I agree, my wording was poor. On forums where we make off-the-cuff remarks, it sometimes happens.
Not a problem, I sit here and write without thinking always 🙂 But I still argue it is off-topic re transient responses, barring any discussion of the increased audibility of pre-responses
Perhaps, but the topic also includes Ben reconciling his views with Earls. I don't have my mods hat on when I say that, it just seems reasonable to me.
Soundbloke, so you say that using linear phase crossovers and FIR filter that produce pre-ring is not OK?
I used IIR and linear phase with FIR and hear something is not ok using FIR, but on the measurements all looks very good. Its like the sound is a little blured. It is tight but the highs and harmonics somehow are not OK.
Can you elaborate please on a gentle slope for phase (what angle for example?) and earlier in the thread you mentioned something about 5hz, what do you mean about that 5hz value?
Thanks!
I used IIR and linear phase with FIR and hear something is not ok using FIR, but on the measurements all looks very good. Its like the sound is a little blured. It is tight but the highs and harmonics somehow are not OK.
Can you elaborate please on a gentle slope for phase (what angle for example?) and earlier in the thread you mentioned something about 5hz, what do you mean about that 5hz value?
Thanks!
I said that there are applications when linear phase filtering causes audible problems: There appears no general rule of which I am aware other than long pre-responses increases the chance of problems occurring. Equally there are some applications where using non-linear phase filters causes problems too.
In crossovers specifically, generating long pre-responses also increases the chances of off-axis "pre-errors" which might also lead to audible problems in certain circumstances - and more likely so if you are a not single listener sat at the ideal listening position. IMHO it does not generate "blurred" sounds, however, so maybe something else is less than optimal?
The 'gentle slope' was a term someone else used, not me. Essentially the slope will be at an angle dictated by the overall delay, and the lower the frequency phase compensation works, so the greater the delay will need to be and the greater the angle. The 5Hz is IMHO where there seems little audible benefit from phase compensating below. But for compensation this low, you need also take account of the proximity effect too.
I also need state that I am not saying 5Hz is audible, just as a cut-off frequency, it serves (me) well in avoiding phase issues at higher frequencies that are audible.
In crossovers specifically, generating long pre-responses also increases the chances of off-axis "pre-errors" which might also lead to audible problems in certain circumstances - and more likely so if you are a not single listener sat at the ideal listening position. IMHO it does not generate "blurred" sounds, however, so maybe something else is less than optimal?
The 'gentle slope' was a term someone else used, not me. Essentially the slope will be at an angle dictated by the overall delay, and the lower the frequency phase compensation works, so the greater the delay will need to be and the greater the angle. The 5Hz is IMHO where there seems little audible benefit from phase compensating below. But for compensation this low, you need also take account of the proximity effect too.
I also need state that I am not saying 5Hz is audible, just as a cut-off frequency, it serves (me) well in avoiding phase issues at higher frequencies that are audible.