Audibility of digital reconstruction filters

[esldude]

There is lots of talk about the effects of digital reconstruction filters. You will often see impulse response or square waves with ringing of the digital filter shown. Talk of pre-echo being unnatural. How other filters will sound better, image better etc. etc.

 

 

 

 

You can see in the ringing the samples go up and down each sample taken. So the ringing must be up near the nyquist frequency. The filters only ring in the transition band between the passband and the stopband. For 44.1khz digital that is between 20 khz and 22.05 khz. I took some impulses and in software did the brick wall filter using 4081 taps (wanting lots of pre-echo). I then took the non filtered signal and inverted it to mix with the filtered signal. What is left is the difference in the signals caused by the filtering. As you can see it all resides above 20 khz.

 

 

 

You can play back this file and hear nothing. Because there is nothing below 20 khz. The two waveforms were identical except for the ultrasonics. So it seems unlikely this ugly looking pre-echo and ringing phenomenon would be heard. It also shows that the reconstruction below 20 khz is correct.

 

I did the same for several music files, and they cancel out below 20 khz as well. So there should be no audible result from the ringing.

 

I also created some impulse files at an 8 khz sample rate. Yes 8000 samples per second. I brick wall filtered it between 3200 hz and 4000 hz. The resulting ringing with the original signal subtracted out was audible as a thin tinny sounding tink at a low level on impulses.. As you would expect from 4 khz ringing. Converted some music to 8khz as well. Some transitions were of course above the filter and set it to ringing. It was barely discernible as modulated noise that mostly coincided with things like cymbals on the drum.

 

Now it is possible the unheard ringing can cause artifacts that would be in the audible range. Excite the resonance in a metal dome tweeter and it may upset its fidelity at a lower level. You get IMD whenever two tones pass through a non linear device. A constantly ringing tweeter that is also audibly non-linear might modulate IMD products as the signal changes. I have seen ESL panels and transformers on tube amps have a resonance between 20-25 khz. That too might inter-modulate the ringing with the heard signal to create artifacts. These artifacts are low in level. The first cycle of ringing by the filters is -21 db from the abrupt transition that causes it. It grows lower with each cycle. But -21 db is only about 1% of the power. So one would think most devices aren’t too stressed by it. And many 'abrupt' changes are lower in level to begin with.

 

One can go to 88 khz or 96 khz sample rates. That would push the ringing up clear of most tweeter or transformer resonances. Likely not a problem for most SS amps either at that level. Beyond that I wonder what could improve with higher sample rates.

 

Now some DACs are made with various filters other than the standard FIR filter. With the idea of having only post echo and various other aspects. Most of those alter phase response and/or frequency response in the upper frequencies. The phase may or may not be audible, the frequency response often is. So when comparing filters of different types you can’t be sure what makes one sound different from another. It is often assumed the pre-echo is one of the larger issues. But as it is too high in frequency to be heard I find that unlikely the reason for the sound difference. One need not think of backward masking if what needs masking is too high in frequency to be heard directly.

 

Now I am no digital filter expert. So comment from those more knowledgeable about what I may have wrong is welcome.

[13mh13]

There is lots of talk about the effects of digital reconstruction filters.

You 're at CA, not HA (hydrogenaudio), and are talkin' Greek to THIS crowd. So go to that place for your query ... and don't come back

[Don Hills]

You 're at CA, not HA (hydrogenaudio), and are talkin' Greek to THIS crowd. So go to that place for your query ... and don't come back

 

Quite right. Everyone knows that different filters sound different. It's night and day.

.

.

.

[sandyk]

It would be helpful if Charles Hanson were able to chime in here, but the timing of this thread isn't conducive to that.

[Fokus]

This broadly matches my experience.

 

But to the OP: one has to consider the ADC filter (anti-aliasing) and the DAC filter (reconstruction) together. Looking at the replay side in isolation is quite pointless.

[esldude]

This broadly matches my experience.

 

But to the OP: one has to consider the ADC filter (anti-aliasing) and the DAC filter (reconstruction) together. Looking at the replay side in isolation is quite pointless.

 

I agree. I seem to remember you work with ADC design. So I would welcome you explaining the matter in terms of ADC to go along with DAC end of things.

[Fokus]

You have to consider the cascade the two filters' impulse responses. If the ADC filter has done its job properly (i.e. fully complies to the sampling theorem's initial requirement) then the DAC filter plays no role (not unless it has been designed to play a role, deviate from transparency, of course).

 

It is interesting to figure in the two dominant ADC filter classes (old-style PCM1630 with analogue, minimum phase, filters versus digital half-band linear phase FIR filtering) with the marketing claims made about DAC filters. Most of these claims would fall flat on their ass.

[esldude]

You have to consider the cascade the two filters' impulse responses. If the ADC filter has done its job properly (i.e. fully complies to the sampling theorem's initial requirement) then the DAC filter plays no role (not unless it has been designed to play a role, deviate from transparency, of course).

 

It is interesting to figure in the two dominant ADC filter classes (old-style PCM1630 with analogue, minimum phase, filters versus digital half-band linear phase FIR filtering) with the marketing claims made about DAC filters. Most of these claims would fall flat on their ass.

 

So if I understand you correctly, the anti-alias filtering on the ADC would leave nothing that could cause additional artifacts in the DAC output with current FIR filtering. Artifacts in this case being ringing-like artifacts. The signal that is recorded will be correctly reconstructed on the DAC side.

 

I have done the brickwall filter on impulse and square waves in software. Take the result and run it through DA/AD and the resulting waveform is the same. It doesn't appear to get altered once the 20 khz plus signal has been removed. Alternatively doing a DA/AD on such signals, taking the result and doing another loop of DA/AD doesn't add to the artifacts on subsequent passes through.

[Jud]

Thanks for raising this topic, Dennis.

 

I think this is an appropriate place to raise a continuing question on filter behavior from another thread. The sequence is (1) my question; (2) Miska's response; (3) Fokus's response.

 

Something that interests me is what appeared to be a mild disagreement between Miska and Fokus in comments in this thread and the 192/48kHz one (though I could easily be misinterpreting).

 

Miska I think is looking at things from a time domain point of view and saying filter ringing should be audible as smearing of transients. Fokus I think is looking at things from a frequency domain point of view and saying for people like me who can hear to perhaps 15-16kHz, ringing should not be audible.

 

Am I accurately restating what each of you has said? Any further thoughts or comments?

 

There are two aspects, one is audibility of the ringing as such, and another thing is it's implications as transient smear. The smear length of typical 1 ms is equivalent of one cycle of 1 kHz sine wave, while the ringing frequency itself is around the Nyquist frequency. It rings for many cycles - the overall smear time is length of the entire ringing period.

 

Like in a video, you may not notice content of a single frame, but if there's "motion blur" (content of previous frames leaking to next one) over 25 consequent frames you will certainly notice it. If you think that a frame coming out now still would have 1% of content from one second ago mixed in. If you have a stationary image, you don't notice anything, but if you have fast movement, it becomes apparent.

 

I am always speaking of and thinking in both domains, as they are the same.

 

Take a linear phase filter with pre-ringing, once more.

 

If the filter cut-off is in the audible range then the pre-ringing is audible as a separate artefact.

 

If you move the filter cut-off frequency upwards then this audible artefact disappears.

 

If someone can provide me with sample music(*) files of two >20kHz filters that are identical, except in their amount of pre-ringing and that can demonstrate an audible difference to a majority of non-naive listeners then I would welcome that. I have been looking and trying for ten years, with listener panels, but so far ... nothing. So I lost all interest, I cook my filters pragmatically (just as Miska, sacrificing some above 18kHz to get to a shorter filter kernel, and always ensuring sufficient alias suppression - yes, I only care about the ADC side of things), and I carry on, concentrating on things that really matter.

 

(* Or even just a killer test sound.)

 

Miska and Fokus agree the ringing itself will not be audible for sufficiently high frequency filter cut-offs. But Miska goes on to say there will be an audible effect of the ringing, consisting of smearing of transients over a relatively long time compared to the ringing frequency (~ the period of a 1kHz sine wave). Fokus does not specifically comment about transient behavior, but does say frequency domain and time domain are the same.

 

So is Miska's remark about audible transient smearing something that others in the field can agree with? Do you instead think it is perhaps measurable but not audible? Or is it something you do not even see in the measurements?

 

Another interesting thing to me is Fokus's remark about sacrificing some response above 18kHz. Is that to get so-called "apodizing" behavior? I note in this connection that if I recall and understand his remarks correctly, Barry Diament, the producer, does not like what he hears from apodizing filters. (Barry likes the Metric Halo ULN-8. Anyone know what type(s) of filter(s) it uses?)

[dc2bluelight]

Just because a square wave shows "ringing" artifacts post filter doesn't mean that it's creating audible distortion. A theoretical square wave is made up of infinite harmonics. Simply limiting its bandwidth to something finite with anything other than a first-order filter will cause visible ringing: Low Pass Filtering Causes Ringing

 

That doesn't mean the artifact is audible. It typically isn't because of it's high frequency and low intensity.

 

Many point to in-band phase shift caused by a high-order LPF as a cause of audibility. However, experiments into the threshold of audibility of in-band phase shift as early as the mid 1980s proved it takes far more deliberate phase shift to become audible than any practical LPF would cause. (There are several AES papers on this, plus this link: Some Experiments With Time)

 

Finally, if anyone would like to take the trouble to compare the sound of digitally generated, mid-band square waves with various degrees of ringing caused by all-pass filtering (phase shift without changing frequency response), those signals can be generated easily in audio editing software. Comparison reveals that remarkably non-square square waves sound identical.

 

Regardless, with the possible exception of early digital recordings with multi-pole analog anti-aliasing filters, modern recording systems have reduced the in-band phase effects well below audibility through the use of digital filters, though they will still display the ringing square wave effects of band limiting.

 

The use of such colloquial terms as "pre-echo" (though often applied to Gibbs oscillation, more appropriately refers to artifacts of bit-rate reduction/compression systems) or "transient smearing" serve largely as strong bias factors, clouding subjective observations. Though the effects are real, the audible threshold of them is actually remarkably high.

 

A possibly more relevant investigation into the audible effects of filters might be through the principles of "spectral contamination" (Jensen and Sokolich, AES, 1988) which relate to the in-band products produced by complex intermodulation occurring in the top octave.

[Miska]

Another interesting thing to me is Fokus's remark about sacrificing some response above 18kHz. Is that to get so-called "apodizing" behavior? I note in this connection that if I recall and understand his remarks correctly, Barry Diament, the producer, does not like what he hears from apodizing filters. (Barry likes the Metric Halo ULN-8. Anyone know what type(s) of filter(s) it uses?)

 

Put simplified, by commonly used terminology apodizing filters are ones that replace original ringing with their own. Practically this can be thought to not apply for ADC filters.

 

AFAIK, it uses AK5394AVS as ADC converter chip. At 192k sampling rate:

Passband: 87.072 kHz

Stopband: 104.928 kHz

Ripple: +-0.007 dB

Stop band attenuation: 120 dB

Group delay: 63 samples

 

To me that is mediocre digital filter. I prefer stop band attenuation around 200 dB and ripple below 0.0000001 dB.

 

Plus a note "The analog modulator samples the input at 6.144 MHz for output word rate of 48 kHz. There is no rejection of input signal which are multiples of sampling frequency (that is: there is no rejection for n x 6.144 MHz +-21.768 kHz, where n=1,2,3...)"

Meaning that you need to make sure that there is absolutely nothing going to ADC chip analog input above 6.111232 MHz.

[Paul R]

Okay, taking your theory as true, what accounts for the fact that different filters produce a different sound? After all, the only difference between filters is post, pre, or no ringing, at least as far as your experiment and derived theory go.

 

Good science has to be able to explain every phenomena. Please do not go into the tired old "expectation bias" stuff, a good solid well backed up hypothesis would be welcome. Elsewise, this is just trolling and does belong somewhere else - like the gasbag forums.

 

-Paul

[Jud]

Okay, taking your theory as true, what accounts for the fact that different filters produce a different sound? After all, the only difference between filters is post, pre, or no ringing, at least as far as your experiment and derived theory go.

 

-Paul

 

?? I think the response was simply concentrating on ringing behavior, not saying that was the only difference among filters.

[Superdad]

?? I think the response was simply concentrating on ringing behavior, not saying that was the only difference among filters.

 

Okay, but as all of us Audirvana+/iZotope SRC users can attest, adjustment of the "ringing" slider (as well as all the other filter parameters) is VERY audible--right down to the limit of the granularity of available adjustment (0 to 1.0 in 0.01 steps; 0 being minimum-phase all post-ring, 1.0 being linear phase equal balance between pre and post).

 

My hearing barely goes out to 14kHz, and the effects I hear are all about the micro-second attack of the notes and how they propagate into the room. For me it rather audibly obvious.

 

It is also quite easy to radically better in s/w the performance of the average DAC with SRC to 176.4 or 352.8 to virtually bypass the chip's internal filter. I am not sure at all why the audibility of filter parameters is still controversial for some.

[dc2bluelight]

Okay, taking your theory as true, what accounts for the fact that different filters produce a different sound?

Science would say first, "Is there a difference?" To answer that, sadly, the experiment must eliminate biases of all kinds. Once established that there is a real and repeatably detectable difference, a hypothesis or three might be formed, then a proof sought.

After all, the only difference between filters is post, pre, or no ringing, at least as far as your experiment and derived theory go.

Ah. If only that were true and it was that simple.

Good science has to be able to explain every phenomena. Please do not go into the tired old "expectation bias" stuff, a good solid well backed up hypothesis would be welcome. Elsewise, this is just trolling and does belong somewhere else - like the gasbag forums.

 

-Paul

 

"Good Science" must by definition take into account the errors in the entire experimental system. Unfortunately, some of those errors include bias in observation, but there are many others as well. If as many errors as possible in the system are to be accounted for, then influence of biases must be included.

 

We could present hypotheses all day, but that would be less than satisfying. A hypotheses is only a beginning, and the scientist must go on to experimentally prove it. Much of this work has been done, though not published in main-stream press.

[Superdad]

Science would say first, "Is there a difference?" To answer that, sadly, the experiment must eliminate biases of all kinds. Once established that there is a real and repeatably detectable difference, a hypothesis or three might be formed, then a proof sought.

 

"Good Science" must by definition take into account the errors in the entire experimental system. Unfortunately, some of those errors include bias in observation, but there are many others as well. If as many errors as possible in the system are to be accounted for, then influence of biases must be included.

 

Oh boy, here we go again. And on the first page. zzzzzzzzz…..

[4est]

Yup, a snooze fest... Gotta wonder why you post this stuff Dennis.

Oh boy, here we go again. And on the first page. zzzzzzzzz…..

[Paul R]

(*sigh*) People can and do hear a difference. That is as close to being an established fact as it is possible for anything to be. Many of us hear a difference when we change the filter ringing parameters.

 

The biggest hypothesis made here is that people are not hearing a difference, or that it is not possible for people to hear a difference. Or more perniciously, that people are imagining such differences due to expectation bias or whatever else you want to call it.

 

Such an extraordinary hypothesis requires extraordinary proof, and would be much better served on the gasbag forums or other places dedicated to such hypothesis. Most people here are more interested in what they hear, than in what someone else thinks they can or cannot hear.

 

-Paul

 

 

 

Science would say first, "Is there a difference?" To answer that, sadly, the experiment must eliminate biases of all kinds. Once established that there is a real and repeatably detectable difference, a hypothesis or three might be formed, then a proof sought.

 

Ah. If only that were true and it was that simple.

 

 

"Good Science" must by definition take into account the errors in the entire experimental system. Unfortunately, some of those errors include bias in observation, but there are many others as well. If as many errors as possible in the system are to be accounted for, then influence of biases must be included.

 

We could present hypotheses all day, but that would be less than satisfying. A hypotheses is only a beginning, and the scientist must go on to experimentally prove it. Much of this work has been done, though not published in main-stream press.

[Superdad]

Yup, a snooze fest... Gotta wonder why you post this stuff Dennis.

 

Hang on here. I was snoozing in response to dc2blurlight's attempt to turn the discussion down the rabbit hole of proof and bias and all that usual nonsense argument we have been over and over, round and round.

 

I don't mind at all Dennis starting a technical discussion and asking questions about the measurement and audibility of digital filters. This is a topic of great interest to me and I especially like it when knowledgable people jump in to explain things from both sides. That's how we all learn. I'd love to see Miska, PeterSt., John Swenson, Hansen, Rankin, Siu, Mallison, etc., along with Don Hills, Mayhem, Fokus and the like all jump in and argue/post/graph this out.

 

I just get bored when conclusions and sides are drawn--and the whole measure versus hear BS--before more than a couple of experienced people even have a chance to jump in.

[Teresa]

You 're at CA, not HA (hydrogenaudio), and are talkin' Greek to THIS crowd. So go to that place for your query ... and don't come back

 

Quite right. Everyone knows that different filters sound different. It's night and day.

 

I agree with both posters would like to third this sentiment. In my opinion this type of theory analysis doesn't really belong in the General Forum perhaps the DAC - Digital to Analog Conversion Forum would be a better place or better yet as 13mh13 suggested Hydrogen Audio.

[4est]

Fair enough, I jumped the gun perhaps. I too would love to actually learn more than see this turn afoul.

Hang on here. I was snoozing in response to dc2blurlight's attempt to turn the discussion down the rabbit hole of proof and bias and all that usual nonsense argument we have been over and over, round and round.

 

I don't mind at all Dennis starting a technical discussion and asking questions about the measurement and audibility of digital filters. This is a topic of great interest to me and I especially like it when knowledgable people jump in to explain things from both sides. That's how we all learn. I'd love to see Miska, PeterSt., John Swenson, Hansen, Rankin, Siu, Mallison, etc., along with Don Hills, Mayhem, Fokus and the like all jump in and argue/post/graph this out.

 

I just get bored when conclusions and sides are drawn--and the whole measure versus hear BS--before more than a couple of experienced people even have a chance to jump in.

[sdolezalek]

I don't mind at all Dennis starting a technical discussion and asking questions about the measurement and audibility of digital filters. This is a topic of great interest to me and I especially like it when knowledgable people jump in to explain things from both sides. That's how we all learn. I'd love to see Miska, PeterSt., John Swenson, Hansen, Rankin, Siu, Mallison, etc., along with Don Hills, Mayhem, Fokus and the like all jump in and argue/post/graph this out.

 

 

+1!

[dc2bluelight]

(*sigh*) People can and do hear a difference. That is as close to being an established fact as it is possible for anything to be. Many of us hear a difference when we change the filter ringing parameters.

...except that's not all you're changing...

 

The biggest hypothesis made here is that people are not hearing a difference, or that it is not possible for people to hear a difference. Or more perniciously, that people are imagining such differences due to expectation bias or whatever else you want to call it.

 

 

Huh. OK, hit the nerve there, didn't I?

 

I never said, "people are not hearing a difference, or that it is not possible for people to hear a difference. Or more perniciously, that people are imagining such differences due to expectation bias". Perhaps a re-read? All I'm saying is we need to know for certain they are hearing a difference. Is that so unreasonable? I also said, if they do hear a difference, we can hypothesize as to why, then go find out for sure. Is that also so unreasonable?

 

I cited some references that showed a few things that have been proven to not be as audible as we thought. Is that the problem?

 

Look, kids, you hear what you hear. I do get a bit tired of betting shot at just for trying to apply a bit of science and reason to find out why.

 

Most people here are more interested in what they hear, than in what someone else thinks they can or cannot hear.

 

I certainly agree with that. But in the quest for audio nirvana, wouldn't it benefit everyone to know what sounds better and why, that tech can be integrated into more stuff?

 

But if it makes everyone happy, I will gladly accept the "evil scientist" label.

 

...and down I go...down my own personal rabbit hole.

[Jud]

I want to concentrate on the audibility of different filters built by experienced people. That's simply because it's quite possible whatever combination of parameters I didn't like the sound of in iZotope SRC (or what I *did* like the sound of) is something no pro would ever do. I would expect less obvious variation in the sound of filters designed by pros.

 

That being said, I was floored by the degree of audible variation among the filter options provided in Miska's HQPlayer software. I would urge anyone curious about the topic of this thread to try it.

[RealAudio]

(*sigh*) People can and do hear a difference. That is as close to being an established fact as it is possible for anything to be. Many of us hear a difference when we change the filter ringing parameters.

 

The biggest hypothesis made here is that people are not hearing a difference, or that it is not possible for people to hear a difference. Or more perniciously, that people are imagining such differences due to expectation bias or whatever else you want to call it.

 

Such an extraordinary hypothesis requires extraordinary proof, and would be much better served on the gasbag forums or other places dedicated to such hypothesis. Most people here are more interested in what they hear, than in what someone else thinks they can or cannot hear.

 

-Paul

 

I think a more precise way to say it that "people claim to perceive a difference". Can they really, well that needs more than just asserting that people can actually hear a difference.