Question about FIR Filtering

[Ghoostknight]

Hello,

i noticed something today that i kinda must have overheared the last time... (i was utilizing a subsonic filter which made it kinda obvious) the question is

Do FIR Filters introduce "preringing" dynamicly depending on what filters you use? for example, would a filter that introduce a large phaseshift in IIR filtering also introduce equally large amount of preringing with FIR?

if this is the case, and i think it is, instead of looking out for phaseshifts you have to look out for increasing preringing with "worse" filters in FIR to keep negatives as low as possible, right? (tho the preringing audibility is debatable and i kinda dont wanna get into this here)

 

[bigshot]

This isn't my area of expertise, but I would think it would depend on the amount of the shift, not the dynamics. And I would also guess that if it was bad enough to be audible, it would affect transients like snare drum hits, not just dynamically loud stuff. You'd have to push it to get it into the audible range.

But I'm sure you're able to convince yourself you can hear anything you really want to.

 

[Ghoostknight]

You'd have to push it to get it into the audible range.

this might be just a problem with a "subsonic filter", my thinking is that a high pass (that usually have quite the phaseshift) that low used as subsonic filter might be extra "dangerous" regarding preringing, if i understand things right the FIR filtering has to add quite a delay (which ends up being worse preringing) to compensate for a phaseshift that low to keep it linear phase, much worse than with higher frequencys, so probably best avoided? tho correct me if im wrong

atleast i noticed quite the change with turning the subsonic filtering on and off and it kinda also sounded like lows fill the "gaps" between louder parts (if they are not far apart) which went away after turning the subsonic high pass off, it kinda sounded more in tune again

 

[bigshot]

Could the subsonic filter have spill into the audible range?

 

[Ghoostknight]

Could the subsonic filter have spill into the audible range?

not really, it was starting at 12-16hz and -3db 10hz

 

[bigshot]

That shouldn't be a problem. If it is, there's something wrong somewhere else.

 

[Ghoostknight]

This isn't my area of expertise, but I would think it would depend on the amount of the shift, not the dynamics.

what exactly you mean by dynamics? like gain settings of filters? these usually have more phaseshift if the gain is higher (like for bell and high shelfs) but high pass phaseshift is usually much worse, specially since the phaseshift stays the way it is, with bell and highshelf filters the phaseshift corrects itself after half the filter bandwidth

i dont wanna go into theoretical audible limits but just answer the simple question if i got things right, FIR introduces more preringing if there is a low frequency phaseshift, right?

the subsonic filter is quite audible, the question is what im hearing
1. subsonics being filtered
2. preringing
3. less distortion on higher frequencys since the speaker is less loaded

to be honest i think im able to detect all 3 to different degrees but now the question is with what compromise i wanna go, specially if my preringing theory is right since i think this was nearly the most audible (atleast it was the most annyoing side effect)

 

[gregorio]

not really, it was starting at 12-16hz and -3db 10hz

So it wasn’t starting at 12-16Hz then. If it was starting at 12-16Hz then at 10Hz the attenuation would have been 0dB, there would have been no attenuation below 12Hz.

i dont wanna go into theoretical audible limits

What’s the difference between theoretical audible limits and practical audible limits? And, if you don’t want to “go into” either then why does the rest of your post do exactly that?

but just answer the simple question if i got things right, FIR introduces more preringing if there is a low frequency phaseshift, right?

What low frequency phase shift, the whole point of FIR filters is to not introduce any phase shift. And, preringing is a function of the signal level being filtered, the steepness/attenuation of the filter and the frequency.

but now the question is with what compromise i wanna go, specially if my preringing theory is right

You don’t seem to have any understanding of filters (why are you applying a filter at 10Hz or 12-16Hz anyway?) plus, every “theory” I’ve seen you mention is not actually a theory, it’s just some idea you’ve invented (out of ignorance) that is contradicted by the actual facts! So the chance that your “preringing theory is right” is extremely remote at best, even though you haven’t really explained it yet!

G

 

[Ghoostknight]

What low frequency phase shift, the whole point of FIR filters is to not introduce any phase shift. And, preringing is a function of the signal level being filtered, the steepness/attenuation of the filter and the frequency.

im talking about filters that would have had a phaseshift with IIR filters
the same filters that introduce phaseshift in IIR also introduce a "equally large" amount of preringing with FIR filtering, right? thats my question

And, preringing is a function of the signal level being filtered, the steepness/attenuation of the filter and the frequency.

i think i have the right thinking then, and lower frequency introduce way more preringing then higher frequencys being filtered, right?

So it wasn’t starting at 12-16Hz then. If it was starting at 12-16Hz then at 10Hz the attenuation would have been 0dB, there would have been no attenuation below 12Hz.

i was playing around with different slopes but usually looked out for that they are starting under 16hz ... the -3 db point was usually around 10hz
i also tried some starting at 16-30hz but this was "too" audible for me, i just wanna take off some of the "bad load" from the speaker not cut the audible frequencys

What’s the difference between theoretical audible limits and practical audible limits? And, if you don’t want to “go into” either then why does the rest of your post do exactly that?

well i wanted to share my experience, feel free to crush my believes/hearing again
i dont wanna go into "audible" stuff because the main question for me is if filter that would have introduced a large phaseshift with IIR filtering, specially in low frequencys, introduce a somewhat relateable amount of preringing with FIR filters
this would mean to avoid the most preringing possible with FIR filtering is to keep the phaseshift under around 100hz that would have occoured with IIR filters still as low as possible with FIR filtering, so FIR filters have the same negative sideeffect of IIR filters, just that its Phaseshift vs amount of Preringing if i understand things right

of course we can discuss absolute audibility but for me its kinda clear, if FIR preringing gets introduced dynamicly depending on filters being used and some can introduce more or less preringing to the whole chain than this is what i heared with playing around with the subsonic high pass, since its theoretically the worse kind of preringing introducing filter

 

[gregorio]

the same filters that introduce phaseshift in IIR also introduce a "equally large" amount of preringing with FIR filtering, right?

How do you figure that out and how do you equate preringing with phase shift?

i think i have the right thinking then, and lower frequency introduce way more preringing then higher frequencys being filtered, right?

If you’re contradicting the facts I’m quoting how on earth can you think you “have the right thinking then”? Again, preringing is a function of the signal level, the steepness/attenuation of the FIR filter and the frequency, NOT of only the frequency!

i was playing around with different slopes but usually looked out for that they are starting under 16hz ... the -3 db point was usually around 10hz

So then the filter must have been starting below 10Hz. Again, why would you want to apply a HP filter at say 8Hz?

well i wanted to share my experience
feel free to crush my believes/hearing again
i dont wanna go into "audible" stuff

Firstly, this is the sound science forum, not the share your experiences forum.

Secondly, you’re not only sharing your experiences, you’re also sharing all sorts of nonsense beliefs/ideas you’ve made-up (contrary to the facts).

Lastly, you keep saying you “don’t wanna go into audible stuff”, while doing exactly that in the very same post! “i also tried some starting at 16-30hz but this was "too" audible for me”.

Honestly, how much signal have you got at 8Hz, how much ringing (even if there were any) do you think your speakers could reproduce at 8-10Hz and even if they could reproduce some, how would a human ever hear it?

G

 

[bigshot]

the subsonic filter is quite audible, the question is what im hearing
1. subsonics being filtered
2. preringing
3. less distortion on higher frequencys since the speaker is less loaded

3 is the purpose of a sub sonic filter. -3dB shouldn’t be enough to cause 2. 1 is possible because ultra low frequencies can be felt. They aren’t imperceptible like super audible frequencies are. But you’d need a sub woofer to get down below 15Hz. Could you have a low frequency hum of some sort?

Most music wouldn’t contain sub sonic frequencies down around 10Hz. Are you using tones or something? It sounds like you’re trying to break the filter or something.

You’re the one that mentioned dynamics. I assumed you meant that it followed the relative level of the music, but maybe you didn’t.

 

[Ghoostknight]

Again, preringing is a function of the signal level, the steepness/attenuation of the FIR filter and the frequency, NOT of only the frequency!

but its worse at lower frequencys because lower frequency phaseshifts equats to more delay needed to make it linear phase which equates to more preringing

So then the filter must have been starting below 10Hz. Again, why would you want to apply a HP filter at say 8Hz?

its called subsonic filter to filter out "DC" or frequencys near DC

Honestly, how much signal have you got at 8Hz, how much ringing (even if there were any) do you think your speakers could reproduce at 8-10Hz and even if they could reproduce some, how would a human ever hear it?

well apparently enough to make it audible..

 

[bigshot]

well apparently enough to make it audible..

And now we’re back to the fact that you aren’t making the least effort to determine if it is audible. You’re trusting your expectation bias that’s fed by your desire to believe that your ears have superhuman abilities to hear the unhearable. You aren’t qualified to say whether it’s audible or not.

 

[71 dB]

The lowest frequencies are interesting. I avoid highpass-filtering sounds at bass frequences, because somehow they tend to loose their integrity. Our hearing is used to sounds having their spectrum density level increase toward the lower frequencies and "too aggressive" highpass-filtering seems to violate this principle. If I need to make space for other bassy sounds in a mix, I use low-shelf-filters that don't go to zero at DC. The phase response is also better. Subsonic filters work at very low frequency (e.g. 5 Hz). So, I combine a low-shelf filter (say -6 dB at 63 Hz) and subsonic filter at 5 Hz. I don't understand the need to highpass filter agressively almost everything to make room for other sounds in the mix. That creates unnatural sound lacking all energy below the filter frequency. Lowering the level by say 6 dB under say 200 Hz with a lowshelf-filter might be enough to create space for other sounds and if it isn't then try 12 dB, but it is good to leave something there unless it's just unwanted noise etc.

 

[bigshot]

High pass filtering of sub sonic frequencies is good for eliminating frequencies that make cheap portable headphones distort at higher frequencies. I don't see any purpose for it for a good home audio system.