As i understand it, a speaker driver is always a minimum phase system. Therefore "errors" in amplitude will always correspond to errors in phase response. If that is true then if i correct a peak using minimum phase EQ, the phase should also be corrected. The definition of correct here is flat FR and phase.
When, if ever, is it then recommended to use linear-phase gain EQ, when the goal is flat FR and phase?
Note that i am not talking about linear phase crossovers, i understand the benefit of that.
When, if ever, is it then recommended to use linear-phase gain EQ, when the goal is flat FR and phase?
Note that i am not talking about linear phase crossovers, i understand the benefit of that.
speaker drivers are non minimum phase in the breakup region as there are multiple acoustic sources. This also applies to speakers where the drivers are not located within a quarter wavelength at crossover. Hence why linear phase speakers tend to use coaxial drivers or multiple entry horns.
Its also worth mentioning that humans are pretty insensitive (at best) to phase anomalies or almost all speakers would sound terrible.
Its also worth mentioning that humans are pretty insensitive (at best) to phase anomalies or almost all speakers would sound terrible.
I see. So an ideal source is minimum phase in theory, but add in acoustic effects such as baffle diffraction and resonances and it becomes non minimum phase. In those cases (ie. all cases) linear phase EQ can make sense.
I'm using rephase and miniDSP HD, so by mixing linear and minimum phase EQ i could get a flatter phase response near the crossover region.
I'm on the fence on the matter of phase audibility. But i figure as long as it doesn't compromise factors that matter more, i can pursue flat phase as a secondary goal.
I'm using rephase and miniDSP HD, so by mixing linear and minimum phase EQ i could get a flatter phase response near the crossover region.
I'm on the fence on the matter of phase audibility. But i figure as long as it doesn't compromise factors that matter more, i can pursue flat phase as a secondary goal.
Last edited:
The problem with trying to correct (some? all?) non-minimum-phase issues with DSP of any type is that the correction will only be valid at one location (e.g. listening position.) In fact, it may make matters worse at other locations. As the old saying goes, sometimes the cure is worse than the disease.
Resonances are minimum phase.
Do you have a reference for this?kipman725 said:
Breakup is based on minimum phase effects, but varies with angle which can't be fixed with EQ so if it is unacceptable, it should be removed.
Almost never recommended and is easy to make things worse 🙂
I have used it when making changes to Mid and side signals to try and correct for phantom centre tonality shift after early reflections have been absorbed. Using minimum phase messes with this so linear works better.
Changing phase only can help with SBIR related nulls at a single listening position if speaker positions or multiple sources can't be used.
Practical example how thing can go wrong with linear-phase EQ and filters, can be seen and also listened: rePhase, a loudspeaker phase linearization, EQ and FIR filtering tool