| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
| |
|
|
|
|
| |
library path, support throughout DynamicLibraryBundle[Info]
|
|
|
|
| |
misleading and wrong info (e.g. on mute)
|
|
|
|
| |
disruption after pause(flush:=false)
|
|
|
|
| |
i.e. ignore audio_queued_ms in such case (Fixes video sync on SCR w/o audio)
|
|
|
|
| |
direct user-exposure and potential data-race
|
|
|
|
| |
w/o video-frame (fixes 'massive' seek'ing)
|
|
|
|
| |
Left behind hacking for AV sync ..
|
|
|
|
|
|
| |
!hasVideoFrame upfront, simplifying sync-block
Fix playSpeed > 1: dt_v (keep playSpeed factor), allow drop frame on no-frames w/ playSpeed > 2
|
|
|
|
| |
stream is not yet initialized or playing, cached clipped value and set at initGL(..)
|
|
|
|
| |
user from selecting video or audio PTS.
|
|
|
|
| |
USE_GLSL_VERTICAL_FLIP, changed names in commit d6cf89f22a5926b437c4430eb166972d90fd92be
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
setUseGLSLVerticalFlip(false) in addition to property 'jogl.gljpanel.noglsl'
Suche setting would enable the slow path of flipping a potentially 'big' framebuffer via the CPU.
The best performance path is using the GLSL shader to vertically flip the FBO,
or - w/o using the shader - to render the GL scene vertically-flipped and set the flag 'flipVertical' to false.
- property "jogl.gljpanel.noverticalflip"
- setter setSkipGLOrientationVerticalFlip(boolean)
If set to <code>true</code>, user needs to flip the OpenGL rendered scene
<i>if {@link #isGLOriented()} == true</i>, e.g. via the projection matrix.<br/>
See constraints of {@link #isGLOriented()}.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(multiple of) audio_dequeued_ms
The case of lagging audio to the audio master-clock (by experience) is a rare and probably synthethic case
of the AV async videos, was
dt_a > MAX_VIDEO_ASYNC && d_apts > 0
now with increased threshold max_adelay = Math.max( 4*audio_dequeued_ms, 4*MAX_VIDEO_ASYNC )
dt_a > max_adelay && d_apts > 0
In conjunction the video-repeat case (video pts > SCR) shall use a higher threshold _when_
detecting, i.e. min1_audio_queued_ms = Math.max( 2*audio_dequeued_ms, 2*MAX_VIDEO_ASYNC )
to ensure enough buffered audio exists (2 audio-frames) before the next getNextTexture()
hits within vsync.
This early detection and min1_audio_queued_ms threshold
is double of the late threshold for video-repeat min0_audio_queued_ms = Math.max( audio_dequeued_ms, MAX_VIDEO_ASYNC ),
when the to-be repeated frame shall be displayed within getNextThreshold().
Failing this requirement (1 audio-frame) will discard it and gather the next video-frame,
allowing to fill the audio-buffer. A subsequent AV sync shall correct the difference.
Strategy is less intervention on less buffered-audio.
This shorter tolerance also reduces some video lag or stuttering on 24fps -> 60fps films.
|
|
|
|
|
|
|
|
|
| |
resample: Use swr_get_out_samples(..) to calculate the required output sample count
Notable when playing audio with e.g. 24k sample rate on a 48k OpenAL ALAudioSink,
the sample duration was cut in half due to erroneous resampling missing half the required samples.
Using swr_get_out_samples(..) resolves this issue pre swr_convert(..), the resampling.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
master-clock, enabling proper AV sync w/ untouched audio
We can finally utilize the added pass through audio PTS, see commits
- GlueGen 52725b4c6525487f93407f529dc0a758b387a4fc
- JOAL 12029f1ec1d8afa576e1ac61655f318cc37c1d16
This enables us to use the audio PTS as the master-clock and adjust video to the untouched audio.
In case no audio is selected/playing or audio is muted,
we sync merely on the system-clock (SCR) w/o audio.
AV granularity is 22ms, however, since the ALAudioSink PTS may be a little late,
it renders even a slightly better sync in case of too early audio (d_apts < 0).
Since video frames are sync'ed to audio, the resync procedure may result
in a hysteresis swinging into sync. This might be notable at start
and when resumed audio or after seek.
We leave the audio frames untouched to reduce processing burden
and allow non-disrupted listening.
Passed AV sync tests
- Five-minute-sync-test.mp4
- Audio-Video-Sync-Test-Calibration-23.98fps-24fps.mp4
- Audio-Video-Sync-Test-2.mkv
|
| |
|
|
|
|
| |
Using `-Djogl.debug.GLMediaPlayer.AVSync -Djogamp.debug.AudioSink` to trace AVSync issues for now.
|
|
|
|
|
| |
FFMPEGMediaPlayer: This also effectively reduces the audio buffer size from 3000ms -> 768ms,
the new default for audio streams with video.
|
|
|
|
|
|
|
|
|
|
|
| |
custom name against GLSL internal 'texture2D', add missing VideoPixelFormat conversions
Always use own textureLookupShader, avoiding confusion or even race conditions.
Validate textureLookupShader custom name against GLSL internal 'texture2D' (illegal override)
Add missing VideoPixelFormat conversions:
- RGB24, ARGB, RGBA, ABGR, BGRA
|
|
|
|
| |
instead hard coded branch
|
| |
|
|
|
|
| |
methods accordingly
|
|
|
|
|
|
|
|
| |
allowing to properly take-down user resources at dispose(GLAutoDrawable)
Used for UISceneDemo20 to stop and release SimpleSineSynth and its ALAudioSink.
The latter causes a bad exit (crash at OpenAL32.dll) on OpenJDK's Window Binary if not stopped!
|
|
|
|
| |
Callbacks as recommended
|
|
|
|
| |
Adapter patterns
|
|
|
|
| |
Callbacks as recommended
|
|
|
|
| |
ReleaseStringUTFChars for GetStringUTFChars() acquired UTF
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
a native dispatch'ed event like key/mouse/touch input
SIGSEGV on use after free of native X11 Display* at XEventsQueued in DisplayDriver.DispatchMessages0.
This potentially happens when an application destroys
the NEWT Window/Display from an action being called directly
from DisplayDriver.DispatchMessages0 (itself), i.e. keyboard or mouse input.
DisplayDriver.DispatchMessages0 stays in the event loop and the next
XEventsQueued call causes a SIGSEGV due to already deleted
display driver connection and hence invalid native X11 Display*.
This issue also exist for other Windowing System drivers,
where the native (dispatch) method sticks to a loop
and still (re)uses the window or display handle.
One is WindowsWindow, where touch events are looped,
but such handler could have closed the window.
Querying the status of a window / display instance before dispatching
is not be good enough
- resource could already be GC'ed, so we also would need to query jobject status
- would imply an addition Java callback
+++
This fix: Having the Java callbacks return
a boolean with the value Window.isNativeValid().
This way the dispatch logic
- can bail out right away w/o using the resource anymore
- must be reviewed by myself due to changed Call{Void->Boolean}*(..)
invocation change.
This review shall resolve potential similar issues.
+++
Tested on X11/Linux/GNU, Windows and MacOS
with new TestDestroyGLAutoDrawableNewtAWT,
which tests all destruction invocation variants.
|
|
|
|
| |
directly w/o mitigation attempts (off-thread)
|
|
|
|
|
|
|
|
| |
'mouseOver' (mouseDragged)
When using a mouse, Scene activates a Shape if mouse is over it (mouseOver).
Hence don't de-activate such Shape via mouse-button clicked or released.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
for sorting w/o actually modifying the position (enogh to be painted on top and for selection)
Also use a simplified comparison using only float relational operations w/o NaN/Inf bit-stuff or epsilon,
as it should be accurate enough for this cause.
This shall also resolve Bug 1454, as we no more modify the position directly
but the local zOffset field .. but this has to be seen (data race).
|
|
|
|
| |
dirty. Only issue validate on childs if theyre shape is dirty
|
|
|
|
| |
VideoPixelFormat conversion for easier debugging
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
aspect-ratio, letter-boxed or zoomed (config) + Bug 1466 Fix color mixing
Bug 1465: Region currently simply bloats a given texture to its region AABBox,
which renders textures with the wrong aspect ratio.
Add facility to program the texture-coordinates to either letter-box
or scaled-up (and cut) true aspect-ratio.
Default shall be zoom (scale-up and cut),
but user shall be able to set a flag in the Region for letter-box.
Have the shader clip texture coordinates properly,
best w/o branching to soothe performance.
See functions.glsl
+++
Bug 1466: Current color mix: texture * color_channel * color_static
is useless in GraphUI.
color_static shall modulate the texture, which works.
But in case of color_channel (attribute/varying)
we want it to be mixed so it can become the more dominant color
for e.g. a border.
Desired is:
color = vec4( mix( tex.rgb * gcu_ColorStatic.rgb, gcv_Color.rgb, gcv_Color.a ),
mix( tex.a * gcu_ColorStatic.a, 1, gcv_Color.a) );
|
|
|
|
|
|
|
|
|
|
| |
Commit c5431f46b7bf64f109315ec78461859dd88f202a
reduced the disableBackgroundErase(..) to SunToolkit's variation which doesn't work on Windows
as it does not act upon the java.awt.Canvas peer post addNotify().
This re-introduces the java.awt.Canvas method via class JAWTUtil.BackgroundEraseControl
and its called only after addNotify() on Windows and ASAP for everyone else.
Method also calles the SunTookit variation just to be sure.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
ad38d1559854985b1131e5b6c7274a392b5bc265
Commit ad38d1559854985b1131e5b6c7274a392b5bc265 introduced XTranslateCoordinates(..) to savely validate
the client-space window position against the parent (root).
Totally missing in this change was the NEWT child window case
since it always used the root-window as the destination.
This change tracks the parent-window (valid parent Window or NULL)
within the JavaWindow struct and either uses the parent-window
if available or the root-window for XTranslateCoordinates(..).
This results in the proper client-space position.
Validated against
- TestGearsES2NewtCanvasAWT
- TestBug1431NewtCanvasAWT
on Debian 12 w/ Java17.
|
| |
|
|
|
|
| |
have the text resize being of fixed aspect-ratio (like the movie)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
symbols (see setFixedLabelSize()),
This allows convenient instantiation of a Button changing its text (or symbol) when toggled, e.g.:
UISceneDemo03:
+ final Button button = new Button(options.renderModes, fontSymbols,
+ fontSymbols.getUTF16String("play_arrow"), fontSymbols.getUTF16String("pause"),
+ buttonWidth, buttonHeight, buttonZOffset);
+ button.setSpacing(symSpacing, fixedSymSize);
+++
setFixedLabelSize(..):
Sets fixed label font size clipped to range [0 .. 1], defaults to {@code 0, 0}.
Use {@code w=0, h=1} when using single symbols from fixed sized symbol fonts!
Use {@link #setSpacing(Vec2f, Vec2f)} to also set spacing.
The fixed label font size is used as the denominator when scaling.{@code max(fixedLabelSize, fontLabelSize)},
hence reasonable values are either {@code 1} to enable using the given font-size
for the axis or {@code 0} to scale up/down the font to match the button box less spacing for the axis.
|
| |
|
|
|
|
| |
experience, add overridable toggleNotify(..) allowing implementations to react before listener.
|
|
|
|
| |
See commit e5eadcdaa615dbeb762885b50435a1f79d6ca895
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
height exceeding overall CharSequence
TypecastFont using `metrics.getAscentFU() - metrics.getDescentFU()` for ascent used for all undefined and whitespace Glyphs w/o a spatial outline,
leads to a potential exceeding height compared to the actual used
bounding box of the rendered text.
This in turn leads to layout issues, e.g. button labels are placed too low.
Solution:
- Whitespace/Undefined: Drop full height 'metrics.getAscentFU() - metrics.getDescentFU()', b/c of non-existing shape height.
- Otherwise, layout on AABBox or created empty shape would pick up such default hhea-table ascent which might exceed actual string height.
Sideeffect would be if user relies on height of a whitespace.
However, knowing this fact - a user may always pick the hhea-table's ascent metric values as exposed in Font and Font.Metric.
This resolves remaining layout issues, including button labels.
|
| |
|
|
|
|
| |
to preserver footprint ..
|
|
|
|
| |
debugging purposes
|
| |
|