diff options
Diffstat (limited to 'include')
-rw-r--r-- | include/binder/Parcel.h | 18 | ||||
-rw-r--r-- | include/media/AudioRecord.h | 19 | ||||
-rw-r--r-- | include/media/IAudioFlinger.h | 2 | ||||
-rw-r--r-- | include/media/mediarecorder.h | 4 | ||||
-rw-r--r-- | include/tts/TtsEngine.h | 167 |
5 files changed, 195 insertions, 15 deletions
diff --git a/include/binder/Parcel.h b/include/binder/Parcel.h index 9087c44..af1490a 100644 --- a/include/binder/Parcel.h +++ b/include/binder/Parcel.h @@ -80,8 +80,11 @@ public: status_t writeStrongBinder(const sp<IBinder>& val); status_t writeWeakBinder(const wp<IBinder>& val); - // doesn't take ownership of the native_handle - status_t writeNativeHandle(const native_handle& handle); + // Place a native_handle into the parcel (the native_handle's file- + // descriptors are dup'ed, so it is safe to delete the native_handle + // when this function returns). + // Doesn't take ownership of the native_handle. + status_t writeNativeHandle(const native_handle* handle); // Place a file descriptor into the parcel. The given fd must remain // valid for the lifetime of the parcel. @@ -114,12 +117,11 @@ public: wp<IBinder> readWeakBinder() const; - // if alloc is NULL, native_handle is allocated with malloc(), otherwise - // alloc is used. If the function fails, the effects of alloc() must be - // reverted by the caller. - native_handle* readNativeHandle( - native_handle* (*alloc)(void* cookie, int numFds, int ints), - void* cookie) const; + // Retrieve native_handle from the parcel. This returns a copy of the + // parcel's native_handle (the caller takes ownership). The caller + // must free the native_handle with native_handle_close() and + // native_handle_delete(). + native_handle* readNativeHandle() const; // Retrieve a file descriptor from the parcel. This returns the raw fd diff --git a/include/media/AudioRecord.h b/include/media/AudioRecord.h index e962db6..106807e 100644 --- a/include/media/AudioRecord.h +++ b/include/media/AudioRecord.h @@ -39,10 +39,15 @@ class AudioRecord { public: - enum stream_type { + // input sources values must always be defined in the range + // [AudioRecord::DEFAULT_INPUT, AudioRecord::NUM_INPUT_SOURCES[ + enum input_source { DEFAULT_INPUT =-1, MIC_INPUT = 0, - NUM_STREAM_TYPES + VOICE_UPLINK_INPUT = 1, + VOICE_DOWNLINK_INPUT = 2, + VOICE_CALL_INPUT = 3, + NUM_INPUT_SOURCES }; static const int DEFAULT_SAMPLE_RATE = 8000; @@ -118,7 +123,7 @@ public: * * Parameters: * - * streamType: Select the audio input to record to (e.g. AudioRecord::MIC_INPUT). + * inputSource: Select the audio input to record to (e.g. AudioRecord::MIC_INPUT). * sampleRate: Track sampling rate in Hz. * format: PCM sample format (e.g AudioSystem::PCM_16_BIT for signed * 16 bits per sample). @@ -140,7 +145,7 @@ public: RECORD_IIR_ENABLE = AudioSystem::TX_IIR_ENABLE }; - AudioRecord(int streamType, + AudioRecord(int inputSource, uint32_t sampleRate = 0, int format = 0, int channelCount = 0, @@ -165,7 +170,7 @@ public: * - NO_INIT: audio server or audio hardware not initialized * - PERMISSION_DENIED: recording is not allowed for the requesting process * */ - status_t set(int streamType = 0, + status_t set(int inputSource = 0, uint32_t sampleRate = 0, int format = 0, int channelCount = 0, @@ -197,6 +202,7 @@ public: int channelCount() const; uint32_t frameCount() const; int frameSize() const; + int inputSource() const; /* After it's created the track is not active. Call start() to @@ -323,7 +329,8 @@ private: audio_track_cblk_t* mCblk; uint8_t mFormat; uint8_t mChannelCount; - uint8_t mReserved[2]; + uint8_t mInputSource; + uint8_t mReserved; status_t mStatus; uint32_t mLatency; diff --git a/include/media/IAudioFlinger.h b/include/media/IAudioFlinger.h index 031335e..bac3d29 100644 --- a/include/media/IAudioFlinger.h +++ b/include/media/IAudioFlinger.h @@ -54,7 +54,7 @@ public: virtual sp<IAudioRecord> openRecord( pid_t pid, - int streamType, + int inputSource, uint32_t sampleRate, int format, int channelCount, diff --git a/include/media/mediarecorder.h b/include/media/mediarecorder.h index b9ea0c6..aebe191 100644 --- a/include/media/mediarecorder.h +++ b/include/media/mediarecorder.h @@ -35,6 +35,10 @@ typedef void (*media_completion_f)(status_t status, void *cookie); enum audio_source { AUDIO_SOURCE_DEFAULT = 0, AUDIO_SOURCE_MIC = 1, + AUDIO_SOURCE_VOICE_UPLINK = 2, + AUDIO_SOURCE_VOICE_DOWNLINK = 3, + AUDIO_SOURCE_VOICE_CALL = 4, + AUDIO_SOURCE_MAX = AUDIO_SOURCE_VOICE_CALL }; enum video_source { diff --git a/include/tts/TtsEngine.h b/include/tts/TtsEngine.h new file mode 100644 index 0000000..06f3820 --- /dev/null +++ b/include/tts/TtsEngine.h @@ -0,0 +1,167 @@ +/* + * Copyright (C) 2009 Google Inc. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include <media/AudioSystem.h> + +// This header defines the interface used by the Android platform +// to access Text-To-Speech functionality in shared libraries that implement speech +// synthesis and the management of resources associated with the synthesis. +// An example of the implementation of this interface can be found in +// FIXME: add path+name to implementation of default TTS engine +// Libraries implementing this interface are used in: +// frameworks/base/tts/jni/android_tts_SpeechSynthesis.cpp + +namespace android { + +// The callback is used by the implementation of this interface to notify its +// client, the Android TTS service, that the last requested synthesis has been +// completed. +// The callback for synthesis completed takes: +// void * - The userdata pointer set in the original synth call +// uint32_t - Track sampling rate in Hz +// audio_format - The AudioSystem::audio_format enum +// int - The number of channels +// int8_t * - A buffer of audio data only valid during the execution of the callback +// size_t - The size of the buffer +// Note about memory management: +// The implementation of TtsEngine is responsible for the management of the memory +// it allocates to store the synthesized speech. After the execution of the callback +// to hand the synthesized data to the client of TtsEngine, the TTS engine is +// free to reuse or free the previously allocated memory. +// This implies that the implementation of the "synthDoneCB" callback cannot use +// the pointer to the buffer of audio samples outside of the callback itself. +typedef void (synthDoneCB_t)(void *, uint32_t, AudioSystem::audio_format, int, int8_t *, size_t); + +class TtsEngine; +extern "C" TtsEngine* getTtsEngine(); + +enum tts_result { + TTS_SUCCESS = 0, + TTS_FAILURE = -1, + TTS_FEATURE_UNSUPPORTED = -2, + TTS_VALUE_INVALID = -3, + TTS_PROPERTY_UNSUPPORTED = -4, + TTS_PROPERTY_SIZE_TOO_SMALL = -5 +}; + +class TtsEngine +{ +public: + // Initialize the TTS engine and returns whether initialization succeeded. + // @param synthDoneCBPtr synthesis callback function pointer + // @return TTS_SUCCESS, or TTS_FAILURE + virtual tts_result init(synthDoneCB_t synthDoneCBPtr); + + // Shut down the TTS engine and releases all associated resources. + // @return TTS_SUCCESS, or TTS_FAILURE + virtual tts_result shutdown(); + + // Interrupt synthesis and flushes any synthesized data that hasn't been output yet. + // This will block until callbacks underway are completed. + // @return TTS_SUCCESS, or TTS_FAILURE + virtual tts_result stop(); + + // Load the resources associated with the specified language. The loaded language will + // only be used once a call to setLanguage() with the same language value is issued. + // Language values are based on the Android conventions for localization as described in + // the Android platform documentation on internationalization. This implies that language + // data is specified in the format xx-rYY, where xx is a two letter ISO 639-1 language code + // in lowercase and rYY is a two letter ISO 3166-1-alpha-2 language code in uppercase + // preceded by a lowercase "r". + // @param value pointer to the language value + // @param size length of the language value + // @return TTS_SUCCESS, or TTS_FAILURE + virtual tts_result loadLanguage(const char *value, const size_t size); + + // Signal the engine to use the specified language. This will force the language to be + // loaded if it wasn't loaded previously with loadLanguage(). + // See loadLanguage for the specification of the language. + // @param value pointer to the language value + // @param size length of the language value + // @return TTS_SUCCESS, or TTS_FAILURE + virtual tts_result setLanguage(const char *value, const size_t size); + + // Retrieve the currently set language, or an empty "value" if no language has + // been set. + // @param[out] value pointer to the retrieved language value + // @param[inout] iosize in: stores the size available to store the language value in *value + // out: stores the size required to hold the language value if + // getLanguage() returned TTS_PROPERTY_SIZE_TOO_SMALL, + // unchanged otherwise. + // @return TTS_SUCCESS, or TTS_PROPERTY_SIZE_TOO_SMALL, or TTS_FAILURE + virtual tts_result getLanguage(char *value, size_t *iosize); + + // Set a property for the the TTS engine + // "size" is the maximum size of "value" for properties "property" + // @param property pointer to the property name + // @param value pointer to the property value + // @param size maximum size required to store this type of property + // @return TTS_PROPERTY_UNSUPPORTED, or TTS_SUCCESS, or TTS_FAILURE, + // or TTS_VALUE_INVALID + virtual tts_result setProperty(const char *property, const char *value, const size_t size); + + // Retrieve a property from the TTS engine + // @param property pointer to the property name + // @param[out] value pointer to the retrieved language value + // @param[inout] iosize in: stores the size available to store the property value + // out: stores the size required to hold the language value if + // getLanguage() returned TTS_PROPERTY_SIZE_TOO_SMALL, + // unchanged otherwise. + // @return TTS_PROPERTY_UNSUPPORTED, or TTS_SUCCESS, or TTS_PROPERTY_SIZE_TOO_SMALL + virtual tts_result getProperty(const char *property, char *value, size_t *iosize); + + // Synthesize the text. + // When synthesis completes, the engine invokes the callback to notify the TTS framework. + // Note about the format of the input: the text parameter may use the following elements + // and their respective attributes as defined in the SSML 1.0 specification: + // * lang + // * say-as: + // o interpret-as + // * phoneme + // * voice: + // o gender, + // o age, + // o variant, + // o name + // * emphasis + // * break: + // o strength, + // o time + // * prosody: + // o pitch, + // o contour, + // o range, + // o rate, + // o duration, + // o volume + // * mark + // Differences between this text format and SSML are: + // * full SSML documents are not supported + // * namespaces are not supported + // Text is coded in UTF-8. + // @param text the UTF-8 text to synthesize + // @param userdata pointer to be returned when the call is invoked + // @return TTS_SUCCESS or TTS_FAILURE + virtual tts_result synthesizeText(const char *text, void *userdata); + + // Synthesize IPA text. When synthesis completes, the engine must call the given callback to notify the TTS API. + // @param ipa the IPA data to synthesize + // @param userdata pointer to be returned when the call is invoked + // @return TTS_FEATURE_UNSUPPORTED if IPA is not supported, otherwise TTS_SUCCESS or TTS_FAILURE + virtual tts_result synthesizeIpa(const char *ipa, void *userdata); +}; + +} // namespace android + |