A MovieAudio is actually any source that provides a sequence of audio samples. More...
#include "movieAudioCursor.h"
Public Member Functions | |
MovieAudioCursor (MovieAudio *src) | |
This constructor returns a null audio stream — a stream of total silence, at 8000 samples per second. More... | |
bool | aborted () const |
int | audio_channels () const |
int | audio_rate () const |
bool | can_seek () const |
bool | can_seek_fast () const |
virtual TypeHandle | force_init_type () |
virtual TypeHandle | get_type () const |
double | length () const |
PT (MovieAudio) get_source() const | |
void | read_samples (int n, Datagram *dg) |
Read audio samples from the stream into a Datagram. More... | |
std::string | read_samples (int n) |
Read audio samples from the stream and returns them as a string. More... | |
virtual void | read_samples (int n, int16_t *data) |
Read audio samples from the stream. More... | |
virtual int | ready () const |
Returns the number of audio samples that are ready to read. More... | |
virtual void | seek (double offset) |
Skips to the specified offset within the file. More... | |
void | skip_samples (int n) |
double | tell () const |
![]() | |
TypedWritableReferenceCount (const TypedWritableReferenceCount ©) | |
virtual ReferenceCount * | as_reference_count () |
Returns the pointer cast to a ReferenceCount pointer, if it is in fact of that type. More... | |
void | operator= (const TypedWritableReferenceCount ©) |
![]() | |
TypedWritable (const TypedWritable ©) | |
virtual int | complete_pointers (TypedWritable **p_list, BamReader *manager) |
Receives an array of pointers, one for each time manager->read_pointer() was called in fillin(). More... | |
vector_uchar | encode_to_bam_stream () const |
Converts the TypedWritable object into a single stream of data using a BamWriter, and returns that data as a bytes object. More... | |
bool | encode_to_bam_stream (vector_uchar &data, BamWriter *writer=nullptr) const |
Converts the TypedWritable object into a single stream of data using a BamWriter, and stores that data in the indicated string. More... | |
virtual void | fillin (DatagramIterator &scan, BamReader *manager) |
This internal function is intended to be called by each class's make_from_bam() method to read in all of the relevant data from the BamFile for the new object. More... | |
virtual void | finalize (BamReader *manager) |
Called by the BamReader to perform any final actions needed for setting up the object after all objects have been read and all pointers have been completed. More... | |
UpdateSeq | get_bam_modified () const |
Returns the current bam_modified counter. More... | |
void | mark_bam_modified () |
Increments the bam_modified counter, so that this object will be invalidated and retransmitted on any open bam streams. More... | |
void | operator= (const TypedWritable ©) |
virtual bool | require_fully_complete () const |
Some objects require all of their nested pointers to have been completed before the objects themselves can be completed. More... | |
virtual void | update_bam_nested (BamWriter *manager) |
Called by the BamWriter when this object has not itself been modified recently, but it should check its nested objects for updates. More... | |
virtual void | write_datagram (BamWriter *manager, Datagram &dg) |
Writes the contents of this object to the datagram for shipping out to a Bam file. More... | |
![]() | |
TypedObject (const TypedObject ©)=default | |
TypedObject * | as_typed_object () |
Returns the object, upcast (if necessary) to a TypedObject pointer. More... | |
const TypedObject * | as_typed_object () const |
Returns the object, upcast (if necessary) to a TypedObject pointer. More... | |
int | get_best_parent_from_Set (const std::set< int > &) const |
int | get_type_index () const |
Returns the internal index number associated with this object's TypeHandle, a unique number for each different type. More... | |
bool | is_exact_type (TypeHandle handle) const |
Returns true if the current object is the indicated type exactly. More... | |
bool | is_of_type (TypeHandle handle) const |
Returns true if the current object is or derives from the indicated type. More... | |
TypedObject & | operator= (const TypedObject ©)=default |
![]() | |
int | get_ref_count () const |
WeakReferenceList * | get_weak_list () const |
Returns the WeakReferenceList associated with this ReferenceCount object. More... | |
bool | has_weak_list () const |
Returns true if this particular ReferenceCount object has a WeakReferenceList created, false otherwise. More... | |
void | local_object () |
This function should be called, once, immediately after creating a new instance of some ReferenceCount-derived object on the stack. More... | |
void | ref () const |
Explicitly increments the reference count. More... | |
bool | ref_if_nonzero () const |
Atomically increases the reference count of this object if it is not zero. More... | |
bool | test_ref_count_integrity () const |
Does some easy checks to make sure that the reference count isn't completely bogus. More... | |
bool | test_ref_count_nonzero () const |
Does some easy checks to make sure that the reference count isn't zero, or completely bogus. More... | |
virtual bool | unref () const |
Explicitly decrements the reference count. More... | |
WeakReferenceList * | weak_ref () |
Adds the indicated PointerToVoid as a weak reference to this object. More... | |
void | weak_unref () |
Removes the indicated PointerToVoid as a weak reference to this object. More... | |
Static Public Member Functions | |
static TypeHandle | get_class_type () |
static void | init_type () |
![]() | |
static TypeHandle | get_class_type () |
static void | init_type () |
static | PT (TypedWritableReferenceCount) decode_from_bam_stream(vector_uchar data |
![]() | |
static bool | decode_raw_from_bam_stream (TypedWritable *&ptr, ReferenceCount *&ref_ptr, vector_uchar data, BamReader *reader=nullptr) |
Reads the bytes created by a previous call to encode_to_bam_stream(), and extracts the single object on those bytes. More... | |
static TypeHandle | get_class_type () |
static void | init_type () |
![]() | |
static TypeHandle | get_class_type () |
static void | init_type () |
This function is declared non-inline to work around a compiler bug in g++ 2.96. More... | |
![]() | |
static TypeHandle | get_class_type () |
static void | init_type () |
Additional Inherited Members | |
![]() | |
static BamReader * | reader = nullptr) |
![]() | |
get_type | |
![]() | |
get_ref_count | |
Returns the current reference count. More... | |
![]() | |
static TypedWritable *const | Null = nullptr |
A MovieAudio is actually any source that provides a sequence of audio samples.
That could include an AVI file, a microphone, or an internet TV station. A MovieAudioCursor is a handle that lets you read data sequentially from a MovieAudio.
Thread safety: each individual MovieAudioCursor must be owned and accessed by a single thread. It is OK for two different threads to open the same file at the same time, as long as they use separate MovieAudioCursor objects.
Definition at line 34 of file movieAudioCursor.h.
MovieAudioCursor::MovieAudioCursor | ( | MovieAudio * | src | ) |
This constructor returns a null audio stream — a stream of total silence, at 8000 samples per second.
To get more interesting audio, you need to construct a subclass of this class.
Definition at line 24 of file movieAudioCursor.cxx.
void MovieAudioCursor::read_samples | ( | int | n, |
Datagram * | dg | ||
) |
Read audio samples from the stream into a Datagram.
N is the number of samples you wish to read. Multiple-channel audio will be interleaved.
This is not particularly efficient, but it may be a convenient way to manipulate samples in python.
Definition at line 73 of file movieAudioCursor.cxx.
Referenced by read_samples().
std::string MovieAudioCursor::read_samples | ( | int | n | ) |
Read audio samples from the stream and returns them as a string.
The samples are stored little-endian in the string. N is the number of samples you wish to read. Multiple-channel audio will be interleaved.
This is not particularly efficient, but it may be a convenient way to manipulate samples in python.
Definition at line 96 of file movieAudioCursor.cxx.
References read_samples().
|
virtual |
Read audio samples from the stream.
N is the number of samples you wish to read. Your buffer must be equal in size to N * channels. Multiple-channel audio will be interleaved.
Reimplemented in FfmpegAudioCursor, FlacAudioCursor, UserDataAudioCursor, and WavAudioCursor.
Definition at line 49 of file movieAudioCursor.cxx.
|
virtual |
Returns the number of audio samples that are ready to read.
This is primarily relevant for sources like microphones which produce samples at a fixed rate. If you try to read more samples than are ready, the result will be silent samples.
Some audio streams do not have a limit on how fast they can produce samples. Such streams will always return 0x40000000 as the ready-count. This may well exceed the length of the audio stream. You therefore need to check length separately.
If the aborted flag is set, that means the ready count is no longer being replenished. For example, a MovieAudioCursor might be reading from an internet radio station, and it might buffer data to avoid underruns. If it loses connection to the radio station, it will set the aborted flag to indicate that the buffer is no longer being replenished. But it is still ok to read the samples that are in the buffer, at least until they run out. Once those are gone, there will be no more.
An audio consumer needs to check the length, the ready status, and the aborted flag.
Reimplemented in UserDataAudioCursor.
Definition at line 161 of file movieAudioCursor.cxx.
|
virtual |
Skips to the specified offset within the file.
If the movie reports that it cannot seek, then this method can still advance by reading samples and discarding them. However, to move backward, can_seek must be true.
If the movie reports that it can_seek, it doesn't mean that it can do so quickly. It may have to rewind the movie and then fast forward to the desired location. Only if can_seek_fast returns true can seek operations be done in constant time.
Seeking may not be precise, because AVI files often have inaccurate indices. After seeking, tell will indicate that the cursor is at the target location. However, in truth, the data you read may come from a slightly offset location.
Reimplemented in FfmpegAudioCursor, UserDataAudioCursor, FlacAudioCursor, and WavAudioCursor.
Definition at line 133 of file movieAudioCursor.cxx.