RPPT Tensor Operations - Audio Augmentations.

RPPT Tensor Operations - Audio Augmentations.#

RPP: RPPT Tensor Operations - Audio Augmentations.
RPPT Tensor Operations - Audio Augmentations.

RPPT Tensor Operations - Audio Augmentations. More...

Functions

RppStatus rppt_non_silent_region_detection_host (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, Rpp32s *srcLengthTensor, Rpp32s *detectedIndexTensor, Rpp32s *detectionLengthTensor, Rpp32f cutOffDB, Rpp32s windowLength, Rpp32f referencePower, Rpp32s resetInterval, rppHandle_t rppHandle)
 Non Silent Region Detection augmentation on HOST backend. More...
 
RppStatus rppt_non_silent_region_detection_gpu (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, Rpp32s *srcLengthTensor, Rpp32s *detectedIndexTensor, Rpp32s *detectionLengthTensor, Rpp32f cutOffDB, Rpp32s windowLength, Rpp32f referencePower, Rpp32s resetInterval, rppHandle_t rppHandle)
 Non Silent Region Detection augmentation on HIP backend. More...
 
RppStatus rppt_to_decibels_host (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, RpptImagePatchPtr srcDims, Rpp32f cutOffDB, Rpp32f multiplier, Rpp32f referenceMagnitude, rppHandle_t rppHandle)
 To Decibels augmentation on HOST backend. More...
 
RppStatus rppt_to_decibels_gpu (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, RpptImagePatchPtr srcDims, Rpp32f cutOffDB, Rpp32f multiplier, Rpp32f referenceMagnitude, rppHandle_t rppHandle)
 To Decibels augmentation on HIP backend. More...
 
RppStatus rppt_pre_emphasis_filter_host (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32s *srcLengthTensor, Rpp32f *coeffTensor, RpptAudioBorderType borderType, rppHandle_t rppHandle)
 Pre Emphasis Filter augmentation on HOST backend. More...
 
RppStatus rppt_pre_emphasis_filter_gpu (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32s *srcLengthTensor, Rpp32f *coeffTensor, RpptAudioBorderType borderType, rppHandle_t rppHandle)
 Pre Emphasis Filter augmentation on HIP backend. More...
 
RppStatus rppt_down_mixing_host (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32s *srcDimsTensor, bool normalizeWeights, rppHandle_t rppHandle)
 Down Mixing augmentation on HOST backend. More...
 
RppStatus rppt_down_mixing_gpu (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32s *srcDimsTensor, bool normalizeWeights, rppHandle_t rppHandle)
 Down Mixing augmentation on HIP backend. More...
 
RppStatus rppt_spectrogram_host (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32s *srcLengthTensor, bool centerWindows, bool reflectPadding, Rpp32f *windowFunction, Rpp32s nfft, Rpp32s power, Rpp32s windowLength, Rpp32s windowStep, rppHandle_t rppHandle)
 Produces a spectrogram from a 1D audio buffer on HOST backend. More...
 
RppStatus rppt_mel_filter_bank_host (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32s *srcDims, Rpp32f maxFreq, Rpp32f minFreq, RpptMelScaleFormula melFormula, Rpp32s numFilter, Rpp32f sampleRate, bool normalize, rppHandle_t rppHandle)
 Mel filter bank augmentation HOST backend. More...
 
RppStatus rppt_resample_host (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32f *inRateTensor, Rpp32f *outRateTensor, Rpp32s *srcDimsTensor, RpptResamplingWindow &window, rppHandle_t rppHandle)
 Resample augmentation on HOST backend. More...
 
RppStatus rppt_resample_gpu (RppPtr_t srcPtr, RpptDescPtr srcDescPtr, RppPtr_t dstPtr, RpptDescPtr dstDescPtr, Rpp32f *inRateTensor, Rpp32f *outRateTensor, Rpp32s *srcDimsTensor, RpptResamplingWindow &window, rppHandle_t rppHandle)
 Resample augmentation on HIP backend. More...
 

Detailed Description

RPPT Tensor Operations - Audio Augmentations.

Function Documentation

◆ rppt_down_mixing_gpu()

RppStatus rppt_down_mixing_gpu ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32s srcDimsTensor,
bool  normalizeWeights,
rppHandle_t  rppHandle 
)

Down Mixing augmentation on HIP backend.

Down Mixing augmentation for audio data

Parameters
[in]srcPtrsource tensor in HIP memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel audio tensor), offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HIP memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 2, offsetInBytes >= 0, dataType = F32)
[in]srcDimsTensorsource audio buffer length and number of channels (1D tensor in HIP/Pinned memory, of size batchSize * 2)
[in]normalizeWeightsbool flag to specify if normalization of weights is needed
[in]rppHandleRPP HIP handle created with rppCreateWithStreamAndBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_down_mixing_host()

RppStatus rppt_down_mixing_host ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32s srcDimsTensor,
bool  normalizeWeights,
rppHandle_t  rppHandle 
)

Down Mixing augmentation on HOST backend.

Down Mixing augmentation for audio data

Parameters
[in]srcPtrsource tensor in HOST memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel audio tensor), offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HOST memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 2, offsetInBytes >= 0, dataType = F32)
[in]srcDimsTensorsource audio buffer length and number of channels (1D tensor in HOST memory, of size batchSize * 2)
[in]normalizeWeightsbool flag to specify if normalization of weights is needed
[in]rppHandleRPP HOST handle created with rppCreateWithBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_mel_filter_bank_host()

RppStatus rppt_mel_filter_bank_host ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32s srcDims,
Rpp32f  maxFreq,
Rpp32f  minFreq,
RpptMelScaleFormula  melFormula,
Rpp32s  numFilter,
Rpp32f  sampleRate,
bool  normalize,
rppHandle_t  rppHandle 
)

Mel filter bank augmentation HOST backend.

Mel filter bank augmentation for audio data

Parameters
[in]srcPtrsource tensor in HOST memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 3, offsetInBytes >= 0, dataType = F32, layout - NFT)
[out]dstPtrdestination tensor in HOST memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 3, offsetInBytes >= 0, dataType = F32, layout - NFT)
[in]srcDimsTensorsource audio buffer length and number of channels (1D tensor in HOST memory, of size batchSize * 2)
[in]maxFreqmaximum frequency if not provided maxFreq = sampleRate / 2
[in]minFreqminimum frequency
[in]melFormulaformula used to convert frequencies from hertz to mel and from mel to hertz (SLANEY / HTK)
[in]numFilternumber of mel filters
[in]sampleRatesampling rate of the audio
[in]normalizeboolean variable that determine whether to normalize weights / not
[in]rppHandleRPP HOST handle created with rppCreateWithBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_non_silent_region_detection_gpu()

RppStatus rppt_non_silent_region_detection_gpu ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
Rpp32s srcLengthTensor,
Rpp32s detectedIndexTensor,
Rpp32s detectionLengthTensor,
Rpp32f  cutOffDB,
Rpp32s  windowLength,
Rpp32f  referencePower,
Rpp32s  resetInterval,
rppHandle_t  rppHandle 
)

Non Silent Region Detection augmentation on HIP backend.

Non Silent Region Detection augmentation for 1D audio buffer
Finds the starting index and length of non silent region in the audio buffer by comparing the calculated short-term power with cutoff value passed

Parameters
[in]srcPtrsource tensor in HIP memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2, offsetInBytes >= 0, dataType = F32)
[in]srcLengthTensorsource audio buffer length (1D tensor in Pinned/HIP memory, of size batchSize)
[out]detectedIndexTensorbeginning index of non silent region (1D tensor in Pinned/HIP memory, of size batchSize)
[out]detectionLengthTensorlength of non silent region (1D tensor in Pinned/HIP memory, of size batchSize)
[in]cutOffDBcutOff in dB below which the signal is considered silent
[in]windowLengthwindow length used for computing short-term power of the signal
[in]referencePowerreference power that is used to convert the signal to dB
[in]resetIntervalnumber of samples after which the moving mean average is recalculated to avoid precision loss
[in]rppHandleRPP HIP handle created with rppCreateWithStreamAndBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_non_silent_region_detection_host()

RppStatus rppt_non_silent_region_detection_host ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
Rpp32s srcLengthTensor,
Rpp32s detectedIndexTensor,
Rpp32s detectionLengthTensor,
Rpp32f  cutOffDB,
Rpp32s  windowLength,
Rpp32f  referencePower,
Rpp32s  resetInterval,
rppHandle_t  rppHandle 
)

Non Silent Region Detection augmentation on HOST backend.

Non Silent Region Detection augmentation for 1D audio buffer
Finds the starting index and length of non silent region in the audio buffer by comparing the calculated short-term power with cutoff value passed

Parameters
[in]srcPtrsource tensor in HOST memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2, offsetInBytes >= 0, dataType = F32)
[in]srcLengthTensorsource audio buffer length (1D tensor in HOST memory, of size batchSize)
[out]detectedIndexTensorbeginning index of non silent region (1D tensor in HOST memory, of size batchSize)
[out]detectionLengthTensorlength of non silent region (1D tensor in HOST memory, of size batchSize)
[in]cutOffDBcutOff in dB below which the signal is considered silent
[in]windowLengthwindow length used for computing short-term power of the signal
[in]referencePowerreference power that is used to convert the signal to dB
[in]resetIntervalnumber of samples after which the moving mean average is recalculated to avoid precision loss
[in]rppHandleRPP HOST handle created with rppCreateWithBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_pre_emphasis_filter_gpu()

RppStatus rppt_pre_emphasis_filter_gpu ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32s srcLengthTensor,
Rpp32f coeffTensor,
RpptAudioBorderType  borderType,
rppHandle_t  rppHandle 
)

Pre Emphasis Filter augmentation on HIP backend.

Pre Emphasis Filter augmentation for audio data

Parameters
[in]srcPtrsource tensor in HIP memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2, offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HIP memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 2, offsetInBytes >= 0, dataType = F32)
[in]srcLengthTensorsource audio buffer length (1D tensor in HIP memory, of size batchSize)
[in]coeffTensorpreemphasis coefficient (1D tensor in Pinned / HIP memory, of size batchSize)
[in]borderTypeborder value policy
[in]rppHandleRPP HIP handle created with rppCreateWithStreamAndBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_pre_emphasis_filter_host()

RppStatus rppt_pre_emphasis_filter_host ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32s srcLengthTensor,
Rpp32f coeffTensor,
RpptAudioBorderType  borderType,
rppHandle_t  rppHandle 
)

Pre Emphasis Filter augmentation on HOST backend.

Pre Emphasis Filter augmentation for audio data

Parameters
[in]srcPtrsource tensor in HOST memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 3, offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HOST memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 3, offsetInBytes >= 0, dataType = F32)
[in]srcLengthTensorsource audio buffer length (1D tensor in HOST memory, of size batchSize)
[in]coeffTensorpreemphasis coefficient (1D tensor in HOST memory, of size batchSize)
[in]borderTypeborder value policy
[in]rppHandleRPP HOST handle created with rppCreateWithBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_resample_gpu()

RppStatus rppt_resample_gpu ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32f inRateTensor,
Rpp32f outRateTensor,
Rpp32s srcDimsTensor,
RpptResamplingWindow window,
rppHandle_t  rppHandle 
)

Resample augmentation on HIP backend.

Resample augmentation for audio data

Parameters
[in]srcPtrsource tensor in HIP memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel audio tensor), offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HIP memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel audio tensor), offsetInBytes >= 0, dataType = F32)
[in]inRateInput sampling rate (1D tensor in Pinned memory, of size batchSize)
[in]outRateOutput sampling rate (1D tensor in Pinned memory, of size batchSize)
[in]srcDimsTensorsource audio buffer length and number of channels (1D tensor in Pinned memory, of size batchSize * 2)
[in]windowResampling window (struct of type RpptRpptResamplingWindow in HIP/Pinned memory)
[in]rppHandleRPP HIP handle created with rppCreateWithStreamAndBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_resample_host()

RppStatus rppt_resample_host ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32f inRateTensor,
Rpp32f outRateTensor,
Rpp32s srcDimsTensor,
RpptResamplingWindow window,
rppHandle_t  rppHandle 
)

Resample augmentation on HOST backend.

Resample augmentation for audio data

Parameters
[in]srcPtrsource tensor in HOST memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel audio tensor), offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HOST memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel audio tensor), offsetInBytes >= 0, dataType = F32)
[in]inRateInput sampling rate (1D tensor in HOST memory, of size batchSize)
[in]outRateOutput sampling rate (1D tensor in HOST memory, of size batchSize)
[in]srcDimsTensorsource audio buffer length and number of channels (1D tensor in HOST memory, of size batchSize * 2)
[in]windowResampling window (struct of type RpptRpptResamplingWindow)
[in]rppHandleRPP HOST handle created with rppCreateWithBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_spectrogram_host()

RppStatus rppt_spectrogram_host ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
Rpp32s srcLengthTensor,
bool  centerWindows,
bool  reflectPadding,
Rpp32f windowFunction,
Rpp32s  nfft,
Rpp32s  power,
Rpp32s  windowLength,
Rpp32s  windowStep,
rppHandle_t  rppHandle 
)

Produces a spectrogram from a 1D audio buffer on HOST backend.

Spectrogram for 1D audio buffer

Parameters
[in]srcPtrsource tensor in HOST memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 3, offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HOST memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 3, offsetInBytes >= 0, dataType = F32, layout - NFT / NTF)
[in]srcLengthTensorsource audio buffer length (1D tensor in HOST memory, of size batchSize)
[in]centerWindowsindicates whether extracted windows should be padded so that the window function is centered at multiples of window_step
[in]reflectPaddingindicates the padding policy when sampling outside the bounds of the signal
[in]windowFunctionsamples of the window function that will be multiplied to each extracted window when calculating the Short Time Fourier Transform (STFT)
[in]nfftsize of the FFT
[in]powerexponent of the magnitude of the spectrum
[in]windowLengthwindow size in number of samples
[in]windowStepstep between the STFT windows in number of samples
[in]rppHandleRPP HOST handle created with rppCreateWithBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_to_decibels_gpu()

RppStatus rppt_to_decibels_gpu ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
RpptImagePatchPtr  srcDims,
Rpp32f  cutOffDB,
Rpp32f  multiplier,
Rpp32f  referenceMagnitude,
rppHandle_t  rppHandle 
)

To Decibels augmentation on HIP backend.

To Decibels augmentation for 1D/2D audio buffer converts magnitude values to decibel values

Parameters
[in]srcPtrsource tensor in HIP memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel/2D audio tensor with 1 channel), offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HIP memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel/2D audio tensor with 1 channel), offsetInBytes >= 0, dataType = F32)
[in]srcDimssource tensor sizes for each element in batch (2D tensor in Pinned/HIP memory, of size batchSize * 2)
[in]cutOffDBminimum or cut-off ratio in dB
[in]multiplierfactor by which the logarithm is multiplied
[in]referenceMagnitudeReference magnitude if not provided maximum value of input used as reference
[in]rppHandleRPP HIP handle created with rppCreateWithStreamAndBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.

◆ rppt_to_decibels_host()

RppStatus rppt_to_decibels_host ( RppPtr_t  srcPtr,
RpptDescPtr  srcDescPtr,
RppPtr_t  dstPtr,
RpptDescPtr  dstDescPtr,
RpptImagePatchPtr  srcDims,
Rpp32f  cutOffDB,
Rpp32f  multiplier,
Rpp32f  referenceMagnitude,
rppHandle_t  rppHandle 
)

To Decibels augmentation on HOST backend.

To Decibels augmentation for 1D/2D audio buffer converts magnitude values to decibel values

Parameters
[in]srcPtrsource tensor in HOST memory
[in]srcDescPtrsource tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel/2D audio tensor with 1 channel), offsetInBytes >= 0, dataType = F32)
[out]dstPtrdestination tensor in HOST memory
[in]dstDescPtrdestination tensor descriptor (Restrictions - numDims = 2 or 3 (for single-channel or multi-channel/2D audio tensor with 1 channel), offsetInBytes >= 0, dataType = F32)
[in]srcDimssource tensor sizes for each element in batch (2D tensor in HOST memory, of size batchSize * 2)
[in]cutOffDBminimum or cut-off ratio in dB
[in]multiplierfactor by which the logarithm is multiplied
[in]referenceMagnitudeReference magnitude if not provided maximum value of input used as reference
[in]rppHandleRPP HOST handle created with rppCreateWithBatchSize()
Returns
A RppStatus enumeration.
Return values
RPP_SUCCESSSuccessful completion.
RPP_ERROR*Unsuccessful completion.