Svep ger oftast lägre brusgolv (om man använder program som räknar ut en impulsrespons baserad på den inspelade filen) men har andra nackdelar. Om du inte använder ett akustikmätningsprogram som skapar impulser så spelar det mindre roll, kör vitt brus (så starkt högtalarna klarar utan att låta konstigt) så borde resultatet vara en rät linje (om du analyserar med FFTn i Wavelab) men se till att bara analysera det område där filen spelas tills precis innan den slutar spela.
"These results show that under high S/N conditions the log-sweep gave best results, while at moderate and low S/N best results were achieved with a pink RPMS excitation. The averaging helps to reduce tail noise level in all cases. Bad results for MLS and pink RPMS excitation are at largest power level. MLS shows large PIR spikes, while pink RPMS pushes loudspeaker to large displacement at low frequencies, thereby producing large distortions. This is the reason why we implemented pink RPMS with variable low-frequency roll-off (20Hz-1000Hz).
Further experiments, with more uneven noise spectrum and temporal patterns (cracks, clicks, hand clapping), have revealed that swept-sine system unevenly distribute noise across the estimated impulse response (typical example shows Fig. 4), and in many cases the averaging do not increases the measurement S/N as expected. In such a cases random phase excitation gives better results." *
* http://www.fesb.hr/~mateljan/arta/papers/im-aaaa2003.pdf sid 8