sp500/log_mel.h File Reference

Mobilint REGULUS v4.1 SP500 DSP API

• sp500

Data Structures | Macros | Functions

log_mel.h File Reference

Whisper-style log-mel spectrogram on the SP500 DSP. More...

#include <sp500/dsp.h>

Go to the source code of this file.

Data Structures
struct	sp500_log_mel_params
	Front-end parameters. More...
struct	sp500_log_mel_result
	Values reported by a log-mel call. More...

Macros
#define	SP500_LOG_MEL_MAX_MELS   256
	Largest supported mel bin count.
#define	SP500_LOG_MEL_MAX_FFT   2048
	Largest supported FFT size.

Functions
void	sp500_log_mel_params_init (sp500_log_mel_params *params)
	Fill params with the Whisper defaults documented above.
int32_t	sp500_log_mel_num_frames (int32_t n_samples, int32_t hop_length)
	Frames produced for n_samples: n_samples / hop_length, or 0 for non-positive arguments. Use it to size the feature buffer.
int32_t	sp500_log_mel_num_freq (int32_t n_fft)
	Frequency bins for n_fft: n_fft / 2 + 1, or 0 if n_fft is not supported. Use it to size a custom mel_filters bank.
sp500_status	sp500_log_mel (const sp500_buffer *audio, const sp500_log_mel_params *params, sp500_buffer *features, sp500_log_mel_result *result)
	Compute the log-mel spectrogram of float32 PCM on the DSP.
sp500_status	sp500_log_mel_cpu (const float *audio, const sp500_log_mel_params *params, float *features, int32_t features_count, sp500_log_mel_result *result)
	Host CPU reference for sp500_log_mel(): same tables, same per-frame maths, no DSP involved.

Detailed Description

Whisper-style log-mel spectrogram on the SP500 DSP.

Turns float32 PCM into the log-mel spectrogram that Whisper and Qwen3-ASR encoders consume. The maths follows the Hugging Face WhisperFeatureExtractor exactly: reflect-padded frames of n_fft samples every hop_length, a periodic Hann window, the power spectrum, a slaney mel filter bank, log10(max(1e-10, mel)), then optionally the Whisper normalisation (clamp at max - 8, then (x + 4) / 4).

Output layout

The result is frame-major: features[t * n_mels + m] for t < n_frames, with n_frames = n_samples / hop_length - the trailing centred frame is dropped, as Hugging Face does. This is the [n_frames, n_mels] layout the MXQ Whisper models take. Hugging Face's own input_features is the transpose.

Example

sp500_dsp_open(0);

sp500_log_mel_params p;

sp500_log_mel_params_init(&p);

p.n_samples = 16000 * 30; // 30 s at 16 kHz

const int32_t frames = sp500_log_mel_num_frames(p.n_samples, p.hop_length);

sp500_buffer audio = sp500_buffer_wrap(pcm, p.n_samples * sizeof(float));

sp500_buffer feats = sp500_buffer_alloc(frames * p.n_mels * sizeof(float),

SP500_ALIGN_DSP);

sp500_log_mel_result r;

if(sp500_log_mel(&audio, &p, &feats, &r) == SP500_OK)

{

// feats.data holds r.n_frames * p.n_mels floats, frame-major.

}

sp500_buffer_release(&feats);

sp500_buffer_alloc

sp500_buffer sp500_buffer_alloc(uint32_t size, uint32_t alignment)

Allocate a zero-copy buffer in shared memory.

Definition dsp.h:286

SP500_OK

@ SP500_OK

Definition dsp.h:87

SP500_ALIGN_DSP

#define SP500_ALIGN_DSP

Smallest alignment the DSP accepts.

Definition dsp.h:192

sp500_dsp_open

sp500_status sp500_dsp_open(uint32_t timeout_ms)

Attach to the DSP. Must succeed before any other DSP call.

sp500_buffer_release

void sp500_buffer_release(sp500_buffer *buffer)

Release a buffer and reset it to unused.

Definition dsp.h:316

sp500_buffer_wrap

sp500_buffer sp500_buffer_wrap(void *data, uint32_t size)

Wrap memory you already own. The call stages one copy in each direction; nothing is taken over and no...

Definition dsp.h:302

sp500_log_mel_params_init

void sp500_log_mel_params_init(sp500_log_mel_params *params)

Fill params with the Whisper defaults documented above.

sp500_log_mel

sp500_status sp500_log_mel(const sp500_buffer *audio, const sp500_log_mel_params *params, sp500_buffer *features, sp500_log_mel_result *result)

Compute the log-mel spectrogram of float32 PCM on the DSP.

sp500_log_mel_num_frames

int32_t sp500_log_mel_num_frames(int32_t n_samples, int32_t hop_length)

Frames produced for n_samples: n_samples / hop_length, or 0 for non-positive arguments....

sp500_buffer

A tensor handed to or received from the DSP.

Definition dsp.h:261

sp500_log_mel_params

Front-end parameters.

Definition log_mel.h:67

sp500_log_mel_params::n_mels

int32_t n_mels

Definition log_mel.h:73

sp500_log_mel_params::n_samples

int32_t n_samples

Definition log_mel.h:68

sp500_log_mel_params::hop_length

int32_t hop_length

Definition log_mel.h:72

sp500_log_mel_result

Values reported by a log-mel call.

Definition log_mel.h:106

Note

A 30 s clip is a few thousand frames of DSP work. If a call returns SP500_ERR_TIMEOUT, raise the deadline with sp500_set_dsp_timeout_ms() rather than retrying - see the warning on that function.

Definition in file log_mel.h.

Macro Definition Documentation

◆ SP500_LOG_MEL_MAX_MELS

#define SP500_LOG_MEL_MAX_MELS   256

Largest supported mel bin count.

Definition at line 54 of file log_mel.h.

◆ SP500_LOG_MEL_MAX_FFT

#define SP500_LOG_MEL_MAX_FFT   2048

Largest supported FFT size.

Definition at line 57 of file log_mel.h.

Function Documentation

◆ sp500_log_mel()

sp500_status sp500_log_mel	(	const sp500_buffer *	audio,
		const sp500_log_mel_params *	params,
		sp500_buffer *	features,
		sp500_log_mel_result *	result )

Compute the log-mel spectrogram of float32 PCM on the DSP.

Parameters

audio	Input samples; at least n_samples * sizeof(float) bytes. 16 kHz mono for Whisper.
params	Front-end parameters; see sp500_log_mel_params.
features	Receives n_frames * n_mels float32, frame-major; at least n_frames * n_mels * sizeof(float) bytes.
result	Receives the frame count and peak. May be NULL.

Returns

SP500_OK on success. SP500_ERR_UNSUPPORTED for an n_fft the DSP cannot transform, or for a mel bank it cannot run - one whose weights are all zero because the frequency range covers no FFT bin, or one too dense to describe. SP500_ERR_INVALID_ARG for any other bad parameter, SP500_ERR_BUFFER_TOO_SMALL for an undersized buffer, SP500_ERR_NO_MEMORY if shared memory is exhausted, SP500_ERR_NOT_READY, SP500_ERR_TIMEOUT or SP500_ERR_DSP.

Note

n_samples < hop_length yields zero frames. That is reported as SP500_OK with result->n_frames == 0 and nothing written.

◆ sp500_log_mel_cpu()

sp500_status sp500_log_mel_cpu	(	const float *	audio,
		const sp500_log_mel_params *	params,
		float *	features,
		int32_t	features_count,
		sp500_log_mel_result *	result )

Host CPU reference for sp500_log_mel(): same tables, same per-frame maths, no DSP involved.

Useful during bring-up to check the DSP result, and as a fallback when the DSP is busy. It does not require sp500_dsp_open().

Parameters

audio	params->n_samples float32 samples.
params	The same parameters the DSP call would take.
features	Receives n_frames * n_mels float32, frame-major.
features_count	Number of floats features can hold.
result	Receives the frame count and peak. May be NULL; host_seconds is the CPU time of this call.

Returns

SP500_OK, SP500_ERR_INVALID_ARG, SP500_ERR_UNSUPPORTED (the same n_fft and mel-bank cases as the DSP call), SP500_ERR_BUFFER_TOO_SMALL or SP500_ERR_NO_MEMORY. It returns the same status as sp500_log_mel() for the same parameters.

---

Generated by  1.14.0