VisUtils Namespace Reference

Functions
bool	IsValidFFTSize (t_size p_size)
t_size	MatchFFTSize (t_size samples)
void	PrepareFFTChunk (audio_chunk const &source, audio_chunk &out, double centerOffset)

Function Documentation

bool VisUtils::IsValidFFTSize

(

t_size

p_size

)

Definition at line 24 of file VisUtils.cpp.

                                       {
        return p_size >= 2 && (p_size & (p_size - 1)) == 0;
    }

t_size VisUtils::MatchFFTSize

(

t_size

samples

)

Definition at line 28 of file VisUtils.cpp.

                                        {
        if (samples <= 2) return 2;
        t_size mask = 1;
        while(!IsValidFFTSize(samples)) {
            samples &= ~mask; mask <<= 1;
        }
        return samples;
    }

void VisUtils::PrepareFFTChunk	(	audio_chunk const &	source,
		audio_chunk &	out,
		double	centerOffset
	)

Turns an arbitrary audio_chunk into a valid chunk to run FFT on, with proper sample count etc.

Parameters

centerOffset

Time offset (in seconds) inside the source chunk to center the output on, in case the FFT window is smaller than input data.

Definition at line 4 of file VisUtils.cpp.

                                                                                             {
        const t_uint32 channels = source.get_channel_count();
        const t_uint32 sampleRate = source.get_sample_rate();
        FB2K_DYNAMIC_ASSERT( sampleRate > 0 );
        out.set_channels(channels, source.get_channel_config());
        out.set_sample_rate(sampleRate);
        const t_size inSize = source.get_sample_count();
        const t_size fftSize = MatchFFTSize(inSize);
        out.set_sample_count(fftSize);
        out.set_data_size(fftSize * channels);
        if (fftSize >= inSize) { //rare case with *REALLY* small input
            pfc::memcpy_t( out.get_data(), source.get_data(), inSize * channels );
            pfc::memset_null_t( out.get_data() + inSize * channels, (fftSize - inSize) * channels );
        } else { //inSize > fftSize, we're using a subset of source chunk for the job, pick a subset around centerOffset.
            const double baseOffset = pfc::max_t<double>(0, centerOffset - 0.5 * (double)fftSize / (double)sampleRate);
            const t_size baseSample = pfc::min_t<t_size>( (t_size) audio_math::time_to_samples(baseOffset, sampleRate), inSize - fftSize);
            pfc::memcpy_t( out.get_data(), source.get_data() + baseSample * channels, fftSize * channels);
        }
    }