summaryrefslogtreecommitdiffstats
path: root/libs
Commit message (Collapse)AuthorAgeFilesLines
* Added a fix from ↵PaulK2011-05-171-1/+1
| | | | | | http://groups.google.com/group/android-building/browse_thread/thread/ed22cf8f6dc8143c to fix the build. Modified: libs/utils/Android.mk
* surfaceflinger and libagl: fix bug #11829Denis 'GNUtoo' Carikli2011-01-098-137/+35
| | | | | | | | | | | | | | | The bug #11829 that can be found here: http://code.google.com/p/android/issues/detail?id=11829 causes various graphic artefacts. According to naseer.ahmed it needed modifications in SurfaceFlinger and libagl The modifications comes from: https://www.codeaurora.org/gitweb/quic/la/?p=platform/frameworks/base.git;a=shortlog;h=refs/heads/froyo_almond at commit 33e6506d96d7602284e18ea8c66dbfd04dc72aa3 Change-Id: I703bace69cb7a41ac8ed69ca1e28b98aa618e0f3 Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
* Merge "audioflinger: Nasty hack for adjusting BCM FM volume" into froyoSteve Kondik2010-11-302-3/+48
|\
| * audioflinger: Nasty hack for adjusting BCM FM volumeSteve Kondik2010-11-292-3/+48
| | | | | | | | | | | | | | | | | | Normally this kind of nasty hack could go into libaudio, but on this particular device, we have no access to the source code of libaudio. Put the nasty hack into audioflinger and wrap it in ifdefs. Change-Id: I5e3e495e3bd6b671823967b61ba5ceb49e59a401
* | CAF Continuous AutofocusGiulio Cervera2010-11-281-0/+6
|/ | | | Change-Id: Idfdcf8ac657669b75c5a6219c2f91d20963bbe5b
* frameworks/base: Destroy the overlay even if there is an errorNaomi Luis2010-11-261-2/+7
| | | | | | | | Allow the overlay object to be destroyed whenever the destroy is invoked. Currently the destroy call returns if there is an error encountered, which results in open data and control channels. Change-Id: I1f2ef4ebb5fb1dcabf05ab50b5bbf6e5e240a63a
* input: Handle extra keys on Vision and Espresso separatelySteve Kondik2010-11-171-0/+6
| | | | | | | | These devices have physical keyboards, but certain keys (? and SYM) are mapped differently. Check the target's name and use the appropriate set of keycodes. Change-Id: Ib91a47beb3681b411bbbd1946d5478334c5f3c33
* camera: Wrap getBufferInfo code in ifdefs.Steve Kondik2010-11-123-10/+20
| | | | | | Define BOARD_CAMERA_USE_GETBUFFERINFO if required. Change-Id: Ieac0bed92c8a1517a6a7730b9636f3591c55bbcb
* Camera: Add support for getBufferInfo APISteve Kondik2010-11-112-1/+37
| | | | | | | | Reworked to avoid ABI breakage. This reverts commit b5d9f64973f06b45423b2216050aa0d2d2513f96. Change-Id: I602faec8e85d80a30dadceed8c7942dc133989b8
* Revert "Revert "audio: Omit FM code when HAVE_FM_RADIO is not set""Steve Kondik2010-11-0713-13/+83
| | | | | | | | | Cleaned it up, and fixed a misplaced ifdef that broke the whole damn thing. This reverts commit 568f674b8ef1a83b7ba948b0e53e0dd86207d115. Change-Id: I99fde73b2cf283dde7e9bcb1e73cf86e553324d3
* Revert "audio: Omit FM code when HAVE_FM_RADIO is not set"Steve Kondik2010-11-0713-96/+13
| | | | This reverts commit c5358c4b94c2953290988be654b284840551b5bd.
* audio: Omit FM code when HAVE_FM_RADIO is not setMichael Webster2010-11-0313-13/+96
| | | | Change-Id: If140f0cd1232f927f4870c5de3f25deac73c3778
* Inotify support for TSLibrmcc2010-10-291-7/+40
| | | | Change-Id: I26a289f00d98b5f010e1a6a9db2c111b349a29b6
* TSLib: Changes made for using TSLibManas Abichandani2010-10-292-48/+142
| | | | | | | | | | | | | These changes are made to make use of TSLib to calibrate events generated from touch screen. ts_read in tslib reads raw events written by touchscreen driver in /dev/input/event0. ts_read then packages all the events between each EV_SYN and sends the final calibrated values to EventHub.cpp along with the list of events raw read. We then send each event seperately in this list to the upper layer. Change-Id: If529e5687620df50cf725edbaa321953df50c0a5
* Disable surface dithering by defaultEran2010-10-171-1/+1
| | | | Change-Id: Ic6daad4beb95941a4323384adea85b6312682927
* frameworks/base: Use textures in LayerDim for smooth menu scrolling.Saurabh Shah2010-10-162-14/+5
| | | | | | | | | Disable the use of DIM_WITH_TEXTURE flag which uses OpenGL textures for a specific platform. If Display Hardware uses SLOW_CONFIG, use OpenGL textures for speedup. This improves scrolling response in LayerDim menus. Change-Id: Icda97eaab3e0745e5b0841a7f5c866283999356d CRs-fixed: 253607
* support FM Radio audio routing in audio flingerxinyu2010-10-1212-7/+155
|
* Initialize dither state during object constructionAntti S. Lankila2010-09-281-0/+7
|
* Revert "Camera: Add support for getBufferInfo API"cyanogen2010-09-232-38/+0
| | | | This reverts commit 57495aafa6ab172c2d2cfe32ca8a4fa57f98cb9b.
* Camera: Add support for getBufferInfo APIKiran Kumar H N2010-09-222-0/+38
| | | | | | | | | Add getBufferInfo() API which gets the recording buffer information from the HAL layer. The opencore uses this API to query the details of the recording buffers allocated by the HAL layer. Add getBufferInfo() to Stub Camera to avoid compilation error for generic/emulator builds.
* Use floating point DSP on armeabi-v7aAntti S. Lankila2010-09-093-108/+264
|
* Add noise-shaped dithering supportAntti S. Lankila2010-09-064-23/+63
| | | | | | | Used rectangular probability distribution function before by accident, depite calculating the high-passed triangular probability density function I intended to use. This commit rectifies the mistake and also adds noise-shaped error feedback similar to JACK.
* Estimate center channel and avoid DSP on itAntti S. Lankila2010-09-022-7/+25
| | | | | Also fixes reverb like I originally intended it and tweaks parameters a little.
* Make Headphone DSP with just 1 biquadAntti S. Lankila2010-09-022-20/+7
| | | | | | | | | | | Subtracting the left and right channels could be argued to separate the center channel from the left/right channel and the DSP therefore gets applied to panned sounds only. This halves the cost of the filtering and eliminates the bass boost issue that bs2b compensates for with a highboost filter. I also adjusted the delay-to-level spatial filter to try to get slightly more consistent perception of panning across the frequency spectrum.
* Initialize track volume more correctlyAntti S. Lankila2010-08-221-2/+2
| | | | prevVolume is scaled by 16 because of the ramping performed with it.
* Various DSP enhancementsAntti S. Lankila2010-08-183-39/+29
| | | | | | | | | | | | | | | | | | | | - Update headset to a 2nd order shelf filter, drop allpass delays A single shelf filter can give a time delay for low frequencies and attenuation for high frequencies yielding a headphone virtualization directly. - Make equalizer out of high shelves. Cascading 4 high shelves and one global gain gives me 5 controllable bands with passthrough characteristics when not used. - Set loudness target to 93 % Also correcting a volume-ramping bug which sometimes caused a slow adjustment to desired volume because of integer truncation and invalid round down/up algorithm. This should eliminate cases where song starts super loud.
* Only monitor recognized uevent devices.Sean McNeil2010-08-181-1/+8
| | | | | | | | | system_server can potentially monitor uevent devices that are of no use to it. For instance, an accelerometer implementation as uevents. This would cause the process to be busy when unnecessary. If a device cannot be classified, don't monitor it. Change-Id: I489cc453bdce7cb376102cba67a9ea470e13292d
* Sound tweaks: Always on speaker 2.0, headphone attenuation, headphone music ↵Jonas Larsson2010-08-161-13/+93
| | | | vol limit, audio focus requests.
* frameworks/base: Swap width and height of temporary buffer only withOmprakash Dhyade2010-08-161-1/+1
| | | | | | | | | | | orientation change Current code swaps the width and height by assuming that aspect ratio of the buffer width and height will be same as that of the layout clip width and height. That is not always true. Change the check to orientation change. Change-Id: Ie387f3a7369025427484e4173cbde7a08df2b9d7
* Use 1st order high shelf instead of RC filter.Antti S. Lankila2010-08-112-2/+20
| | | | | | This admittedly sounds almost the same, but the advantage of shelving is with more high-frequency content being mixed. That being said, the used shelving is reasonably steep (16 dB).
* Spread the equal loudness contour upwards.Antti S. Lankila2010-08-111-1/+1
|
* Take some ideas from bs2b.Antti S. Lankila2010-08-112-33/+4
| | | | | | | | | | | Bs2b uses the low-pass filter to generate the delay as a byproduct of its operation. The filter is generally set around 700 Hz where it produces about 250 microseconds of delay. In addition to that, the crossfeed is now mixed at -6 dB level. The reverb is unchanged, although the new configuration gives it larger role in forming an impression of a soundstage. The design is still not complete, but good enough for a commit.
* HACK: Allow use of Eclair MemoryDealer for compatibility.Steve Kondik2010-08-085-1/+441
| | | | Take #2.
* Revert "HACK: Allow use of Eclair MemoryDealer for compatibility."Steve Kondik2010-08-086-444/+1
| | | | Patchset is broken. Will recommit a single working patch.
* HACK: Allow use of Eclair MemoryDealer for compatibility.Steve Kondik2010-08-086-1/+444
|
* Spacing cleanupkmobs2010-08-072-21/+20
|
* Added support for user selectable calibration.niloc2010-08-073-6/+49
| | | | | | Added support for user-defined calibration. Added support for user-defined calibration.
* Revert "Camera: Adding framework support for Geotagging."Steve Kondik2010-08-071-1/+1
| | | | This reverts commit b769bca07fb9ec14d8a6b09905a38bdc4ecc8864.
* Increase virtual head size slightly.Antti S. Lankila2010-08-061-2/+2
|
* Integrate reverb and headphone effects.Antti S. Lankila2010-08-062-133/+100
|
* Make reverb buffers ping-pong between L&R channelsAntti S. Lankila2010-08-061-5/+7
| | | | | | | Additionally move the estimation boost clamper into a more sensible base in the codebase. Signed-off-by: Antti S. Lankila <alankila@gmail.com>
* Rebalance dithering, volume ramping and DRCAntti S. Lankila2010-08-062-29/+29
| | | | | | | | | | | | | | | | - Move dithering to actually happen in the ditherAndClamp() procedure to reduce memory pressure. - Make volume ramping occur according to opposite logic: increases of volume are slow, decreases rapid. That is a balance that makes far more sense and spares user's ears far better. - Avoid extreme boosts by mapping the dynamic range boost curve on a nonlinear function x - x^3/3, for -1 <= x <= 0. This cuts off 1/3rd of the entire boosting region and avoids the artifacts arising from hardclipping this function. Signed-off-by: Antti S. Lankila <alankila@gmail.com>
* Add reverb modeAntti S. Lankila2010-08-062-24/+199
| | | | | | Add a reverb module based on the Haas stereo widening effect. Signed-off-by: Antti S. Lankila <alankila@gmail.com>
* Remove boost limit from bottom.Antti S. Lankila2010-08-062-15/+15
| | | | | | | | I noticed there was some distortion arising from the fixed bottom level set in the level estimator, so I removed such limits. I also enhanced the estimation quality for the nearly inaudible sound levels. Signed-off-by: Antti S. Lankila <alankila@gmail.com>
* Added N1 Calibration profile to rendering options.niloc2010-08-041-0/+1
|
* Added N1 Calibration profile to rendering options.niloc2010-08-041-0/+3
|
* Fix rattling of very quiet soundsAntti S. Lankila2010-08-031-1/+4
| | | | | An integer truncation issue resulted in discontinuous estimates of sound level when the audio stream sound level was very quiet.
* Do not use track's frameCount but the buffer's frameCount.Antti S. Lankila2010-08-032-7/+7
| | | | This change makes compression effect about 10 times faster.
* Fix some shortcomings of the audio level estimationAntti S. Lankila2010-08-033-32/+71
| | | | | | | | | | | | Use crude approximation of the equal loudness curves (a single 2nd order bandpass at 1 kHz). Because we must be stateless, the filter is always reset first, which somewhat damages the behavior. Limit adjustment range to 30 dB to avoid problems when sound goes very quiet. Make ramp up fast and ramp down slow. This gives reasonable response to transients between periods of silence.
* Optimizations and cleanups.Antti S. Lankila2010-08-012-54/+108
| | | | | | | | | | | | | | | | | | Optimization: - write ARM assembly routine for biquads, using one load per pair of coefficients. Correctness fix: - clamp sound level estimate below at -50 dB because of integer truncation at estimateLevel() method. Cosmetic changes: - add headings in .cpp to make it easier to keep track of file structure. - remove trailing whitespace - rearrange some functions in logical order.
generated by cgit v1.1 at 2025-08-15 21:53:40 +0000