On the subject of "ringing"

[Shadders]

23 minutes ago, Fokus said:

 

Wrong.

 

Take this input signal, sampled at 96kHz. It occupies a band from DC to 47 kHz, i.e. it is properly band-limited and thus a legal signal in the 96k space. Observe its clean, monotonic leading edge.

 

 

If we downsample this signal to 48kHz with a steep linear phase filter cutting at 24 kHz we get this

 

 

Our linear phase filter has imprinted its pre-ringing at its transition frequency.

 

 

Hi Fokus,

I am not wrong. Please read what i have written. I stated a 192kHz sample rate, and a filter with a cut off at 50kHz.

Maybe, stop trying to tell everyone they are wrong and read what is written instead.

Regards,

Shadders.

[Shadders]

26 minutes ago, Hifi Bob said:

Shall dropbox the files soon, when I have a moment.

Hi Hifi Bob,

Thanks.

Regards,

Shadders.

[Hifi Bob]

2 hours ago, Hifi Bob said:

 

sox castanets-2496.flac ringing-1.flac rate 44100
sox castanets-2496.flac ringing-2.flac rate -s 44100
sox castanets-2496.flac ringing-3.flac rate -b 99.7 44100

 

 

 

Here are the files:  https://www.dropbox.com/s/sofy0eia1apjglz/ringing.zip?dl=0

[Shadders]

3 minutes ago, Hifi Bob said:

 

Here are the files:  https://www.dropbox.com/s/sofy0eia1apjglz/ringing.zip?dl=0

Hi Hifi Bob,

Thanks. i have downloaded.

Regards,

Shadders.

[Fokus]

1 hour ago, Shadders said:

 

Maybe, stop trying to tell everyone they are wrong and read what is written

 

Not everyone. You are often wrong, and you really don’t get it.

 

Your example had a 40 to 60 kHz signal and a 50 kHz filter.  Mine was a 0 to 47 k signal and a 24 k filter, no difference. Sampling rate plays no role, this is only about the filtering. That I used a donwsampling step is for practical reasons only, as this is for me today the quickest way of generating the required steep filters.

[Shadders]

1 minute ago, Fokus said:

 

Not everyone. You are often wrong, and you really don’t get it.

 

Your example had a 40 to 60 kHz signal and a 50 kHz filter.  Mine was a 0 to 47 k signal and a 24 k filter, no difference. Sampling rate plays no role, this is only about the filtering. That I used a donwsampling step is for practical reasons only, as this is for me today the quickest way of generating the required steep filters.

Hi Fokus,

I am not wrong. You mis-read or misinterpret what is written to create a false response.

Please provide evidence where i am often wrong ?. Thanks.

 

I have mode NO mention of downsampling. Why you persist in repeating this, despite i am not talking about it, and my example does NOT state anything about downsampling, shows you intentionally lie about others to state that they are wrong.

 

Regards,

Shadders.

[mansr]

1 minute ago, Shadders said:

I have mode NO mention of downsampling. Why you persist in repeating this, despite i am not talking about it, and my example does NOT state anything about downsampling, shows you intentionally lie about others to state that they are wrong.

The downsampling in Fokus' example is incidental and irrelevant. Use the 'sinc' effect in Sox to apply filters without resampling. You'll get the same effect.

[Shadders]

12 minutes ago, mansr said:

The downsampling in Fokus' example is incidental and irrelevant. Use the 'sinc' effect in Sox to apply filters without resampling. You'll get the same effect.

Hi Mansr,

OK- i will examine later.

What my example shows is that when the filter is operating in its steady state response, then as longs as there are no gaps in the samples, then no matter what frequency is energising the filter, there is no ringing. (In general - if you energise with a impulse within the signal - then that will cause ringing - but for standard audio in the ranges pertaining to normal audio - it does not occur)

 

It has been stated by people on this forum that "hitting the filter with the filter cut off frequency causes ringing" - not exact words, but close.

 

This is not correct. If the filter is working in the steady state response region, then there is no ringing. A sound file that has no gaps introduces no ringing except at the start and end of the file.

 

I have proven this using Octave - i generated a sweep frequency from 100Hz to 20kHz with a constant linear rate of change (no discontinuities) for a 192kHz sample rate, cut off frequency at 10kHz approx, and you can see the passband, transition band and stop band, and from the difference signal (output - input, ensuring time alignment), then there is no ringing in the sweep signal.

 

There is ringing at the beginning and end of the signal as i have referred to above - transient response + steady state response.

 

If you implement a CPFSK (Continuous Phase Frequency Shift Keying) signal, where the transition between the frequencies is at the zero-crossing, then that does introduce ringing at every transition in frequency - simply because of a discontinuity between the frequencies.

 

I will respond to Fokus's response with the ringing example.

 

Regards,

Shadders.

[Shadders]

2 hours ago, Fokus said:

 

Wrong.

 

Take this input signal, sampled at 96kHz. It occupies a band from DC to 47 kHz, i.e. it is properly band-limited and thus a legal signal in the 96k space. Observe its clean, monotonic leading edge.

 

 

If we downsample this signal to 48kHz with a steep linear phase filter cutting at 24 kHz we get this

 

 

Our linear phase filter has imprinted its pre-ringing at its transition frequency.

 

 

Hi Fokus,

All you have done here is created a signal with finite number of samples that has a bandwidth of 47kHz. This is a short (by comparison) signal.

 

You have then passed it through a linear phase filter which you state is steep.

 

What you have not provided is the relevant lengths of the 47kHz signal nor the number of taps of the steep linear phase filter.

 

All you have done in the  filtering to prepare to downsample , is to energise the linear phase filter with a short signal where that signal length is comparable to the filter length. What you are seeing is the predominant transient response and possibly some of the steady state response (without knowing the signal length and filter length - not able to state accurately).

 

The last picture you can see that the filter response IS NOT 100% symmetrical, but very nearly. This indicates that the transient response is dominant.

 

So all you are seeing is a transient response of the filter - which does indeed, include pre ringing and post ringing, where the post ringing is slightly off compared to the pre ringing due to the input signal affecting the post ringing more.

 

This engineering aspect of transient response and steady state response is simple University 2nd year Electronic Engineering analysis.

 

Happens in the analogue world too, but without the pre ringing.

 

Regards,

Shadders.

[mansr]

6 minutes ago, Shadders said:

Happens in the analogue world too, but without the pre ringing.

I'm tempted to build an analogue filter with "pre-ringing" just because.

[Shadders]

Just now, mansr said:

I'm tempted to build an analogue filter with "pre-ringing" just because.

Hi mansr,

Yes - but you would have to intend to build one - does not happen with the usual topologies used in any signal conditioning. It is a rarity.

Regards,

Shadders.

[psjug]

4 hours ago, Fokus said:

 

Take this input signal, sampled at 96kHz. It occupies a band from DC to 47 kHz, i.e. it is properly band-limited and thus a legal signal in the 96k space. Observe its clean, monotonic leading edge.

 

 

 

I know you know this, but you should point out that this input signal (band limited impulse) is very unnatural and should not be in a recording.  If you play it is sounds pretty nasty.  It may look like ringing a bell, but it sure won't sound like it!

[sandyk]

 

Some time back, IIRC, () John Kenny provided samples of pre ringing and normal, and most were able to hear the difference.

I can't remember the details, so John Kenny should be able to provide the details of his test if he was asked.

[esldude]

4 hours ago, mansr said:

I'm tempted to build an analogue filter with "pre-ringing" just because.

It would be worth it if people believed you.  Unfortunately even though it should enlighten many people about how such things work I fear it wouldn't.  

[buonassi]

just got caught up.  in reading the last couple pages, I keep seeing attempts to prove ringing doesn't exist or shouldn't matter if it does.  I'm not even going to try and pretend I can follow some of the posts as I'm not a physicist or electrical engineer.   

 

if ringing doesn't exist, or doesn't matter, then why are there different filter designs and why do they sound different to me?  Are the opponents of ringing going to assert that I'm imagining the differences?  That "expectation bias" is to blame?  

 

I'm not going to state that one filter is better than another.  But am I the only one who can admit to hearing a difference among them?  Have all the hours I've scrutinized iZotope upsampling parameters, observing the rolloff steepness in an RTA, and noted sonic differences all been a hard lesson in realizing I'm "drinking the coolaid"?

[Hifi Bob]

11 minutes ago, buonassi said:

just got caught up.  in reading the last couple pages, I keep seeing attempts to prove ringing doesn't exist or shouldn't matter if it does.  I'm not even going to try and pretend I can follow some of the posts as I'm not a physicist or electrical engineer.   

 

if ringing doesn't exist, or doesn't matter, then why are there different filter designs and why do they sound different to me?

Ringing does exist, as I showed in the spectrograms above, but it seems only with very steep filters.

 

If you can hear it, then it’s probably due to a non-linearity in your playback chain—possibly a damaged tweeter.

[fas42]

45 minutes ago, buonassi said:

if ringing doesn't exist, or doesn't matter, then why are there different filter designs and why do they sound different to me?  Are the opponents of ringing going to assert that I'm imagining the differences?  That "expectation bias" is to blame?  

 

 

They can sound different, because the electronic behaviour needed to action them may impact the analogue side of the audio chain; this is in the same realm as, how a source PC is set up alters the quality of sound.

 

A fairly recent experience for me was listening to a rather ambitious, older CDP that had a good range of switchable filters available - which was the best setting? Answer: no filtering!! Every time any filter, at any setting was used, the sound degraded, and also varied in the degree of negative impact - nulling all of that extra electronic activity yielded the optimum quality. IOW, if you're going to go fancy, you must get the implementation 100% clean, or you could easily be worse off ...

[buonassi]

5 minutes ago, Hifi Bob said:

Ringing does exist, as I showed in the spectrograms above, but it seems only with very steep filters.

 

If you can hear it, then it’s probably due to a non-linearity in your playback chain—possibly a damaged tweeter.

single dynamic driver per channel headphones - focal elear.  Same on my HD600s and even on my phase aligned balanced armature in ears (JH/AK Angie).  

 

A properly setup speaker system (sans sub) should uncover ringing given that frequencies above 100hz travel roughly at the same speed.  Yet I read somewhere that it's not as evident on a free field speaker system given reflections, room effects, standing waves at low frequencies, etc.  But that was purely conjecture.  

 

[Shadders]

4 hours ago, buonassi said:

just got caught up.  in reading the last couple pages, I keep seeing attempts to prove ringing doesn't exist or shouldn't matter if it does.  I'm not even going to try and pretend I can follow some of the posts as I'm not a physicist or electrical engineer.   

 

if ringing doesn't exist, or doesn't matter, then why are there different filter designs and why do they sound different to me?  Are the opponents of ringing going to assert that I'm imagining the differences?  That "expectation bias" is to blame?  

 

I'm not going to state that one filter is better than another.  But am I the only one who can admit to hearing a difference among them?  Have all the hours I've scrutinized iZotope upsampling parameters, observing the rolloff steepness in an RTA, and noted sonic differences all been a hard lesson in realizing I'm "drinking the coolaid"?

Hi,

The discussion is not that ringing does not occur, it does, and only occurs at the beginning and end of a track if there are no gaps in the track.

I analysed the castanets from Hifi Bob - and ringing does occur.There are gaps between the castanet sounds. So if you are listening to a rock track, for example, then there is no ringing (as long as there are no gaps - very low sounds followed by a burst of music)

If listening to a track with quiet passages, and there is a burst of sound, then ringing does occur (pre-ringing).

If the filter length is long, then the pre ringing is long, and likewise, if the filter length is short, then the ringing is short. This is based on symmetrical filters. I used the fir1 filter in Octave to simulate.

Regards,

Shadders.

[Spacehound]

4 hours ago, buonassi said:

just got caught up.  in reading the last couple pages, I keep seeing attempts to prove ringing doesn't exist or shouldn't matter if it does.  I'm not even going to try and pretend I can follow some of the posts as I'm not a physicist or electrical engineer.   

 

if ringing doesn't exist, or doesn't matter, then why are there different filter designs and why do they sound different to me?  Are the opponents of ringing going to assert that I'm imagining the differences?  That "expectation bias" is to blame?  

 

I'm not going to state that one filter is better than another.  But am I the only one who can admit to hearing a difference among them?  Have all the hours I've scrutinized iZotope upsampling parameters, observing the rolloff steepness in an RTA, and noted sonic differences all been a hard lesson in realizing I'm "drinking the coolaid"?

My DAC has several different  filters and also supports MQA. Though  thankfully it automatically switches all the MQA stuff off  completely when it is not playing an MQA file and reverts to the filter you had previously selected.

 

Which  filter is "best"? Dunno, I don't have a clue how it is supposed to sound.

 

And neither do you   

[semente]

5 hours ago, buonassi said:

just got caught up.  in reading the last couple pages, I keep seeing attempts to prove ringing doesn't exist or shouldn't matter if it does.  I'm not even going to try and pretend I can follow some of the posts as I'm not a physicist or electrical engineer.   

 

if ringing doesn't exist, or doesn't matter, then why are there different filter designs and why do they sound different to me?  Are the opponents of ringing going to assert that I'm imagining the differences?  That "expectation bias" is to blame?  

 

I'm not going to state that one filter is better than another.  But am I the only one who can admit to hearing a difference among them?  Have all the hours I've scrutinized iZotope upsampling parameters, observing the rolloff steepness in an RTA, and noted sonic differences all been a hard lesson in realizing I'm "drinking the coolaid"?

 

I am generally quite sceptical and wasn't expecting to hear a clear difference but that's what happened when I recently compared linear vs. minimum phase using HQPLAYER poly-sinc-xtr filters. Even thought about running a poll to have an idea of people's preferences.

I prefer linear phase but if I compare it with minimum phase there's something not quite right about it... I wonder if @Miska has a filter with 1/3 pre-ringing and 2/3 post- (is it even possible?).

[Miska]

9 minutes ago, semente said:

I prefer linear phase but if I compare it with minimum phase there's something not quite right about it... I wonder if @Miska has a filter with 1/3 pre-ringing and 2/3 post- (is it even possible?).

 

For PCM yes, "asymFIR". And on my TODO-list for SDM outputs too. From my personal listening experience point of view, I feel it is more like "bad of both worlds" instead of "good of both worlds".

 

[semente]

1 minute ago, Miska said:

 

For PCM yes, "asymFIR". And on my TODO-list for SDM outputs too. From my personal listening experience point of view, I feel it is more like "bad of both worlds" instead of "good of both worlds".

 

Ok, thanks. I'll give it a listen.

[adamdea]

2 hours ago, Shadders said:

The discussion is not that ringing does not occur, it does, and only occurs at the beginning and end of a track if there are no gaps in the track.

....

If listening to a track with quiet passages, and there is a burst of sound, then ringing does occur (pre-ringing).

 

I think the second formulation is more helpful. The talk of gaps in tracks is confusing 

 

It really can happen in the middle of a track. And since you can't ever have complete silence it follows that it does not need to be a series of zero value samples before the impulse. There must be a threshold below which it is buried in the noise. I'm not sure whether that means there is no pre-ringing or merely that it is disguised. Certainly though we know that the front half of the filter is not just creating spuriae it is reconstructing signal.

 

Is there any way of generalising when the linear phase filter will create pre-ringing? Clearly there needs to be spectral content in the transition band of the filter.  But is there no pre-ringing when it is in not visible, or is it just disguised/buried?  It seems that there must be pre-ringing when there is spectral content in the transition band of the transient event but no such spectral content before the transient (so that the effect of the filter must be spurious), but what if there is? 

 

In any event, the $64,000 question for me is whether even if we could hear pre-ringing it could possibly account for the sense a lot of people have that somehow conventionally produced 16/44 was not enough and that other things sound better. As far as I can see the whole pre-ringing conjecture was borne out of desperation because it is the only real "problem" which can be identified with the original 16/44 spec once implementation issues are taken out of the equation. Now the problem for me has always been that what people claim to experience with higher resolution tracks (or alternative filtering strategies for 16/44)   just does not seem to map to the circumstances in which ringing really does occur. A load of ancient tracks with no hf signal worth mentioning were reissued as sacd. Many piano recordings have sod all hf either. 

 

The possibility that people can detect the ringing in some recordings at some moments is IMHO entirely beside the point if ringing cannot account for claims which are made. 

[psjug]

2 hours ago, Shadders said:

 

If listening to a track with quiet passages, and there is a burst of sound, then ringing does occur (pre-ringing).

 

Archimago's test did not find any ringing in this kind of scenario; the only ringing he found was in bad, clipped recordings that are therefore not properly band limited .  If there is ringing in properly made  recordings of music, even with a burst as you describe, someone should show it doing a test similar to Archimago's.