コード例 #1
0
// this is the 32-bit perform method for Max 5 and earlier
t_int *specFlatness_perform(t_int *w)
{
    int i, n;
    double timef;
	// DO NOT CALL post IN HERE, but you can call defer_low (not defer)
	
	// args are in a vector, sized as specified in specFlatness_dsp method
	// w[0] contains &specFlatness_perform, so we start at w[1]
	t_specFlatness *x = (t_specFlatness *)(w[1]);
	//t_float *inL = (t_float *)(w[2]);
	//t_float *outL = (t_float *)(w[3]);
    t_sample *in = (t_float *)(w[2]);
	n = (int)w[3];
	
	// shift signal buffer contents back.
	for(i=0; i<x->window; i++)
		x->signal_R[i] = x->signal_R[i+n];
    
	// write new block to end of signal buffer.
	for(i=0; i<n; i++)
		x->signal_R[(int)x->window+i] = in[i];
    
	//x->lastDspTime = clock_getlogicaltime();
    clock_getftime(&timef);
    x->lastDspTime =  timef;
		
	// you have to return the NEXT pointer in the array OR MAX WILL CRASH
	return w + 4;
}
コード例 #2
0
// this is 64-bit perform method for Max 6
void specFlatness_perform64(t_specFlatness *x, t_object *dsp64, double **ins, long numins, double **outs, long numouts, long sampleframes, long flags, void *userparam)
{
    int i, n;
    double timef;
	//t_double *inL = ins[0];		// we get audio for each inlet of the object from the **ins argument
    t_sample *in = ins[0];
	//t_double *outL = outs[0];	// we get audio for each outlet of the object from the **outs argument
	n = sampleframes;
    
    // shift signal buffer contents back.
	for(i=0; i<x->window; i++)
		x->signal_R[i] = x->signal_R[i+n];
    
	// write new block to end of signal buffer.
	for(i=0; i<n; i++)
		x->signal_R[(int)x->window+i] = in[i];
    
	//x->lastDspTime = clock_getlogicaltime();
    clock_getftime(&timef);
    x->lastDspTime =  timef;

	
	// this perform method simply copies the input to the output, offsetting the value
	//while (n--)
	//	*outL++ = *inL++ + x->offset;
}
コード例 #3
0
// this function is called when the DAC is enabled, and "registers" a function for the signal chain in Max 5 and earlier.
// In this case we register the 32-bit, "specFlatness_perform" method.
void specFlatness_dsp(t_specFlatness *x, t_signal **sp, short *count)
{
    int i;
    double timef;
	post("my sample rate is: %f", sp[0]->s_sr);
	
	// dsp_add
	// 1: (t_perfroutine p) perform method
	// 2: (long argc) number of args to your perform method
	// 3...: argc additional arguments, all must be sizeof(pointer) or long
	// these can be whatever, so you might want to include your object pointer in there
	// so that you have access to the info, if you need it.
	dsp_add(specFlatness_perform, 4, x, sp[0]->s_vec, sp[1]->s_vec, sp[0]->s_n);
    
    
    // compare n to stored n and recalculate filterbank if different
	if(sp[0]->s_sr != (x->sr*x->overlap) || sp[0]->s_n != x->n)
	{
		x->signal_R = (t_sample *)t_resizebytes_(x->signal_R, (x->window+x->n) * sizeof(t_sample), (x->window+sp[0]->s_n) * sizeof(t_sample));
        
		x->sr = sp[0]->s_sr/x->overlap;
		x->n = sp[0]->s_n;
		//x->lastDspTime = clock_getlogicaltime();
        clock_getftime(&timef);
        x->lastDspTime =  timef;
        
		// init signal buffer
		for(i=0; i<(x->window+x->n); i++)
			x->signal_R[i] = 0.0;
                
    	post("specFlatness~: window size: %i. overlap: %i. sampling rate: %i, block size: %i", (int)x->window, x->overlap, (int)x->sr, (int)x->n);
	};
}
コード例 #4
0
// this is the Max 6 version of the dsp method -- it registers a function for the signal chain in Max 6,
// which operates on 64-bit audio signals.
void barkSpec_dsp64(t_barkSpec *x, t_object *dsp64, short *count, double samplerate, long maxvectorsize, long flags)
{
    int i, oldNumFilters;
    double timef;
	post("my sample rate is: %f", samplerate);
	
	// instead of calling dsp_add(), we send the "dsp_add64" message to the object representing the dsp chain
	// the dsp_add64 arguments are:
	// 1: the dsp64 object passed-in by the calling function
	// 2: a pointer to your object
	// 3: a pointer to your 64-bit perform method
	// 4: flags to alter how the signal chain handles your object -- just pass 0
	// 5: a generic pointer that you can use to pass any additional data to your perform method
	
	object_method(dsp64, gensym("dsp_add64"), x, barkSpec_perform64, 0, NULL);
    
    
    // compare n to stored n and recalculate filterbank if different
	if(samplerate != (x->sr*x->overlap) || maxvectorsize != x->n)
	{
		x->signal_R = (t_sample *)t_resizebytes_(x->signal_R, (x->window+x->n) * sizeof(t_sample), (x->window+maxvectorsize) * sizeof(t_sample));
        
		x->sr = samplerate/x->overlap;
		x->n = maxvectorsize;
        
		//x->lastDspTime = clock_getlogicaltime();
        clock_getftime(&timef);
        x->lastDspTime =  timef;
        
		// init signal buffer
		for(i=0; i<(x->window+x->n); i++)
			x->signal_R[i] = 0.0;
        
    	oldNumFilters = x->numFilters;
		x->numFilters = tIDLib_getBarkFilterSpacing(&x->x_filterFreqs, x->sizeFilterFreqs, x->barkSpacing, x->sr);
        
		x->sizeFilterFreqs = x->numFilters+2;
        
		tIDLib_createFilterbank(x->x_filterFreqs, &x->x_filterbank, oldNumFilters, x->numFilters, x->window, x->sr);
        
		post("barkSpec~: window size: %i. sampling rate: %i, block size: %i", (int)x->window, (int)x->sr, (int)x->n);
	};
}
/*! \param tm absolute time (in seconds)
    \param data user data
    \param dopast if set events with times lying in the past will be triggered immediately, if not set they are ignored
    \return true on success
*/
bool flext::Timer::At(double tm,void *data,bool dopast)
{
    userdata = data;
    period = 0;
#if FLEXT_SYS == FLEXT_SYS_PD 
    const double systm = clock_gettimesince(0);
    double df = tm*1000.-systm;
    if(dopast && df < 0) df = 0;
    if(df >= 0)
        clock_delay(clk,df);
#elif FLEXT_SYS == FLEXT_SYS_MAX
    const double ms = tm*1000.;
    double cur;
    clock_getftime(&cur);
    if(cur <= ms)
        clock_fdelay(clk,ms-cur);
    else if(dopast) // trigger timer is past
        clock_fdelay(clk,0);
#else
    #error Not implemented
#endif
    return true;
}
コード例 #6
0
static void specFlatness_tilde_bang(t_specFlatness *x)
{
    int i, j, window, windowHalf, bangSample;
    t_sample *signal_R, *signal_I;
    t_float dividend, divisor, windowHalfPlusOneRecip, flatness, *windowFuncPtr;
	double currentTime,timef =0.0;
    
    window = x->window;
    windowHalf = window*0.5;
    windowHalfPlusOneRecip = 1.0/(t_float)(windowHalf+1);
	
	// create local memory
   	signal_R = (t_sample *)t_getbytes(window*sizeof(t_sample));
	signal_I = (t_sample *)t_getbytes((windowHalf+1)*sizeof(t_sample));
    
	//currentTime = clock_gettimesince(x->lastDspTime);
    clock_getftime(&timef);
    currentTime =  timef - x->lastDspTime;
    
	bangSample = (int)(((currentTime/1000.0)*x->sr)+0.5); // round
    
	if (bangSample < 0)
        bangSample = 0;
    else if ( bangSample >= x->n )
        bangSample = x->n - 1;
    
	// construct analysis window using bangSample as the end of the window
	for(i=0, j=bangSample; i<window; i++, j++)
		signal_R[i] = x->signal_R[j];
    
	// set window function
	windowFuncPtr = x->hann; //default case to get rid of compile warning
    
	switch(x->windowFunction)
	{
		case 0:
			break;
		case 1:
			windowFuncPtr = x->blackman;
			break;
		case 2:
			windowFuncPtr = x->cosine;
			break;
		case 3:
			windowFuncPtr = x->hamming;
			break;
		case 4:
			windowFuncPtr = x->hann;
			break;
		default:
			break;
	};
    
	// if windowFunction == 0, skip the windowing (rectangular)
	if(x->windowFunction>0)
		for(i=0; i<window; i++, windowFuncPtr++)
			signal_R[i] *= *windowFuncPtr;
    
	mayer_realfft(window, signal_R);
	tIDLib_realfftUnpack(window, windowHalf, signal_R, signal_I);
	tIDLib_power(windowHalf+1, signal_R, signal_I);
    
	if(!x->powerSpectrum)
		tIDLib_mag(windowHalf+1, signal_R);
    
	dividend=1; // to get the product of all terms for geometric mean
	divisor=0;
	flatness=0;
	
	// geometric mean
	// take the nth roots first so as not to lose data.
	for(i=0; i<=windowHalf; i++)
        x->nthRoots[i] = pow(signal_R[i], windowHalfPlusOneRecip);
    
	// take the product of nth roots
	// what to do with values that are zero?  for now, ignoring them.	
	for(i=0; i<=windowHalf; i++)
		if(x->nthRoots[i] != 0)
			dividend *= x->nthRoots[i];
    
	for(i=0; i<=windowHalf; i++)
		divisor += signal_R[i];
	
	divisor *= windowHalfPlusOneRecip; // arithmetic mean
	
	if(divisor==0)
		divisor = 1;
	
	flatness = dividend/divisor;
	
	outlet_float(x->x_flatness, flatness);
    
	// free local memory
	t_freebytes(signal_R, window*sizeof(t_sample));
	t_freebytes(signal_I, (windowHalf+1)*sizeof(t_sample));
}
コード例 #7
0
void *specFlatness_new(t_symbol *s, long argc, t_atom *argv)
{
	t_specFlatness *x = (t_specFlatness *)object_alloc(specFlatness_class);
    int i, isPow2;
    double timef = 0.0;
	//s=s;
    
	if (x) {
		dsp_setup((t_pxobject *)x, 1);	// MSP inlets: arg is # of inlets and is REQUIRED! 
										// use 0 if you don't need inlets
		//outlet_new(x, "signal"); 		// signal outlet (note "signal" rather than NULL)
		//x->offset = 0.0;
        
        
        x->x_flatness = floatout(x);
        //x->x_flatness = outlet_new(&x->x_obj, &s_float);        
        
        if(argc > 1)
        {
            x->window = atom_getfloat(argv);
            isPow2 = (int)x->window && !( ((int)x->window-1) & (int)x->window );
            
            if(!isPow2)
            {
                error("requested window size is not a power of 2. default value of 1024 used instead");
                x->window = 1024;
            };
                      
        }
        else if(argc > 0)
        {
            x->window = atom_getfloat(argv);
            isPow2 = (int)x->window && !( ((int)x->window-1) & (int)x->window );
            
            if(!isPow2)
            {
                error("requested window size is not a power of 2. default value of 1024 used instead");
                x->window = 1024;
            };
            
            
        }
        else
        {
            x->window = 1024;
        };
        
        x->sr = 44100.0;
        x->n = 64.0;
        x->overlap = 1;
        x->windowFunction = 4; // 4 is hann window
        x->powerSpectrum = 0;
        //x->lastDspTime = clock_getlogicaltime();
        clock_getftime(&timef);
        x->lastDspTime = timef;
        
        x->signal_R = (t_sample *)t_getbytes_((x->window+x->n) * sizeof(t_sample));
        x->nthRoots = (t_float *)t_getbytes((x->window*0.5+1)*sizeof(t_float));
        
        for(i=0; i<(x->window+x->n); i++)
            x->signal_R[i] = 0.0;
        
        x->blackman = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        x->cosine = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        x->hamming = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        x->hann = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        
        // initialize signal windowing functions
        tIDLib_blackmanWindow(x->blackman, x->window);
        tIDLib_cosineWindow(x->cosine, x->window);
        tIDLib_hammingWindow(x->hamming, x->window);
        tIDLib_hannWindow(x->hann, x->window);
        
        post("specFlatness~: window size: %i", (int)x->window);
        
	}
	return (x);
}
コード例 #8
0
static void barkSpec_tilde_bang(t_barkSpec *x)
{
    int i, j, window, windowHalf, bangSample;
	t_atom *listOut;
	t_sample *signal_R, *signal_I;
	t_float *windowFuncPtr;
	double currentTime, timef = 0.0;
    
	window = x->window;
	windowHalf = window*0.5;
    
	// create local memory
	listOut = (t_atom *)t_getbytes_(x->numFilters*sizeof(t_atom));
	signal_R = (t_sample *)t_getbytes_(window*sizeof(t_sample));
	signal_I = (t_sample *)t_getbytes_((windowHalf+1)*sizeof(t_sample));
    
	//currentTime = clock_gettimesince(x->lastDspTime);
    clock_getftime(&timef);
    currentTime =  timef - x->lastDspTime;
    
	bangSample = (int)(((currentTime/1000.0)*x->sr)+0.5); // round
    
	if (bangSample < 0)
		bangSample = 0;
	else if ( bangSample >= x->n )
		bangSample = x->n - 1;
    
	// construct analysis window using bangSample as the end of the window
	for(i=0, j=bangSample; i<window; i++, j++)
		signal_R[i] = x->signal_R[j];
    
	// set window function
	windowFuncPtr = x->hann; //default case to get rid of compile warning
    
	switch(x->windowFunction)
	{
		case 0:
			break;
		case 1:
			windowFuncPtr = x->blackman;
			break;
		case 2:
			windowFuncPtr = x->cosine;
			break;
		case 3:
			windowFuncPtr = x->hamming;
			break;
		case 4:
			windowFuncPtr = x->hann;
			break;
		default:
			break;
	};
    
	// if windowFunction == 0, skip the windowing (rectangular)
	if(x->windowFunction>0)
		for(i=0; i<window; i++, windowFuncPtr++)
			signal_R[i] *= *windowFuncPtr;
    
	mayer_realfft(window, signal_R);
	tIDLib_realfftUnpack(window, windowHalf, signal_R, signal_I);
	tIDLib_power(windowHalf+1, signal_R, signal_I);
    
	// power spectrum sometimes generates lower scores than magnitude. make it optional.
	if(!x->powerSpectrum)
		tIDLib_mag(windowHalf+1, signal_R);
    
	tIDLib_filterbankMultiply(signal_R, x->normalize, x->filterAvg, x->x_filterbank, x->numFilters);
    
	for(i=0; i<x->numFilters; i++)
		SETFLOAT(listOut+i, signal_R[i]);
    
	outlet_list(x->x_featureList, 0, x->numFilters, listOut);
    
	// free local memory
	t_freebytes_(listOut, x->numFilters*sizeof(t_atom));
	t_freebytes_(signal_R, window*sizeof(t_sample));
	t_freebytes_(signal_I, (windowHalf+1)*sizeof(t_sample));    
    
}
コード例 #9
0
void *barkSpec_new(t_symbol *s, long argc, t_atom *argv)
{
	t_barkSpec *x = (t_barkSpec *)object_alloc(barkSpec_class);
    int i, isPow2;
    double timef = 0.0;
	//s=s;
    
	if (x) {
		dsp_setup((t_pxobject *)x, 1);	// MSP inlets: arg is # of inlets and is REQUIRED! 
										// use 0 if you don't need inlets
		//outlet_new(x, "signal"); 		// signal outlet (note "signal" rather than NULL)
		//x->offset = 0.0;
        
        
        x->x_featureList = listout(x);
        //x->x_featureList = outlet_new(&x->x_obj, &s_float);        
        
        if(argc > 1)
        {
            x->window = atom_getlong(argv);  // should perform a check for >64 && power of two
            isPow2 = (int)x->window && !( ((int)x->window-1) & (int)x->window );
           
            if(!isPow2)
            {
                error("requested window size is not a power of 2. default value of 1024 used instead.");
                x->window = 1024;
            };
            
            x->barkSpacing = atom_getfloat(argv+1);
        }
        else if(argc > 0)
        {
            x->window = atom_getlong(argv);
            isPow2 = (int)x->window && !( ((int)x->window-1) & (int)x->window );
            
            if(!isPow2)
            {
                error("requested window size is not a power of 2. default value of 1024 used instead.");
                x->window = 1024;
            };
            
            x->barkSpacing = 0.5;
        }
        else
        {
            x->window = 1024;
            x->barkSpacing = 0.5;
        }
        
        x->sr = 44100.0;
        x->n = 64.0;
        x->overlap = 1;
        x->windowFunction = 4; // 4 is hann window
        x->powerSpectrum = 0; // choose mag (0) or power (1) spec
        x->normalize = 1; // this is generally a good thing, but should be off for concatenative synth
        x->filterAvg = 0;
        //x->lastDspTime = clock_getlogicaltime();
        clock_getftime(&timef);
        x->lastDspTime = timef;
        x->sizeFilterFreqs = 0;
        x->numFilters = 0; // this is just an init size that will be updated in createFilterbank anyway.
        
        x->signal_R = (t_sample *)t_getbytes_((x->window+x->n)*sizeof(t_sample));
        
        // initialize signal buffer
        for(i=0; i<x->window+x->n; i++)
            x->signal_R[i] = 0.0;
        
        x->blackman = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        x->cosine = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        x->hamming = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        x->hann = (t_float *)t_getbytes_(x->window*sizeof(t_float));
        
        // initialize signal windowing functions
        tIDLib_blackmanWindow(x->blackman, x->window);
        tIDLib_cosineWindow(x->cosine, x->window);
        tIDLib_hammingWindow(x->hamming, x->window);
        tIDLib_hannWindow(x->hann, x->window);
        
        // grab memory
        x->x_filterbank = (t_filter *)t_getbytes_(0);
        x->x_filterFreqs = (t_float *)t_getbytes_(0);
        
        x->numFilters = tIDLib_getBarkFilterSpacing(&x->x_filterFreqs, x->sizeFilterFreqs, x->barkSpacing, x->sr);
        
        x->sizeFilterFreqs = x->numFilters+2;
        
        tIDLib_createFilterbank(x->x_filterFreqs, &x->x_filterbank, 0, x->numFilters, x->window, x->sr);
        
	}
	return (x);
}