Lines Matching refs:coefShift
288 * below coefShift. See fir() for details.
324 * @param phaseWrapLimit is #polyphases<<coefShift, where #polyphases is the number of polyphases
325 * in the polyphase filter. Likewise, #polyphases can be obtained as (phaseWrapLimit>>coefShift).
327 * @param coefShift gives the bit alignment of the polyphase index in the phase parameter.
355 * The filter polyphase index is given by indexP = phase >> coefShift. Due to
359 * The fractional siting between the polyphase indices is given by the bits below coefShift:
361 * lerpP = phase << 32 - coefShift >> 1; // for 32 bit unsigned phase multiply
362 * lerpP = phase << 32 - coefShift >> 17; // for 16 bit unsigned phase multiply
366 * lerpP = phase << sizeof(phase)*8 - coefShift
371 * lerpP = (phase << 32 - coefShift) / (1 << 32); // floating point equivalent
378 const int coefShift, const int halfNumCoefs, const TC* const coefs,
389 uint32_t indexP = phase >> coefShift;
390 uint32_t indexN = (phaseWrapLimit - phase) >> coefShift;
402 uint32_t indexP = phase >> coefShift;
403 uint32_t indexN = (phaseWrapLimit - phase - 1) >> coefShift; // one's complement.
423 TC lerpP = TC(phase << (sizeof(phase)*8 - coefShift)) * scale;
428 uint32_t lerpP = phase << (sizeof(phase)*8 - coefShift)
