libgig  3.3.0.svn4
gig::Sample Class Reference

Encapsulates sample waves used for playback. More...

#include <gig.h>

Inheritance diagram for gig::Sample:
DLS::Sample DLS::Resource

Public Member Functions

buffer_t LoadSampleData ()
 Loads (and uncompresses if needed) the whole sample wave into RAM. More...
 
buffer_t LoadSampleData (unsigned long SampleCount)
 Reads (uncompresses if needed) and caches the first SampleCount numbers of SamplePoints in RAM. More...
 
buffer_t LoadSampleDataWithNullSamplesExtension (uint NullSamplesCount)
 Loads (and uncompresses if needed) the whole sample wave into RAM. More...
 
buffer_t LoadSampleDataWithNullSamplesExtension (unsigned long SampleCount, uint NullSamplesCount)
 Reads (uncompresses if needed) and caches the first SampleCount numbers of SamplePoints in RAM. More...
 
buffer_t GetCache ()
 Returns current cached sample points. More...
 
void ReleaseSampleData ()
 Frees the cached sample from RAM if loaded with LoadSampleData() previously. More...
 
void Resize (int iNewSize)
 Resize sample. More...
 
unsigned long SetPos (unsigned long SampleCount, RIFF::stream_whence_t Whence=RIFF::stream_start)
 Sets the position within the sample (in sample points, not in bytes). More...
 
unsigned long GetPos ()
 Returns the current position in the sample (in sample points). More...
 
unsigned long Read (void *pBuffer, unsigned long SampleCount, buffer_t *pExternalDecompressionBuffer=NULL)
 Reads SampleCount number of sample points from the current position into the buffer pointed by pBuffer and increments the position within the sample. More...
 
unsigned long ReadAndLoop (void *pBuffer, unsigned long SampleCount, playback_state_t *pPlaybackState, DimensionRegion *pDimRgn, buffer_t *pExternalDecompressionBuffer=NULL)
 Reads SampleCount number of sample points from the position stored in pPlaybackState into the buffer pointed by pBuffer and moves the position within the sample respectively, this method honors the looping informations of the sample (if any). More...
 
unsigned long Write (void *pBuffer, unsigned long SampleCount)
 Write sample wave data. More...
 
GroupGetGroup () const
 Returns pointer to the Group this Sample belongs to. More...
 
virtual void UpdateChunks ()
 Apply sample and its settings to the respective RIFF chunks. More...
 
unsigned long GetSize ()
 Returns sample size. More...
 
unsigned long Read (void *pBuffer, unsigned long SampleCount)
 Reads SampleCount number of sample points from the current position into the buffer pointed by pBuffer and increments the position within the sample. More...
 
ResourceGetParent ()
 
void GenerateDLSID ()
 Generates a new DLSID for the resource. More...
 
virtual void CopyAssign (const Resource *orig)
 Make a deep copy of the Resource object given by orig and assign it to this object. More...
 

Static Public Member Functions

static buffer_t CreateDecompressionBuffer (unsigned long MaxReadSize)
 Allocates a decompression buffer for streaming (compressed) samples with Sample::Read(). More...
 
static void DestroyDecompressionBuffer (buffer_t &DecompressionBuffer)
 Free decompression buffer, previously created with CreateDecompressionBuffer(). More...
 

Public Attributes

uint32_t Manufacturer
 Specifies the MIDI Manufacturer's Association (MMA) Manufacturer code for the sampler intended to receive this file's waveform. If no particular manufacturer is to be specified, a value of 0 should be used. More...
 
uint32_t Product
 Specifies the MIDI model ID defined by the manufacturer corresponding to the Manufacturer field. If no particular manufacturer's product is to be specified, a value of 0 should be used. More...
 
uint32_t SamplePeriod
 Specifies the duration of time that passes during the playback of one sample in nanoseconds (normally equal to 1 / Samples Per Second, where Samples Per Second is the value found in the format chunk), don't bother to update this attribute, it won't be saved. More...
 
uint32_t MIDIUnityNote
 Specifies the musical note at which the sample will be played at it's original sample rate. More...
 
uint32_t FineTune
 Specifies the fraction of a semitone up from the specified MIDI unity note field. A value of 0x80000000 means 1/2 semitone (50 cents) and a value of 0x00000000 means no fine tuning between semitones. More...
 
smpte_format_t SMPTEFormat
 Specifies the Society of Motion Pictures and Television E time format used in the following SMPTEOffset field. If a value of 0 is set, SMPTEOffset should also be set to 0. More...
 
uint32_t SMPTEOffset
 The SMPTE Offset value specifies the time offset to be used for the synchronization / calibration to the first sample in the waveform. This value uses a format of 0xhhmmssff where hh is a signed value that specifies the number of hours (-23 to 23), mm is an unsigned value that specifies the number of minutes (0 to 59), ss is an unsigned value that specifies the number of seconds (0 to 59) and ff is an unsigned value that specifies the number of frames (0 to -1). More...
 
uint32_t Loops
 Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: Number of defined sample loops. So far only seen single loops in gig files - please report if you encounter more!) More...
 
uint32_t LoopID
 Specifies the unique ID that corresponds to one of the defined cue points in the cue point list (only if Loops > 0), as the Gigasampler format only allows one loop definition at the moment, this attribute isn't really useful for anything. More...
 
loop_type_t LoopType
 Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: The type field defines how the waveform samples will be looped.) More...
 
uint32_t LoopStart
 Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: The start value specifies the offset [in sample points] in the waveform data of the first sample to be played in the loop [only if Loops > 0].) More...
 
uint32_t LoopEnd
 Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: The end value specifies the offset [in sample points] in the waveform data which represents the end of the loop [only if Loops > 0].) More...
 
uint32_t LoopSize
 Caution: Use the respective fields in the DimensionRegion instead of this one! (Intended purpose: Length of the looping area [in sample points] which is equivalent to More...
 
uint32_t LoopFraction
 The fractional value specifies a fraction of a sample at which to loop. This allows a loop to be fine tuned at a resolution greater than one sample. A value of 0 means no fraction, a value of 0x80000000 means 1/2 of a sample length. 0xFFFFFFFF is the smallest fraction of a sample that can be represented. More...
 
uint32_t LoopPlayCount
 Number of times the loop should be played (a value of 0 = infinite). More...
 
bool Compressed
 If the sample wave is compressed (probably just interesting for instrument and sample editors, as this library already handles the decompression in it's sample access methods anyway). More...
 
uint32_t TruncatedBits
 For 24-bit compressed samples only: number of bits truncated during compression (0, 4 or 6) More...
 
bool Dithered
 For 24-bit compressed samples only: if dithering was used during compression with bit reduction. More...
 
uint16_t FormatTag
 Format ID of the waveform data (should be DLS_WAVE_FORMAT_PCM for DLS1 compliant files, this is also the default value if Sample was created with Instrument::AddSample()). More...
 
uint16_t Channels
 Number of channels represented in the waveform data, e.g. 1 for mono, 2 for stereo (defaults to 1=mono if Sample was created with Instrument::AddSample() previously). More...
 
uint32_t SamplesPerSecond
 Sampling rate at which each channel should be played (defaults to 44100 if Sample was created with Instrument::AddSample() previously). More...
 
uint32_t AverageBytesPerSecond
 The average number of bytes per second at which the waveform data should be transferred (Playback software can estimate the buffer size using this value). More...
 
uint16_t BlockAlign
 The block alignment (in bytes) of the waveform data. Playback software needs to process a multiple of BlockAlign bytes of data at a time, so the value of BlockAlign can be used for buffer alignment. More...
 
uint16_t BitDepth
 Size of each sample per channel (only if known sample data format is used, 0 otherwise). More...
 
unsigned long SamplesTotal
 Reflects total number of sample points (only if known sample data format is used, 0 otherwise), do not bother to change this value, it will not be saved. More...
 
uint FrameSize
 Reflects the size (in bytes) of one single sample point (only if known sample data format is used, 0 otherwise). Caution: with the current version of libgig you have to upate this field by yourself whenever you change one of the following fields: Channels, BitDepth ! Ignoring this might lead to undesired behavior when i.e. calling Resize(), SetPos(), Write() or Read(). More...
 
Info * pInfo
 Points (in any case) to an Info object, providing additional, optional infos and comments. More...
 
dlsid_t * pDLSID
 Points to a dlsid_t structure if the file provided a DLS ID else is NULL. More...
 

Protected Member Functions

 Sample (File *pFile, RIFF::List *waveList, unsigned long WavePoolOffset, unsigned long fileNo=0)
 Constructor. More...
 
 ~Sample ()
 Destructor. More...
 
unsigned long GuessSize (unsigned long samples)
 
unsigned long WorstCaseMaxSamples (buffer_t *pDecompressionBuffer)
 

Protected Attributes

GrouppGroup
 pointer to the Group this sample belongs to (always not-NULL) More...
 
unsigned long FrameOffset
 Current offset (sample points) in current sample frame (for decompression only). More...
 
unsigned long * FrameTable
 For positioning within compressed samples only: stores the offset values for each frame. More...
 
unsigned long SamplePos
 For compressed samples only: stores the current position (in sample points). More...
 
unsigned long SamplesInLastFrame
 For compressed samples only: length of the last sample frame. More...
 
unsigned long WorstCaseFrameSize
 For compressed samples only: size (in bytes) of the largest possible sample frame. More...
 
unsigned long SamplesPerFrame
 For compressed samples only: number of samples in a full sample frame. More...
 
buffer_t RAMCache
 Buffers samples (already uncompressed) in RAM. More...
 
unsigned long FileNo
 File number (> 0 when sample is stored in an extension file, 0 when it's in the gig) More...
 
RIFF::ChunkpCk3gix
 
RIFF::ChunkpCkSmpl
 
uint32_t crc
 CRC-32 checksum of the raw sample data. More...
 
RIFF::ListpWaveList
 
RIFF::ChunkpCkData
 
RIFF::ChunkpCkFormat
 
unsigned long ulWavePoolOffset
 
ResourcepParent
 
RIFF::ListpResourceList
 

Static Protected Attributes

static unsigned int Instances = 0
 Number of instances of class Sample. More...
 
static buffer_t InternalDecompressionBuffer
 Buffer used for decompression as well as for truncation of 24 Bit -> 16 Bit samples. More...
 

Friends

class File
 
class Region
 
class Group
 

Detailed Description

Encapsulates sample waves used for playback.

In case you created a new sample with File::AddSample(), you should first update all attributes with the desired meta informations (amount of channels, bit depth, sample rate, etc.), then call Resize() with the desired sample size, followed by File::Save(), this will create the mandatory RIFF chunk which will hold the sample wave data and / or resize the file so you will be able to Write() the sample data directly to disk.

Caution: for gig synthesis, most looping relevant information are retrieved from the respective DimensionRegon instead from the Sample itself. This was made for allowing different loop definitions for the same sample under different conditions.

Definition at line 526 of file gig.h.

Constructor & Destructor Documentation

gig::Sample::Sample ( File pFile,
RIFF::List waveList,
unsigned long  WavePoolOffset,
unsigned long  fileNo = 0 
)
protected

Constructor.

Load an existing sample or create a new one. A 'wave' list chunk must be given to this constructor. In case the given 'wave' list chunk contains a 'fmt', 'data' (and optionally a '3gix', 'smpl') chunk, the format and sample data will be loaded from there, otherwise default values will be used and those chunks will be created when File::Save() will be called later on.

Parameters
pFile- pointer to gig::File where this sample is located (or will be located)
waveList- pointer to 'wave' list chunk which is (or will be) associated with this sample
WavePoolOffset- offset of this sample data from wave pool ('wvpl') list chunk
fileNo- number of an extension file where this sample is located, 0 otherwise

Definition at line 369 of file gig.cpp.

References DLS::Sample::BitDepth, DLS::Sample::Channels, CHUNK_ID_3GIX, CHUNK_ID_EWAV, CHUNK_ID_INAM, CHUNK_ID_SMPL, Compressed, Dithered, FileNo, FineTune, FrameOffset, FrameTable, gig::File::GetGroup(), RIFF::List::GetSubChunk(), INITIAL_SAMPLE_BUFFER_SIZE, Instances, InternalDecompressionBuffer, gig::loop_type_normal, LoopEnd, LoopFraction, LoopID, LoopPlayCount, Loops, LoopSize, LoopStart, LoopType, Manufacturer, MIDIUnityNote, gig::buffer_t::NullExtensionSize, pCk3gix, pCkSmpl, pGroup, DLS::Resource::pInfo, Product, gig::buffer_t::pStart, RAMCache, RIFF::Chunk::Read(), RIFF::Chunk::ReadInt16(), RIFF::Chunk::ReadInt32(), SamplePeriod, SamplePos, DLS::Sample::SamplesPerSecond, DLS::Info::SetFixedStringLengths(), RIFF::Chunk::SetPos(), gig::buffer_t::Size, gig::smpte_format_no_offset, SMPTEFormat, SMPTEOffset, and TruncatedBits.

gig::Sample::~Sample ( )
protectedvirtual

Destructor.

Removes RIFF chunks associated with this Sample and frees all memory occupied by this sample.

Reimplemented from DLS::Sample.

Definition at line 1295 of file gig.cpp.

References FrameTable, Instances, InternalDecompressionBuffer, gig::buffer_t::pStart, RAMCache, and gig::buffer_t::Size.

Member Function Documentation

void DLS::Resource::CopyAssign ( const Resource orig)
virtualinherited

Make a deep copy of the Resource object given by orig and assign it to this object.

Parameters
orig- original Resource object to be copied from

Definition at line 534 of file DLS.cpp.

References DLS::Info::CopyAssign(), and DLS::Resource::pInfo.

Referenced by DLS::Region::CopyAssign(), and DLS::Instrument::CopyAssignCore().

buffer_t gig::Sample::CreateDecompressionBuffer ( unsigned long  MaxReadSize)
static

Allocates a decompression buffer for streaming (compressed) samples with Sample::Read().

If you are using more than one streaming thread in your application you HAVE to create a decompression buffer for EACH of your streaming threads and provide it with the Sample::Read() call in order to avoid race conditions and crashes.

You should free the memory occupied by the allocated buffer(s) once you don't need one of your streaming threads anymore by calling DestroyDecompressionBuffer().

Parameters
MaxReadSize- the maximum size (in sample points) you ever expect to read with one Read() call
Returns
allocated decompression buffer
See Also
DestroyDecompressionBuffer()

Definition at line 1257 of file gig.cpp.

References gig::buffer_t::NullExtensionSize, gig::buffer_t::pStart, and gig::buffer_t::Size.

void gig::Sample::DestroyDecompressionBuffer ( buffer_t DecompressionBuffer)
static

Free decompression buffer, previously created with CreateDecompressionBuffer().

Parameters
DecompressionBuffer- previously allocated decompression buffer to free

Definition at line 1274 of file gig.cpp.

References gig::buffer_t::NullExtensionSize, gig::buffer_t::pStart, and gig::buffer_t::Size.

void DLS::Resource::GenerateDLSID ( )
inherited

Generates a new DLSID for the resource.

Definition at line 487 of file DLS.cpp.

References DLS::dlsid_t::abData, DLS::Resource::pDLSID, DLS::dlsid_t::ulData1, DLS::dlsid_t::usData2, and DLS::dlsid_t::usData3.

Referenced by gig::File::AddInstrument(), and gig::File::File().

buffer_t gig::Sample::GetCache ( )

Returns current cached sample points.

A buffer_t structure will be returned which contains address pointer to the begin of the cache and the size of the cached sample data in bytes. Use LoadSampleData() to cache a specific amount of sample points in RAM.

Returns
buffer_t structure with current cached sample points
See Also
LoadSampleData();

Definition at line 699 of file gig.cpp.

References gig::buffer_t::NullExtensionSize, gig::buffer_t::pStart, RAMCache, and gig::buffer_t::Size.

Referenced by LoadSampleDataWithNullSamplesExtension().

Group * gig::Sample::GetGroup ( ) const

Returns pointer to the Group this Sample belongs to.

In the .gig format a sample always belongs to one group. If it wasn't explicitly assigned to a certain group, it will be automatically assigned to a default group.

Returns
Sample's Group (never NULL)

Definition at line 1291 of file gig.cpp.

References pGroup.

unsigned long gig::Sample::GetPos ( )

Returns the current position in the sample (in sample points).

Definition at line 811 of file gig.cpp.

References Compressed, DLS::Sample::FrameSize, RIFF::Chunk::GetPos(), DLS::Sample::pCkData, and SamplePos.

Referenced by ReadAndLoop().

unsigned long DLS::Sample::GetSize ( )
inherited

Returns sample size.

Returns the sample wave form's data size (in sample points). This is actually the current, physical size (converted to sample points) of the RIFF chunk which encapsulates the sample's wave data. The returned value is dependant to the current FrameSize value.

Returns
number of sample points or 0 if FormatTag != DLS_WAVE_FORMAT_PCM
See Also
FrameSize, FormatTag

Definition at line 797 of file DLS.cpp.

References DLS_WAVE_FORMAT_PCM, DLS::Sample::FormatTag, DLS::Sample::FrameSize, RIFF::Chunk::GetSize(), and DLS::Sample::pCkData.

Referenced by DLS::Sample::Write(), and Write().

unsigned long gig::Sample::GuessSize ( unsigned long  samples)
inlineprotected

Definition at line 586 of file gig.h.

References DLS::Sample::BitDepth, DLS::Sample::Channels, and WorstCaseFrameSize.

Referenced by Read().

buffer_t gig::Sample::LoadSampleData ( )

Loads (and uncompresses if needed) the whole sample wave into RAM.

Use ReleaseSampleData() to free the memory if you don't need the cached sample data anymore.

Returns
buffer_t structure with start address and size of the buffer in bytes
See Also
ReleaseSampleData(), Read(), SetPos()

Definition at line 591 of file gig.cpp.

References LoadSampleDataWithNullSamplesExtension(), and DLS::Sample::SamplesTotal.

buffer_t gig::Sample::LoadSampleData ( unsigned long  SampleCount)

Reads (uncompresses if needed) and caches the first SampleCount numbers of SamplePoints in RAM.

Use ReleaseSampleData() to free the memory space if you don't need the cached samples anymore. There is no guarantee that exactly SampleCount samples will be cached; this is not an error. The size will be eventually truncated e.g. to the beginning of a frame of a compressed sample. This is done for efficiency reasons while streaming the wave by your sampler engine later. Read the Size member of the buffer_t structure that will be returned to determine the actual cached samples, but note that the size is given in bytes! You get the number of actually cached samples by dividing it by the frame size of the sample:

* buffer_t buf = pSample->LoadSampleData(acquired_samples);
* long cachedsamples = buf.Size / pSample->FrameSize;
*
Parameters
SampleCount- number of sample points to load into RAM
Returns
buffer_t structure with start address and size of the cached sample data in bytes
See Also
ReleaseSampleData(), Read(), SetPos()

Definition at line 617 of file gig.cpp.

References LoadSampleDataWithNullSamplesExtension().

buffer_t gig::Sample::LoadSampleDataWithNullSamplesExtension ( uint  NullSamplesCount)

Loads (and uncompresses if needed) the whole sample wave into RAM.

Use ReleaseSampleData() to free the memory if you don't need the cached sample data anymore. The method will add NullSamplesCount silence samples past the official buffer end (this won't affect the 'Size' member of the buffer_t structure, that means 'Size' always reflects the size of the actual sample data, the buffer might be bigger though). Silence samples past the official buffer are needed for differential algorithms that always have to take subsequent samples into account (resampling/interpolation would be an important example) and avoids memory access faults in such cases.

Parameters
NullSamplesCount- number of silence samples the buffer should be extended past it's data end
Returns
buffer_t structure with start address and size of the buffer in bytes
See Also
ReleaseSampleData(), Read(), SetPos()

Definition at line 640 of file gig.cpp.

References DLS::Sample::SamplesTotal.

Referenced by LoadSampleData().

buffer_t gig::Sample::LoadSampleDataWithNullSamplesExtension ( unsigned long  SampleCount,
uint  NullSamplesCount 
)

Reads (uncompresses if needed) and caches the first SampleCount numbers of SamplePoints in RAM.

Use ReleaseSampleData() to free the memory space if you don't need the cached samples anymore. There is no guarantee that exactly SampleCount samples will be cached; this is not an error. The size will be eventually truncated e.g. to the beginning of a frame of a compressed sample. This is done for efficiency reasons while streaming the wave by your sampler engine later. Read the Size member of the buffer_t structure that will be returned to determine the actual cached samples, but note that the size is given in bytes! You get the number of actually cached samples by dividing it by the frame size of the sample:

* buffer_t buf = pSample->LoadSampleDataWithNullSamplesExtension(acquired_samples, null_samples);
* long cachedsamples = buf.Size / pSample->FrameSize;
*

The method will add NullSamplesCount silence samples past the official buffer end (this won't affect the 'Size' member of the buffer_t structure, that means 'Size' always reflects the size of the actual sample data, the buffer might be bigger though). Silence samples past the official buffer are needed for differential algorithms that always have to take subsequent samples into account (resampling/interpolation would be an important example) and avoids memory access faults in such cases.

Parameters
SampleCount- number of sample points to load into RAM
NullSamplesCount- number of silence samples the buffer should be extended past it's data end
Returns
buffer_t structure with start address and size of the cached sample data in bytes
See Also
ReleaseSampleData(), Read(), SetPos()

Definition at line 676 of file gig.cpp.

References DLS::Sample::FrameSize, GetCache(), gig::buffer_t::NullExtensionSize, gig::buffer_t::pStart, RAMCache, Read(), DLS::Sample::SamplesTotal, SetPos(), and gig::buffer_t::Size.

unsigned long DLS::Sample::Read ( void *  pBuffer,
unsigned long  SampleCount 
)
inherited

Reads SampleCount number of sample points from the current position into the buffer pointed by pBuffer and increments the position within the sample.

Use this method and SetPos() if you don't want to load the sample into RAM, thus for disk streaming.

Parameters
pBufferdestination buffer
SampleCountnumber of sample points to read

Definition at line 873 of file DLS.cpp.

References DLS_WAVE_FORMAT_PCM, DLS::Sample::FormatTag, DLS::Sample::FrameSize, DLS::Sample::pCkData, and RIFF::Chunk::Read().

unsigned long gig::Sample::Read ( void *  pBuffer,
unsigned long  SampleCount,
buffer_t pExternalDecompressionBuffer = NULL 
)

Reads SampleCount number of sample points from the current position into the buffer pointed by pBuffer and increments the position within the sample.

The sample wave stream will be decompressed on the fly if using a compressed sample. Use this method and SetPos() if you don't want to load the sample into RAM, thus for disk streaming.

Caution: If you are using more than one streaming thread, you have to use an external decompression buffer for EACH streaming thread to avoid race conditions and crashes!

For 16 bit samples, the data in the buffer will be int16_t (using native endianness). For 24 bit, the buffer will contain three bytes per sample, little-endian.

Parameters
pBufferdestination buffer
SampleCountnumber of sample points to read
pExternalDecompressionBuffer(optional) external buffer to use for decompression
Returns
number of successfully read sample points
See Also
SetPos(), CreateDecompressionBuffer()

Definition at line 1025 of file gig.cpp.

References DLS::Sample::BitDepth, DLS::Sample::Channels, Compressed, FrameOffset, DLS::Sample::FrameSize, GuessSize(), InternalDecompressionBuffer, DLS::Sample::pCkData, gig::buffer_t::pStart, RIFF::Chunk::Read(), RIFF::Chunk::RemainingBytes(), SamplePos, SamplesInLastFrame, SamplesPerFrame, DLS::Sample::SamplesTotal, RIFF::Chunk::SetPos(), gig::buffer_t::Size, RIFF::stream_backward, RIFF::stream_ready, TruncatedBits, and WorstCaseMaxSamples().

Referenced by LoadSampleDataWithNullSamplesExtension(), and ReadAndLoop().

unsigned long gig::Sample::ReadAndLoop ( void *  pBuffer,
unsigned long  SampleCount,
playback_state_t pPlaybackState,
DimensionRegion pDimRgn,
buffer_t pExternalDecompressionBuffer = NULL 
)

Reads SampleCount number of sample points from the position stored in pPlaybackState into the buffer pointed by pBuffer and moves the position within the sample respectively, this method honors the looping informations of the sample (if any).

The sample wave stream will be decompressed on the fly if using a compressed sample. Use this method if you don't want to load the sample into RAM, thus for disk streaming. All this methods needs to know to proceed with streaming for the next time you call this method is stored in pPlaybackState. You have to allocate and initialize the playback_state_t structure by yourself before you use it to stream a sample:

* gig::playback_state_t playbackstate;
* playbackstate.position = 0;
* playbackstate.reverse = false;
* playbackstate.loop_cycles_left = pSample->LoopPlayCount;
*

You don't have to take care of things like if there is actually a loop defined or if the current read position is located within a loop area. The method already handles such cases by itself.

Caution: If you are using more than one streaming thread, you have to use an external decompression buffer for EACH streaming thread to avoid race conditions and crashes!

Parameters
pBufferdestination buffer
SampleCountnumber of sample points to read
pPlaybackStatewill be used to store and reload the playback state for the next ReadAndLoop() call
pDimRgndimension region with looping information
pExternalDecompressionBuffer(optional) external buffer to use for decompression
Returns
number of successfully read sample points
See Also
CreateDecompressionBuffer()

Definition at line 850 of file gig.cpp.

References DLS::Sample::FrameSize, GetPos(), gig::playback_state_t::loop_cycles_left, gig::loop_type_backward, gig::loop_type_bidirectional, gig::loop_type_normal, DLS::sample_loop_t::LoopLength, LoopPlayCount, DLS::sample_loop_t::LoopStart, DLS::sample_loop_t::LoopType, gig::playback_state_t::position, DLS::Sampler::pSampleLoops, Read(), gig::playback_state_t::reverse, DLS::Sampler::SampleLoops, and SetPos().

void gig::Sample::ReleaseSampleData ( )

Frees the cached sample from RAM if loaded with LoadSampleData() previously.

See Also
LoadSampleData();

Definition at line 714 of file gig.cpp.

References gig::buffer_t::NullExtensionSize, gig::buffer_t::pStart, RAMCache, and gig::buffer_t::Size.

void gig::Sample::Resize ( int  iNewSize)

Resize sample.

Resizes the sample's wave form data, that is the actual size of sample wave data possible to be written for this sample. This call will return immediately and just schedule the resize operation. You should call File::Save() to actually perform the resize operation(s) "physically" to the file. As this can take a while on large files, it is recommended to call Resize() first on all samples which have to be resized and finally to call File::Save() to perform all those resize operations in one rush.

The actual size (in bytes) is dependant to the current FrameSize value. You may want to set FrameSize before calling Resize().

Caution: You cannot directly write (i.e. with Write()) to enlarged samples before calling File::Save() as this might exceed the current sample's boundary!

Also note: only DLS_WAVE_FORMAT_PCM is currently supported, that is FormatTag must be DLS_WAVE_FORMAT_PCM. Trying to resize samples with other formats will fail!

Parameters
iNewSize- new sample wave data size in sample points (must be greater than zero)
Exceptions
DLS::Excecptionif FormatTag != DLS_WAVE_FORMAT_PCM or if iNewSize is less than 1
gig::Exceptionif existing sample is compressed
See Also
DLS::Sample::GetSize(), DLS::Sample::FrameSize, DLS::Sample::FormatTag, File::Save()

Definition at line 751 of file gig.cpp.

References Compressed, and DLS::Sample::Resize().

unsigned long gig::Sample::SetPos ( unsigned long  SampleCount,
RIFF::stream_whence_t  Whence = RIFF::stream_start 
)

Sets the position within the sample (in sample points, not in bytes).

Use this method and Read() if you don't want to load the sample into RAM, thus for disk streaming.

Although the original Gigasampler engine doesn't allow positioning within compressed samples, I decided to implement it. Even though the Gigasampler format doesn't allow to define loops for compressed samples at the moment, positioning within compressed samples might be interesting for some sampler engines though. The only drawback about my decision is that it takes longer to load compressed gig Files on startup, because it's neccessary to scan the samples for some mandatory informations. But I think as it doesn't affect the runtime efficiency, nobody will have a problem with that.

Parameters
SampleCountnumber of sample points to jump
Whenceoptional: to which relation SampleCount refers to, if omited RIFF::stream_start is assumed
Returns
the new sample position
See Also
Read()

Definition at line 777 of file gig.cpp.

References Compressed, FrameOffset, DLS::Sample::FrameSize, FrameTable, DLS::Sample::pCkData, SamplePos, DLS::Sample::SamplesTotal, RIFF::Chunk::SetPos(), RIFF::stream_backward, RIFF::stream_curpos, RIFF::stream_end, and RIFF::stream_start.

Referenced by LoadSampleDataWithNullSamplesExtension(), and ReadAndLoop().

void gig::Sample::UpdateChunks ( )
virtual

Apply sample and its settings to the respective RIFF chunks.

You have to call File::Save() to make changes persistent.

Usually there is absolutely no need to call this method explicitly. It will be called automatically when File::Save() was called.

Exceptions
DLS::Exceptionif FormatTag != DLS_WAVE_FORMAT_PCM or no sample data was provided yet
gig::Exceptionif there is any invalid sample setting

Reimplemented from DLS::Sample.

Definition at line 467 of file gig.cpp.

References RIFF::List::AddSubChunk(), CHUNK_ID_3GIX, CHUNK_ID_SMPL, FineTune, RIFF::List::GetSubChunk(), RIFF::Chunk::LoadChunkData(), LoopEnd, LoopFraction, LoopID, LoopPlayCount, Loops, LoopStart, LoopType, Manufacturer, MIDIUnityNote, pCk3gix, pCkSmpl, pGroup, DLS::Resource::pParent, Product, DLS::Sample::pWaveList, SamplePeriod, DLS::Sample::SamplesPerSecond, SMPTEFormat, SMPTEOffset, and DLS::Sample::UpdateChunks().

unsigned long gig::Sample::WorstCaseMaxSamples ( buffer_t pDecompressionBuffer)
inlineprotected

Definition at line 603 of file gig.h.

References SamplesPerFrame, gig::buffer_t::Size, and WorstCaseFrameSize.

Referenced by Read().

unsigned long gig::Sample::Write ( void *  pBuffer,
unsigned long  SampleCount 
)

Write sample wave data.

Writes SampleCount number of sample points from the buffer pointed by pBuffer and increments the position within the sample. Use this method to directly write the sample data to disk, i.e. if you don't want or cannot load the whole sample data into RAM.

You have to Resize() the sample to the desired size and call File::Save() before using Write().

Note: there is currently no support for writing compressed samples.

For 16 bit samples, the data in the source buffer should be int16_t (using native endianness). For 24 bit, the buffer should contain three bytes per sample, little-endian.

Parameters
pBuffer- source buffer
SampleCount- number of sample points to write
Exceptions
DLS::Exceptionif current sample size is too small
gig::Exceptionif sample is compressed
See Also
DLS::LoadSampleData()

Definition at line 1214 of file gig.cpp.

References DLS::Sample::BitDepth, Compressed, DLS::Sample::FrameSize, DLS::Resource::GetParent(), RIFF::Chunk::GetPos(), RIFF::Chunk::GetSize(), DLS::Sample::GetSize(), DLS::Sample::pCkData, gig::File::SetSampleChecksum(), and RIFF::Chunk::Write().

Friends And Related Function Documentation

friend class File
friend

Definition at line 608 of file gig.h.

friend class Group
friend

Definition at line 610 of file gig.h.

friend class Region
friend

Definition at line 609 of file gig.h.

Member Data Documentation

uint32_t DLS::Sample::AverageBytesPerSecond
inherited

The average number of bytes per second at which the waveform data should be transferred (Playback software can estimate the buffer size using this value).

Definition at line 379 of file DLS.h.

Referenced by DLS::Sample::Sample(), and DLS::Sample::UpdateChunks().

uint16_t DLS::Sample::BitDepth
inherited

Size of each sample per channel (only if known sample data format is used, 0 otherwise).

Definition at line 381 of file DLS.h.

Referenced by GuessSize(), Read(), DLS::Sample::Sample(), Sample(), DLS::Sample::UpdateChunks(), and Write().

uint16_t DLS::Sample::BlockAlign
inherited

The block alignment (in bytes) of the waveform data. Playback software needs to process a multiple of BlockAlign bytes of data at a time, so the value of BlockAlign can be used for buffer alignment.

Definition at line 380 of file DLS.h.

Referenced by DLS::Sample::Sample(), and DLS::Sample::UpdateChunks().

uint16_t DLS::Sample::Channels
inherited

Number of channels represented in the waveform data, e.g. 1 for mono, 2 for stereo (defaults to 1=mono if Sample was created with Instrument::AddSample() previously).

Definition at line 377 of file DLS.h.

Referenced by GuessSize(), Read(), DLS::Sample::Sample(), Sample(), DLS::Sample::UpdateChunks(), and gig::File::UpdateChunks().

bool gig::Sample::Compressed

If the sample wave is compressed (probably just interesting for instrument and sample editors, as this library already handles the decompression in it's sample access methods anyway).

Definition at line 543 of file gig.h.

Referenced by GetPos(), Read(), Resize(), Sample(), SetPos(), and Write().

uint32_t gig::Sample::crc
protected

CRC-32 checksum of the raw sample data.

Definition at line 580 of file gig.h.

bool gig::Sample::Dithered

For 24-bit compressed samples only: if dithering was used during compression with bit reduction.

Definition at line 545 of file gig.h.

Referenced by Sample().

unsigned long gig::Sample::FileNo
protected

File number (> 0 when sample is stored in an extension file, 0 when it's in the gig)

Definition at line 577 of file gig.h.

Referenced by gig::Region::GetSampleFromWavePool(), and Sample().

uint32_t gig::Sample::FineTune

Specifies the fraction of a semitone up from the specified MIDI unity note field. A value of 0x80000000 means 1/2 semitone (50 cents) and a value of 0x00000000 means no fine tuning between semitones.

Definition at line 532 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint16_t DLS::Sample::FormatTag
inherited

Format ID of the waveform data (should be DLS_WAVE_FORMAT_PCM for DLS1 compliant files, this is also the default value if Sample was created with Instrument::AddSample()).

Definition at line 376 of file DLS.h.

Referenced by DLS::Sample::GetSize(), DLS::Sample::Read(), DLS::Sample::Resize(), DLS::Sample::Sample(), DLS::Sample::SetPos(), DLS::Sample::UpdateChunks(), and DLS::Sample::Write().

unsigned long gig::Sample::FrameOffset
protected

Current offset (sample points) in current sample frame (for decompression only).

Definition at line 570 of file gig.h.

Referenced by Read(), Sample(), and SetPos().

uint DLS::Sample::FrameSize
inherited

Reflects the size (in bytes) of one single sample point (only if known sample data format is used, 0 otherwise). Caution: with the current version of libgig you have to upate this field by yourself whenever you change one of the following fields: Channels, BitDepth ! Ignoring this might lead to undesired behavior when i.e. calling Resize(), SetPos(), Write() or Read().

Definition at line 383 of file DLS.h.

Referenced by GetPos(), DLS::Sample::GetSize(), LoadSampleDataWithNullSamplesExtension(), DLS::Sample::Read(), Read(), ReadAndLoop(), DLS::Sample::Resize(), DLS::Sample::Sample(), DLS::Sample::SetPos(), SetPos(), DLS::Sample::Write(), and Write().

unsigned long* gig::Sample::FrameTable
protected

For positioning within compressed samples only: stores the offset values for each frame.

Definition at line 571 of file gig.h.

Referenced by Sample(), SetPos(), and ~Sample().

unsigned int gig::Sample::Instances = 0
staticprotected

Number of instances of class Sample.

Definition at line 567 of file gig.h.

Referenced by Sample(), and ~Sample().

buffer_t gig::Sample::InternalDecompressionBuffer
staticprotected

Buffer used for decompression as well as for truncation of 24 Bit -> 16 Bit samples.

Definition at line 568 of file gig.h.

Referenced by Read(), Sample(), and ~Sample().

uint32_t gig::Sample::LoopEnd

Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: The end value specifies the offset [in sample points] in the waveform data which represents the end of the loop [only if Loops > 0].)

Definition at line 539 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::LoopFraction

The fractional value specifies a fraction of a sample at which to loop. This allows a loop to be fine tuned at a resolution greater than one sample. A value of 0 means no fraction, a value of 0x80000000 means 1/2 of a sample length. 0xFFFFFFFF is the smallest fraction of a sample that can be represented.

Definition at line 541 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::LoopID

Specifies the unique ID that corresponds to one of the defined cue points in the cue point list (only if Loops > 0), as the Gigasampler format only allows one loop definition at the moment, this attribute isn't really useful for anything.

Definition at line 536 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::LoopPlayCount

Number of times the loop should be played (a value of 0 = infinite).

Definition at line 542 of file gig.h.

Referenced by ReadAndLoop(), Sample(), and UpdateChunks().

uint32_t gig::Sample::Loops

Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: Number of defined sample loops. So far only seen single loops in gig files - please report if you encounter more!)

Definition at line 535 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::LoopSize

Caution: Use the respective fields in the DimensionRegion instead of this one! (Intended purpose: Length of the looping area [in sample points] which is equivalent to

.)

Definition at line 540 of file gig.h.

Referenced by Sample().

uint32_t gig::Sample::LoopStart

Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: The start value specifies the offset [in sample points] in the waveform data of the first sample to be played in the loop [only if Loops > 0].)

Definition at line 538 of file gig.h.

Referenced by Sample(), and UpdateChunks().

loop_type_t gig::Sample::LoopType

Caution: Use the respective field in the DimensionRegion instead of this one! (Intended purpose: The type field defines how the waveform samples will be looped.)

Definition at line 537 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::Manufacturer

Specifies the MIDI Manufacturer's Association (MMA) Manufacturer code for the sampler intended to receive this file's waveform. If no particular manufacturer is to be specified, a value of 0 should be used.

Definition at line 528 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::MIDIUnityNote

Specifies the musical note at which the sample will be played at it's original sample rate.

Definition at line 531 of file gig.h.

Referenced by Sample(), and UpdateChunks().

RIFF::Chunk* gig::Sample::pCk3gix
protected

Definition at line 578 of file gig.h.

Referenced by Sample(), and UpdateChunks().

RIFF::Chunk* DLS::Sample::pCkFormat
protectedinherited

Definition at line 396 of file DLS.h.

Referenced by DLS::Sample::Sample(), and DLS::Sample::UpdateChunks().

RIFF::Chunk* gig::Sample::pCkSmpl
protected

Definition at line 579 of file gig.h.

Referenced by Sample(), and UpdateChunks().

dlsid_t* DLS::Resource::pDLSID
inherited

Points to a dlsid_t structure if the file provided a DLS ID else is NULL.

Definition at line 328 of file DLS.h.

Referenced by DLS::Resource::GenerateDLSID(), DLS::Resource::Resource(), DLS::Resource::UpdateChunks(), and DLS::Resource::~Resource().

Group* gig::Sample::pGroup
protected

pointer to the Group this sample belongs to (always not-NULL)

Definition at line 569 of file gig.h.

Referenced by gig::Group::AddSample(), GetGroup(), Sample(), and UpdateChunks().

Info* DLS::Resource::pInfo
inherited

Points (in any case) to an Info object, providing additional, optional infos and comments.

Definition at line 327 of file DLS.h.

Referenced by gig::File::AddInstrument(), DLS::Resource::CopyAssign(), gig::File::File(), gig::Instrument::Instrument(), DLS::Resource::Resource(), Sample(), DLS::Resource::UpdateChunks(), and DLS::Resource::~Resource().

Resource* DLS::Resource::pParent
protectedinherited
RIFF::List* DLS::Resource::pResourceList
protectedinherited

Definition at line 336 of file DLS.h.

Referenced by DLS::Resource::Resource(), and DLS::Resource::UpdateChunks().

uint32_t gig::Sample::Product

Specifies the MIDI model ID defined by the manufacturer corresponding to the Manufacturer field. If no particular manufacturer's product is to be specified, a value of 0 should be used.

Definition at line 529 of file gig.h.

Referenced by Sample(), and UpdateChunks().

RIFF::List* DLS::Sample::pWaveList
protectedinherited
buffer_t gig::Sample::RAMCache
protected

Buffers samples (already uncompressed) in RAM.

Definition at line 576 of file gig.h.

Referenced by GetCache(), LoadSampleDataWithNullSamplesExtension(), ReleaseSampleData(), Sample(), and ~Sample().

uint32_t gig::Sample::SamplePeriod

Specifies the duration of time that passes during the playback of one sample in nanoseconds (normally equal to 1 / Samples Per Second, where Samples Per Second is the value found in the format chunk), don't bother to update this attribute, it won't be saved.

Definition at line 530 of file gig.h.

Referenced by Sample(), and UpdateChunks().

unsigned long gig::Sample::SamplePos
protected

For compressed samples only: stores the current position (in sample points).

Definition at line 572 of file gig.h.

Referenced by GetPos(), Read(), Sample(), and SetPos().

unsigned long gig::Sample::SamplesInLastFrame
protected

For compressed samples only: length of the last sample frame.

Definition at line 573 of file gig.h.

Referenced by Read().

unsigned long gig::Sample::SamplesPerFrame
protected

For compressed samples only: number of samples in a full sample frame.

Definition at line 575 of file gig.h.

Referenced by Read(), and WorstCaseMaxSamples().

uint32_t DLS::Sample::SamplesPerSecond
inherited

Sampling rate at which each channel should be played (defaults to 44100 if Sample was created with Instrument::AddSample() previously).

Definition at line 378 of file DLS.h.

Referenced by DLS::Sample::Sample(), Sample(), DLS::Sample::UpdateChunks(), and UpdateChunks().

unsigned long DLS::Sample::SamplesTotal
inherited

Reflects total number of sample points (only if known sample data format is used, 0 otherwise), do not bother to change this value, it will not be saved.

Definition at line 382 of file DLS.h.

Referenced by LoadSampleData(), LoadSampleDataWithNullSamplesExtension(), Read(), DLS::Sample::Sample(), and SetPos().

smpte_format_t gig::Sample::SMPTEFormat

Specifies the Society of Motion Pictures and Television E time format used in the following SMPTEOffset field. If a value of 0 is set, SMPTEOffset should also be set to 0.

Definition at line 533 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::SMPTEOffset

The SMPTE Offset value specifies the time offset to be used for the synchronization / calibration to the first sample in the waveform. This value uses a format of 0xhhmmssff where hh is a signed value that specifies the number of hours (-23 to 23), mm is an unsigned value that specifies the number of minutes (0 to 59), ss is an unsigned value that specifies the number of seconds (0 to 59) and ff is an unsigned value that specifies the number of frames (0 to -1).

Definition at line 534 of file gig.h.

Referenced by Sample(), and UpdateChunks().

uint32_t gig::Sample::TruncatedBits

For 24-bit compressed samples only: number of bits truncated during compression (0, 4 or 6)

Definition at line 544 of file gig.h.

Referenced by Read(), and Sample().

unsigned long DLS::Sample::ulWavePoolOffset
protectedinherited
unsigned long gig::Sample::WorstCaseFrameSize
protected

For compressed samples only: size (in bytes) of the largest possible sample frame.

Definition at line 574 of file gig.h.

Referenced by GuessSize(), and WorstCaseMaxSamples().


The documentation for this class was generated from the following files: