub-folders; Add lavc54.lavf54.lavu52 (version 9) header files
lavc54.lavf54.lavu52:
- Begin testing w/ version 9
- Header files are binary incompatible (data types)
Default is still lavc53.lavf53.lavu51
|
|
|
|
| |
finding it and dumping a DEBUG message
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add EOS detection, setAudioVolume(..)
GLMediaPlayerImpl.initStreamGL(..):
Only require a minimum texture count of 2,
which is the bare minimum to allow our algorithm to work,
i.e. having a 'lastFrame' and avail/playing ringbuffer have each one frame.
Android's MediaPlayer API can only deal w/ one SurfaceTexture,
hence we have to fake a second SurfaceTextureFrame w/ same content
to allow our implementation to work w/ the threaded decoder (min 2 frames).
|
|
|
|
| |
allowing to change the audio volume.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
non-frame based AudioSink's to deal w/ desired queue sizes.
- Rename AudioSink.initSink(..) -> AudioSink.init(..)
- Move: "int initialFrameCount, int frameGrowAmount, int frameLimit" to
"int initialQueueSize, int queueGrowAmount, int queueLimit"
based on milliseconds instead of frame count.
- Passing hint 'float frameDuration' to calculate frame count for fame based audio sink, i.e. ALAudioSink.
- Adding sensible static final default values
- AudioDataFormat: Add convenient conversion routines (samples/bytes/frame-count)
- FFMPEGMediaPlayer: Retrieve audio frame size in samples per channel, pass it to AudioSink.init(..)
to properly calculate frame count/limits based on duration.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Multithreaded decoding and API should be considered stable by now,
minor changes may apply if Android/OMX impl. requires it.
We still need to solve TODO's as listed below, copied from 474ce65081ecd452215bc07ab866666cb11ca8b1.
+++
- *TextureFrame OO changes:
- TextureFrame extends TimeFrameI
- GLMediaPlayerImpl*
- Adapt to Ringbuffer changes of GlueGen commit f9f881e59c78e3036cb3f956bc97cfc3197f620d
- Fix impl. method's API doc
- getNextTextureImpl(..) returns video PTS
- Fix audio-only playback
- frame dropping shall only happen if:
- previous frame has not been dropped
- frame is too later
- one decoded frame is already available
- Don't block for decoder anymore:
- nextFrame = "videoFramesDecoded.getBlocking() -> videoFramesDecoded.get()";
No 'next decoded frame avail' only could mean:
- slow decoding/hardware
- slow transport
hence we shall not block rendering.
- Add DEBUG output if using last frame
- Add integer property 'jogl.debug.GLMediaPlayer.StreamWorker.delay' in milliseconds
to simulate slow decoding, i.e. delay is added in StreamWorker after decoding
before pushing new frame to Ringbuffer.
- FFMPEGMediaPlayer:
- audioFrameLimitWithVideo 128 -> 64
- audioFrameLimitAudioOnly 128 -> 32
- uses AudioSink's 'enqueueData(int pts, ByteBuffer bytes, int byteCount)'
- fixes for audio-only playback
+++
Working Tests: MovieSimple and MovieCube
TODO-1: Fix
- Android
- OMXGLMediaPlayer
TODO-2:
- Fix issue where async audio frames arrive much later than 1st video frame, i.e. around 300ms.
- Default TextureCount .. maybe 3 ?
- Adding Audio synchronization ?
- Find 'truth' about correlation of audio and video PTS values,
currently, we assume both to be unrelated ?
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Reuses ALAudioFrames to ease GC, Ringbuffer changes
- Adapt to Ringbuffer changes of GlueGen commit f9f881e59c78e3036cb3f956bc97cfc3197f620d
- Favor AudioSink 'AudioFrame enqueueData(int pts, ByteBuffer bytes, int byteCount)',
- Impl. shall reuse AudioFrame's instead of creating them on the fly
- User shall simply pass the net data required, while receiving an internal AudioFrame
- Add byte/time calc to AudioDataFormat:
- Add getDuration(byteCount) and getByteCount(ms).
- *AudioFrame OO changes:
- abstract AudioFrame extends TimeFrameI
- allow setting of all components to reuse instanced (GC clean)
- ALAudioSink reuses ALAudioFrames to ease GC:
- Remove creating temporary objects to ease GC
- ALAudioFrame holds ALBuffer name, remove ActiveBuffer type.
- Use ALAudioFrame similar to TextureFrame in GLMediaPlayerImpl,
i.e. fill them in 'full' Ringbuffer and move them in-between 'full'/'playing' Ringbuffer.
-
|
|
|
|
| |
basen on integer milliseconds.
|
|
|
|
| |
work well w/ URI encoded names.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Update/fix GLMediaPlayer API doc
- GLMediaEventListener: Add event bits for all state changes to be delivered via attributesChanged(..)
- StreamWorker / Decoder Thread:
- Use StreamWorker only !
- Handle exceptions on StreamWorker via StreamException
- Handles stream initialization and decoding (-> initStream(..))
- Split initGLStream(..) -> initStream(..) + initGL(GL)
- allow initStream(..)'s implementation being executed on StreamWorker
- allow GL initialization to be 'postponed' when stream is read,
i.e. non blocking stream initialization (UI .. etc)
- Handle EOS via END_OF_STREAM_PTS -> pause/event
- Video: Use lock-free LFRingbuffer, similar to
ALAudioSink (commit f18a94b3defef16e98badd6d99f2422609aa56c5)
+++
- FFMPEGDynamicLibraryBundleInfo
- Add avcodec's:
- avcodec_get_frame_defaults, avcodec_free_frame (54.28.0), avcodec_flush_buffers,
- Add avutil's:
- av_frame_unref (55.0.0)
- Add avformat's:
- avformat_seek_file (??)
+++
- FFMPEGMediaPlayer Native:
- add 'snoop' video frames for a/v frame count relation.
disabled per default, since no more needed due to ALAudioSink's
grow-buffer usage of LFRingbuffer.
- use sp_avcodec_free_frame if available
- 'useRefCountedFrames=1' for libav 55.0 to cache more than one audio frame,
not used since ALAudioSink's OpenAL usage does not require it (copies data once).
Note: the above snooped-video frame count is used here.
- use only one cached audio-frame (-> see above, OpenAL copies data once),
while reusing the NIO buffer!
- Perform OpenGL sync (glFinish) in native code!
- find proper PTS value, i.e. either frame's PTS or DTS,
see 'PTSStats'.
- FFMPEGMediaPlayer Java:
- use private fields
- simplified code due to above changes.
+++
Working Tests: MovieSimple and MovieCube
TODO-1: Fix
- Android
- OMXGLMediaPlayer
TODO-2:
- Fix issue where async audio frames arrive much later than 1st video frame, i.e. around 300ms.
- Default TextureCount .. maybe 3 ?
- Adding Audio synchronization ?
- Find 'truth' about correlation of audio and video PTS values,
currently, we assume both to be unrelated ?
|
|
|
|
| |
getNextTexture(..), may blocks .. or not, depending on implementation and state.
|
|
|
|
| |
existing object name.
|
|
|
|
| |
frameLimit allowing an optional used Ringbuffer to grow in implementation.
|
|
|
|
| |
30475c6bbeb9a5d48899b281ead8bb305679028d
|