[old tech]

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.

It is quite useful for LP playback with some records that may be slightly warped, that is the main reason why most high end phono pre-amps (or amps) have a subsonic filter function. Subsonic sound can also create distortions in audible frequencies either directly or through feedback through the turntable/cart set up. A good playback system will reveal the artefacts of some less than perfect LPs, hence the subsonic filter.

 

[Steve999]

Interesting discussion!!

When I used to use turntables and receivers had “subsonic” filter buttons pressing the filter button would get rid of all kinds of rumbly and grungy sounds that made themselves known in quiet passages or between songs when playback was at high volumes.

When I recorded my LPs to digital in the early oughts I used a “subsonic” filter in the software to get rid of those kinds of sounds in the recorded signal. However if I used that filter in conjunction with some types of click and crackle removal software in a certain order I’d get some bizarre kinds of low-level cyclical wishy wowy whooshing sounds so I just learned not to do that in exactly the wrong way.

I put “subsonic” in quotes because some of the sound, more felt than heard perhaps, could still be heard to some significant degree.

FWIW & etc.

 

[gregorio]

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

There is no phase shift with a linear phase filter, hence why it’s called a linear phase filter and, you are falsely equating phase shift with ringing. To get preringing you need a high signal level and a steep (many pole) filter. Studio mics are very insensitive at such low freqs, so unless you kick the mic stand you get very little 8-10Hz content. Although it’s true that you get more preringing with a linear phase filter in lower frequencies, more than what? 5x0 still equals zero, so how have you ascertained that “more pre-ringing” results in any pre-ringing at all and even if it does, that it’s at a level which will have any affect at all on speakers (that can’t reproduce it)?

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

Why are you trying to filter out freqs at or near DC in the first place and why are you using the wrong type of filter to achieve it? Subsonic (or rumble) filters, as mentioned by others, are sometimes used on/after turntables, which can produce low freq rumble due to warped vinyl and/or TT problems. But obviously you’re talking about digital audio, not vinyl, because you cannot implement a FIR filter in the analogue domain. So again, why are you trying to filter freqs near DC? Even if you have a valid reason (which you haven’t explained/demonstrated) why are you using a linear phase filter in the first place, when a minimum phase filter is the correct tool for the job? And lastly, as 71dB stated, the correct setting would be a 1 or 2 pole (6dB or 12dB per octave roll-off) filter.

well apparently enough to make it audible..

“i dont wanna go into theoretical audible limits”. And: “i dont wanna go into "audible" stuff”. - How is contradicting yourself a valid argument? And, why don’t you settle the argument you’re having with yourself BEFORE you argue with the rest of us?

As you are consistently making audibility claims though: How have you ascertained audibility? Presumably you haven’t, as you routinely claim audibility with no controlled testing for all sorts of inaudible nonsense. Humans can’t hear below about 20Hz and although we can “feel” freqs lower than that, we need extremely high SPLs to do so, levels that would cause hearing damage when replaying a music recording. Theoretically there could be an audible effect, IMD much higher in the freq spectrum for example but again, you’d need significant levels of very LF content and certain speakers susceptible to it.

You are filtering something that almost certainly never needs filtering, are using the wrong filter for the job with probably the wrong filter settings as well and claiming audibility with nothing (no evidence) except your expectation of pre-ringing from using the wrong filter type! Again, it’s all just nonsense!!

G

 

[Ghoostknight]

So again, why are you trying to filter freqs near DC?

because its "audible", "for whatever reason"

why are you using a linear phase filter in the first place, when a minimum phase filter is the correct tool for the job?

why is that? how is a phaseshift in the most crucial area "the right thing todo" since its audible too

And lastly, as 71dB stated, the correct setting would be a 1 or 2 pole (6dB or 12dB per octave roll-off) filter

who says i havent tried those?

There is no phase shift with a linear phase filter, hence why it’s called a linear phase filter and, you are falsely equating phase shift with ringing. To get preringing you need a high signal level and a steep (many pole) filter.

well i say the sentence i said before again
"does a IIR filter with X amount of phaseshift equate to a similar high level of preringing with FIR filter" which it apparently does
1. high pole filter in IIR = larger phaseshift
2. high pole filter in FIR = larger amount of preringing

+ the fact that lower frequency has worse influence on preringing, so a high pole subsonic filter is the worse thing todo hence why it was audible.
yes i also noticed that high pole filters make it worse but it wasnt inaudible at 12db slope either

Humans can’t hear below about 20Hz and although we can “feel” freqs lower than that, we need extremely high SPLs to do so, levels that would cause hearing damage when replaying a music recording. Theoretically there could be an audible effect, IMD much higher in the freq spectrum for example but again, you’d need significant levels of very LF content and certain speakers susceptible to it.

well apparently its enough to make it "audible" in the sense of i "notice" them wether i hear or feel them doesnt play any meaningful role here, maybe its just reduced IMD idk

 

[bigshot]

If it’s audible, it isn’t because of ringing or phase. I’m betting it’s user error.

 

[Ghoostknight]

If it’s audible, it isn’t because of ringing or phase. I’m betting it’s user error.

and what error i could possibly have been implementing?

also im quite unsure about the "the preringing at 10hz isnt audible" since normally on a high pass around 10hz the phaseshift in IIR filters start way higher then that like 30hz, tho im just guessing, isnt it possible that preringing is introduced in the same range the phaseshift of IIR filters were happening?

is the amount of phaseshift and at what frequency it starts etc directly correlated to the amount of preringing FIR introduces?

 

[71 dB]

And lastly, as 71dB stated, the correct setting would be a 1 or 2 pole (6dB or 12dB per octave roll-off) filter.

G

Actually, I didn't state this. In some circumstances those filter settings might be right for the job, but I was talking about low-shelf-filters.

If a minimum phase HP filter is used, I think the filter order can be the higher the lower the cut-off frequency is: For example 1st order at 20 Hz, 2nd order at 10 Hz, 3rd order at 7 Hz, 4th order at 5 Hz etc.

Having a 6-12 dB step down at 40 Hz with a low-shelf filter and a 3rd order subsonic filter at 7 Hz in series could be one "musical" solution to reduce signal energy below 50 Hz or so. That's what I have figured out, but then again I am just an amateur. Professionals probably do things some other ways...

 

[Ghoostknight]

what i dont get @gregorio if
1. preringing is inaudible
2. phaseshift is inaudible
3. subsonics are inaudible
4. filtering dc out could potentially help reduce IMD

whats wrong with implementing a subsonic filter in the digital domain?

 

[71 dB]

and what error i could possibly have been implementing?

also im quite unsure about the "the preringing at 10hz isnt audible" since normally on a high pass around 10hz the phaseshift in IIR filters start way higher then that like 30hz, tho im just guessing, isnt it possible that preringing is introduced in the same range the phaseshift of IIR filters were happening?

Ringing happens at the filter frequency. It is very local in frequency and that's why it "rings" (consists of very narrow band of frequencies). Phase shifts however spread in frequency and indeed a 10 Hz subsonic filter can have substantial phase shift at 30 Hz, but the ringing happens at 10 Hz.

is the amount of phaseshift and at what frequency it starts etc directly correlated to the amount of preringing FIR introduces?

They are related. Higher order filters create more phase shift and ringing (not all FIRs are minimum phase. FIRs can have any phase response).

 

[71 dB]

whats wrong with implementing a subsonic filter in the digital domain?

Nothing wrong with it, but if your digital domain is clean of low frequency garbage there isn't a need to use one.

 

[gregorio]

because its "audible", "for whatever reason"

Sure it is. An octave below the limit of human audibility is audible to an audiophile, ergo audiophiles aren’t human. Glad we sorted that out!

why is that? how is a phaseshift in the most crucial area "the right thing todo" since its audible too

The most crucial area is 10Hz? Better get on to all the universities, text books, encyclopaedias, scientists and sound engineers then, those idiots all think the crucial area is around 3kHz, nowhere near 10Hz.

the fact that lower frequency has worse influence on preringing, so a high pole subsonic filter is the worse thing todo hence why it was audible.

Firstly, the “worse thing to do” to what? It’s the worse thing to do to a full-scale or near full-scale signal and that’s what you have at 10Hz is it? As it isn’t, you have no idea if you’re actually getting any pre-ringing at all, let alone if your speakers are reproducing anything in response even if there were!

Secondly, if a high pole, low freq linear phase filter is “the worse thing to do”, why on earth are you doing it?

Lastly, “hence why it was audible” is the best example of expectation bias I’ve seen! That assertion isn’t even necessarily true in the most critical hearing band, let alone an octave outside the least sensitive hearing band!

So: What if there is some low level ringing, what if my speakers can magically reproduce 10Hz, what if you’re not human and can hear it, what if you’ve got crappy speakers creating audible IMD from low level very LF signals, what if it’s not expectation bias, what if a box of rubbish connected to ground really is magic, what if pigs can fly, the earth is flat and science/engineers are wrong about everything? That’s a lot of “what ifs” … I know, to play it safe I’ll do “the worse thing to do”. Flawless audiophile logic there, well done!!

G

 

[Ghoostknight]

The most crucial area is 10Hz? Better get on to all the universities, text books, encyclopaedias, scientists and sound engineers then, those idiots all think the crucial area is around 3kHz, nowhere near 10Hz.

well i can just speak for me but a phaseshift is way more audible under 100hz than 3khz

Firstly, the “worse thing to do” to what? It’s the worse thing to do to a full-scale or near full-scale signal and that’s what you have at 10Hz is it? As it isn’t, you have no idea if you’re actually getting any pre-ringing at all, let alone if your speakers are reproducing anything in response even if there were!

well just listen to some youtube demos about preringing, thats the effect im getting, soo im pretty sure its preringing but there is no point in argueing with you

Secondly, if a high pole, low freq linear phase filter is “the worse thing to do”, why on earth are you doing it?

well i created this thread to get the theory right, which apparently correlates with what im hearing, so i can avoid it
and i just wanted to get the theory right, not hearing some "oh but its all inaudible" bs

 

[bigshot]

and what error i could possibly have been implementing?

Lack of level matching, direct A/B switching and blind comparison averaged over multiple trials. It may sound different... it may sound the same... but you'll never know until you eliminate the most likely source of the error... your personal bias. Until you determine it exists objectively, there's absolutely no point hatching up crackpot theories about what might be causing it.

It also could be defective equipment of course. If after carefully applying controls to my listening comparison, I was getting audible differences with a filter setting like that, I would suspect something was broken somewhere.

 

[Ghoostknight]

Lack of level matching, direct A/B switching

how is the level supposed to change with just implementing a highpass? maybe the overall "pressure-level" changes because of lack of certain frequencys but beside that?

im also able to switch things in "realtime" with the eq im using

It may sound different... it may sound the same... but you'll never know until you eliminate the most likely source of the error... your personal bias. Until you determine it exists objectively, there's absolutely no point hatching up crackpot theories about what might be causing it.

well for me its enough if i can hear a change that seem consistent and not some "random stuff" my brain might make up

It also could be defective equipment of course. If after carefully applying controls to my listening comparison, I was getting audible differences with a filter setting like that, I would suspect something was broken somewhere.

ah well i dont know... i dont think so

 

[gregorio]

well i can just speak for me but a phaseshift is way more audible under 100hz than 3khz

Then you’re admitting that either you’ve got some serious hearing damage or that you’re not human. So either you need an urgent visit to an audiologist or do what I suggested and contact some universities, countless thousands of scientists would be more than interested in a person who isn’t human!

well just listen to some youtube demos about preringing, thats the effect im getting, soo im pretty sure its preringing but there is no point in argueing with you

What YouTube demos about pre-ringing? Presumably, as you’ve quoted it in another thread, you mean this one: The “effect you’re getting“ in that video is indeed easily audible BUT the HP filter is at around 200Hz, is high order, well above the speaker low freq roll-off, over 3 octaves above the 20Hz hearing threshold and is applied to a near full-scale signal.

You are right about one thing though, “there is no point arguing with me” because you are talking about 10Hz (rather than 200Hz), which is well below the low freq roll-off of your speakers, an octave below the threshold of human audibility AND is a low level signal as opposed to near full-scale signal. Your argument is utter nonsense!

well i created this thread to get the theory right, which apparently correlates with what im hearing, so i can avoid it

Exactly, glad you admitted it! … This quote is hilarious and the fact you don’t seem to realise it’s even funny just makes it even more hilarious!!

Let’s get your quote in context: You don’t actually ascertain “what you’re hearing”, whether you’re actually hearing any difference or just imagining one. Then you invent some ridiculous idea (which you falsely call a “theory”) that is contradicted by the established facts but “apparently correlates” with what you’re imagining and then you do “the worse thing to do”, “so you can avoid it [what you’re imagining]”. How does ANY of that sound remotely sane/rational even to you?

G