A specialised type of AudioFormatReader that uses a MemoryMappedFile to read directly from an audio file. More...
#include <juce_MemoryMappedAudioFormatReader.h>
Public Member Functions | |
virtual AudioChannelSet | getChannelLayout () |
Get the channel layout of the audio stream. More... | |
const File & | getFile () const noexcept |
Returns the file that is being mapped. More... | |
const String & | getFormatName () const noexcept |
Returns a description of what type of format this is. More... | |
Range< int64 > | getMappedSection () const noexcept |
Returns the sample range that's currently memory-mapped and available for reading. More... | |
size_t | getNumBytesUsed () const |
Returns the number of bytes currently being mapped. More... | |
virtual void | getSample (int64 sampleIndex, float *result) const noexcept=0 |
Returns the samples for all channels at a given sample position. More... | |
bool | mapEntireFile () |
Attempts to map the entire file into memory. More... | |
virtual bool | mapSectionOfFile (Range< int64 > samplesToMap) |
Attempts to map a section of the file into memory. More... | |
bool | read (int *const *destSamples, int numDestChannels, int64 startSampleInSource, int numSamplesToRead, bool fillLeftoverChannelsWithCopies) |
Reads samples from the stream. More... | |
void | read (AudioBuffer< float > *buffer, int startSampleInDestBuffer, int numSamples, int64 readerStartSample, bool useReaderLeftChan, bool useReaderRightChan) |
Fills a section of an AudioBuffer from this reader. More... | |
virtual void | readMaxLevels (int64 startSample, int64 numSamples, Range< float > *results, int numChannelsToRead) |
Finds the highest and lowest sample levels from a section of the audio stream. More... | |
virtual void | readMaxLevels (int64 startSample, int64 numSamples, float &lowestLeft, float &highestLeft, float &lowestRight, float &highestRight) |
Finds the highest and lowest sample levels from a section of the audio stream. More... | |
virtual bool | readSamples (int **destSamples, int numDestChannels, int startOffsetInDestBuffer, int64 startSampleInFile, int numSamples)=0 |
Subclasses must implement this method to perform the low-level read operation. More... | |
int64 | searchForLevel (int64 startSample, int64 numSamplesToSearch, double magnitudeRangeMinimum, double magnitudeRangeMaximum, int minimumConsecutiveSamples) |
Scans the source looking for a sample whose magnitude is in a specified range. More... | |
void | touchSample (int64 sample) const noexcept |
Touches the memory for the given sample, to force it to be loaded into active memory. More... | |
Public Attributes | |
unsigned int | bitsPerSample = 0 |
The number of bits per sample, e.g. More... | |
InputStream * | input |
The input stream, for use by subclasses. More... | |
int64 | lengthInSamples = 0 |
The total number of samples in the audio stream. More... | |
StringPairArray | metadataValues |
A set of metadata values that the reader has pulled out of the stream. More... | |
unsigned int | numChannels = 0 |
The total number of channels in the audio stream. More... | |
double | sampleRate = 0 |
The sample-rate of the stream. More... | |
bool | usesFloatingPointData = false |
Indicates whether the data is floating-point or fixed. More... | |
Protected Member Functions | |
MemoryMappedAudioFormatReader (const File &file, const AudioFormatReader &details, int64 dataChunkStart, int64 dataChunkLength, int bytesPerFrame) | |
Creates an MemoryMappedAudioFormatReader object. More... | |
int64 | filePosToSample (int64 filePos) const noexcept |
Converts a byte position in the file to a sample index. More... | |
int64 | sampleToFilePos (int64 sample) const noexcept |
Converts a sample index to a byte position in the file. More... | |
const void * | sampleToPointer (int64 sample) const noexcept |
Converts a sample index to a pointer to the mapped file memory. More... | |
template<typename SampleType , typename Endianness > | |
Range< float > | scanMinAndMaxInterleaved (int channel, int64 startSampleInFile, int64 numSamples) const noexcept |
Used by AudioFormatReader subclasses to scan for min/max ranges in interleaved data. More... | |
Static Protected Member Functions | |
static void | clearSamplesBeyondAvailableLength (int **destSamples, int numDestChannels, int startOffsetInDestBuffer, int64 startSampleInFile, int &numSamples, int64 fileLengthInSamples) |
Used by AudioFormatReader subclasses to clear any parts of the data blocks that lie beyond the end of their available length. More... | |
Protected Attributes | |
int | bytesPerFrame |
int64 | dataChunkStart |
int64 | dataLength |
File | file |
std::unique_ptr< MemoryMappedFile > | map |
Range< int64 > | mappedSection |
A specialised type of AudioFormatReader that uses a MemoryMappedFile to read directly from an audio file.
This allows for incredibly fast random-access to sample data in the mapped region of the file, but not all audio formats support it - see AudioFormat::createMemoryMappedReader().
Note that before reading samples from a MemoryMappedAudioFormatReader, you must first call mapEntireFile() or mapSectionOfFile() to ensure that the region you want to read has been mapped.
{Audio}
|
protected |
Creates an MemoryMappedAudioFormatReader object.
Note that before attempting to read any data, you must call mapEntireFile() or mapSectionOfFile() to ensure that the region you want to read has been mapped.
|
inlinestaticprotectedinherited |
Used by AudioFormatReader subclasses to clear any parts of the data blocks that lie beyond the end of their available length.
References int(), jassert, and juce::zeromem().
|
inlineprotectednoexcept |
Converts a byte position in the file to a sample index.
|
virtualinherited |
Get the channel layout of the audio stream.
|
inlinenoexcept |
Returns the file that is being mapped.
|
inlinenoexceptinherited |
Returns a description of what type of format this is.
E.g. "AIFF"
Returns the sample range that's currently memory-mapped and available for reading.
|
inline |
Returns the number of bytes currently being mapped.
|
pure virtualnoexcept |
Returns the samples for all channels at a given sample position.
The result array must be large enough to hold a value for each channel that this reader contains.
bool juce::MemoryMappedAudioFormatReader::mapEntireFile | ( | ) |
Attempts to map the entire file into memory.
|
virtual |
Attempts to map a section of the file into memory.
|
inherited |
Reads samples from the stream.
destSamples | an array of buffers into which the sample data for each channel will be written. If the format is fixed-point, each channel will be written as an array of 32-bit signed integers using the full range -0x80000000 to 0x7fffffff, regardless of the source's bit-depth. If it is a floating-point format, you should cast the resulting array to a (float**) to get the values (in the range -1.0 to 1.0 or beyond) If the format is stereo, then destSamples[0] is the left channel data, and destSamples[1] is the right channel. The numDestChannels parameter indicates how many pointers this array contains, but some of these pointers can be null if you don't want to read data for some of the channels |
numDestChannels | the number of array elements in the destChannels array |
startSampleInSource | the position in the audio file or stream at which the samples should be read, as a number of samples from the start of the stream. It's ok for this to be beyond the start or end of the available data - any samples that are out-of-range will be returned as zeros. |
numSamplesToRead | the number of samples to read. If this is greater than the number of samples that the file or stream contains. the result will be padded with zeros |
fillLeftoverChannelsWithCopies | if true, this indicates that if there's no source data available for some of the channels that you pass in, then they should be filled with copies of valid source channels. E.g. if you're reading a mono file and you pass 2 channels to this method, then if fillLeftoverChannelsWithCopies is true, both destination channels will be filled with the same data from the file's single channel. If fillLeftoverChannelsWithCopies was false, then only the first channel would be filled with the file's contents, and the second would be cleared. If there are many channels, e.g. you try to read 4 channels from a stereo file, then the last 3 would all end up with copies of the same data. |
|
inherited |
Fills a section of an AudioBuffer from this reader.
This will convert the reader's fixed- or floating-point data to the buffer's floating-point format, and will try to intelligently cope with mismatches between the number of channels in the reader and the buffer.
|
virtualinherited |
Finds the highest and lowest sample levels from a section of the audio stream.
This will read a block of samples from the stream, and measure the highest and lowest sample levels from the channels in that section, returning these as normalised floating-point levels.
startSample | the offset into the audio stream to start reading from. It's ok for this to be beyond the start or end of the stream. |
numSamples | how many samples to read |
results | this array will be filled with Range values for each channel. The array must contain numChannels elements. |
numChannelsToRead | the number of channels of data to scan. This must be more than zero, but not more than the total number of channels that the reader contains |
Reimplemented in juce::AudioSubsectionReader.
|
virtualinherited |
Finds the highest and lowest sample levels from a section of the audio stream.
This will read a block of samples from the stream, and measure the highest and lowest sample levels from the channels in that section, returning these as normalised floating-point levels.
startSample | the offset into the audio stream to start reading from. It's ok for this to be beyond the start or end of the stream. |
numSamples | how many samples to read |
lowestLeft | on return, this is the lowest absolute sample from the left channel |
highestLeft | on return, this is the highest absolute sample from the left channel |
lowestRight | on return, this is the lowest absolute sample from the right channel (if there is one) |
highestRight | on return, this is the highest absolute sample from the right channel (if there is one) |
|
pure virtualinherited |
Subclasses must implement this method to perform the low-level read operation.
Callers should use read() instead of calling this directly.
destSamples | the array of destination buffers to fill. Some of these pointers may be null |
numDestChannels | the number of items in the destSamples array. This value is guaranteed not to be greater than the number of channels that this reader object contains |
startOffsetInDestBuffer | the number of samples from the start of the dest data at which to begin writing |
startSampleInFile | the number of samples into the source data at which to begin reading. This value is guaranteed to be >= 0. |
numSamples | the number of samples to read |
Implemented in juce::AudioCDReader, juce::AudioSubsectionReader, and juce::BufferingAudioReader.
|
inlineprotectednoexcept |
Converts a sample index to a byte position in the file.
|
inlineprotectednoexcept |
Converts a sample index to a pointer to the mapped file memory.
References juce::addBytesToPointer().
|
inlineprotectednoexcept |
Used by AudioFormatReader subclasses to scan for min/max ranges in interleaved data.
References juce::addBytesToPointer(), juce::AudioData::Pointer< SampleFormat, Endianness, InterleavingType, Constness >::findMinAndMax(), and JUCE_DECLARE_NON_COPYABLE_WITH_LEAK_DETECTOR.
|
inherited |
Scans the source looking for a sample whose magnitude is in a specified range.
This will read from the source, either forwards or backwards between two sample positions, until it finds a sample whose magnitude lies between two specified levels.
If it finds a suitable sample, it returns its position; if not, it will return -1.
There's also a minimumConsecutiveSamples setting to help avoid spikes or zero-crossing points when you're searching for a continuous range of samples
startSample | the first sample to look at |
numSamplesToSearch | the number of samples to scan. If this value is negative, the search will go backwards |
magnitudeRangeMinimum | the lowest magnitude (inclusive) that is considered a hit, from 0 to 1.0 |
magnitudeRangeMaximum | the highest magnitude (inclusive) that is considered a hit, from 0 to 1.0 |
minimumConsecutiveSamples | if this is > 0, the method will only look for a sequence of this many consecutive samples, all of which lie within the target range. When it finds such a sequence, it returns the position of the first in-range sample it found (i.e. the earliest one if scanning forwards, the latest one if scanning backwards) |
|
noexcept |
Touches the memory for the given sample, to force it to be loaded into active memory.
|
inherited |
The number of bits per sample, e.g.
16, 24, 32.
|
protected |
|
protected |
|
protected |
|
protected |
|
inherited |
The input stream, for use by subclasses.
|
inherited |
The total number of samples in the audio stream.
|
protected |
|
inherited |
A set of metadata values that the reader has pulled out of the stream.
Exactly what these values are depends on the format, so you can check out the format implementation code to see what kind of stuff they understand.
|
inherited |
The total number of channels in the audio stream.
|
inherited |
The sample-rate of the stream.
|
inherited |
Indicates whether the data is floating-point or fixed.