Sallen Key Butterworth?

[gregorio]

I think you have some sort of issue with communicating with other people. Not entirely sure what it is, but it prevents you from engaging in any sort of back and forth discussion.

You’re the one who made a false assertion and have “some sort of issue” communicating when it’s refuted. You could just accept it and move on, or you could argue on a scientific/factual basis if my refutation were incorrect. Unfortunately you do neither and instead argue irrelevant and personal ad hominems.

Go on thinking whatever you want, and I'll go on thinking what I think.

Indeed, I’ve already said you can think or say whatever you want and I’ll “go on” refuting it, if you post it in this subforum and it’s incorrect/false.

The irony is that we are both saying the same thing.

Really, where did I say those two filters would provide an audibly transparent result?

G

 

[bigshot]

Has anyone else ever commented on the way you communicate?

 

[AnalogEuphoria]

So the equipment I’m talking about here is the Cambridge audio Minx XI 2.1 All in one system which I use with my tv on optical.

Specs​

Power output 40 watts rms (into 8 Ohms)

DAC Wolfson WM8728 24-bit 96kHz

THD (unweighted) <0.01% @ 1kHz <0.09% 20Hz - 20kHz

Frequency response (-1dB) 5Hz - 50kHz

S/N ratio (ref 1W) -85dB relative to 1W A weighted

Total correlated jitter <250pS

My concern here was with the digital filter used if it’s not great would the weak attenuation cause audible aliasing, then should I need to “upgrade” to external unit but as I’ve said it sounds great to me no matter if it’s movies or music playback. To reiterate, yes the interpolation may not be correct but are any of these “poor” filters actually doing anything that could be audible or paranoia over nothing? As we know only poorly mastered music rings anyway and even then it’s higher up than we can hear so that wouldn’t contribute to it.

 

[AnalogEuphoria]

delete

 

[bigshot]

You're already into the range of overkill. Don't worry about it.

 

[VNandor]

My concern here was with the digital filter used if it’s not great would the weak attenuation cause audible aliasing, then should I need to “upgrade” to external unit but as I’ve said it sounds great to me no matter if it’s movies or music playback. To reiterate, yes the interpolation may not be correct but are any of these “poor” filters actually doing anything that could be audible or paranoia over nothing?

I'll try to keep my answer short. Your concerns are unfounded for a variety of reasons:

1) A poor interpolation filter used for reconstruction would never cause aliasing. Aliasing potentially happens during AD conversion. This is not a semantic difference, the frequencies above nyquist remaining due to a poor interpolation filter would not wrap back to the audible range due to aliasing. Frequencies remaining above nyquist due to improper filtering do wrap back during AD conversion though.

2) You already have your device and it sounds good. Technically, this might not be enough, for example if it could potentially damage your other gear but that is very unlikely, especially if you consider the next point:

3) Here's the DAC chip's datasheet. There's nothing that implies a horribly bad interpolation filter in place. Both of the available interpolation filters certainly provide enough attenuation to prevent any potential problems I'm aware of. Also an interesting note, as far as I know, these filter responses can not come from a 2 pole butterworth filter, I don't know where you got that information about your DAC but that is wrong.

 

[AnalogEuphoria]

I'll try to keep my answer short. Your concerns are unfounded for a variety of reasons:

1) A poor interpolation filter would never cause aliasing. Aliasing potentially happens during AD conversion. This is not a semantic difference, the frequencies above nyquist remaining due to a poor interpolation filter would not wrap back to the audible range due to aliasing. Frequencies remaining above nyquist due to improper filtering do wrap back during AD conversion though.

2) You already have your device and it sounds good. Technically, this might not be enough, for example if it could potentially damage your other gear but that is very unlikely, especially if you consider the next point:

3) Here's the DAC chip's datasheet. There's nothing that implies a horribly bad interpolation filter in place. Both of the available interpolation filters certainly provide enough attenuation to prevent any potential problems I'm aware of. Also an interesting note, as far as I know, these filter responses can not come from a 2 pole butterworth filter, I don't know where you got that information about your DAC but that is wrong.

Sorry if I wasn’t clear the filter response picture I posted of minimum phase wasnt from this DAC specifically but I was curious about which of the two that the butterworth is closest to.

 

[bigshot]

Both of those charted responses are essentially the same to human ears. If it sounds good through your system, you’re golden. Don’t overthink it.

 

[AnalogEuphoria]

Thanks you guys have been helpful

 

[AnalogEuphoria]

While we’re on the subject I have a question for people far more intelligent than me, some believe that a filter is not needed on a DAC because your ears can actually filter out any “junk” naturally is this true?

 

[71 dB]

While we’re on the subject I have a question for people far more intelligent than me, some believe that a filter is not needed on a DAC because your ears can actually filter out any “junk” naturally is this true?

Not that I think I am far more intelligent than you, but I answer anyway: Our ears do filter out the "junk" because it is outside our hearing range. The problem is that the junk is there whether we hear it or not and it can create other issues. Audio devices may have difficulties with excessive amount of ultrasonic content and this can lead to intermodulation distortion that generates distortion in audible range (audible junk)! Also, the "junk" can even act as high frequency electric interference for other electronic devices around. Not using a deconstruction filter on a DAC is "possible", but kind of ugly/anarchistic and theoretically wrong. Better use proper filtering and do it correctly without "junk."

The non-filtering DACs are audiophool snake oil products trying to be something special and "pure." As it has been said many times in this thread, reconstruction filtering is not a real problem in audio in 21th century. Different types of reconstruction filters can be told apart from each other only with NOS DACs (44.1 kHz sampling), but even then it is not about fidelity, but mild changes in flavour (for me it is a few percent change in the apparent width of the soundstage). In oversampling DACs reconstruction filters are totally transparent for human ears.

 

[gregorio]

My concern here was with the digital filter used if it’s not great would the weak attenuation cause audible aliasing …

As VNandor stated, there wouldn’t be any aliasing, let alone audible aliasing. The problem for the audiophile industry is that even in the late 1980’s many DAC chips were effectively audibly flawless and by around the mid/late 1990’s they pretty much all were, even the DACs found in relatively cheap consumer CD and DVD players and even under quite extreme listening test conditions. To keep charging premium prices, the audiophile manufacturers had to keep coming up with spurious “problems” and providing effectively snake oil solutions for these non-problems. In other words, an anti-image filter does not need to be “great” to be audibly transparent it just needs to be adequate and they have been adequate for many years. A “great” filter today might look impressive on paper but makes no audible difference, even filters that are comparatively “weak” still have no audible artefacts. The datasheet for your relatively old (2008) DAC chip does indicate a filter with “weak” attenuation compared to current ones; -60dB attenuation is not as “great” as more than -100dB attenuation, but 60dB attenuation is still a level reduction of 1,000 times!

While we’re on the subject I have a question for people far more intelligent than me, some believe that a filter is not needed on a DAC because your ears can actually filter out any “junk” naturally is this true?

It’s not really a question of intelligence, just some knowledge of how digital audio works and relates to what humans can hear. This isn’t as easy as it sounds though, because how digital audio works (and what humans can hear) is routinely misrepresented in the audiophile world and can be somewhat unintuitive anyway, which is why the audiophile misinformation is so widely accepted/successful.

A consequence of the sampling/digitisation process is the creation of “images” mirrored above and below the sampling frequency (and multiples of the sampling freq), so even from the time the sampling theory was first publicly proposed (in 1927 by Harry Nyquist) a brick-wall filter was mandated at half the sampling frequency (now called the “Nyquist Frequency”). Not having this mandated filter is therefore a “broken” design. While it’s true that the human ear effectively filters out content above 20kHz (more like 16kHz in adults) the signal output from the DAC obviously has to pass through amps and speakers before it gets to your ears. Amps and speakers produce an artefact called IMD (Inter-Modulation Distortion), which is the creation of spurious “sum and difference” tones. These tones/frequencies are not harmonically related to the input frequencies (and are therefore more noticeable/audible) and can occur within the audible spectrum even if the input signal only contains frequencies outside the audible spectrum. In any half decent amp or speakers this IMD should be well below audibility. However, all amps and speakers have an optimal operating bandwidth and the higher the input frequency/the closer to it’s optimal bandwidth limit, the more IMD they will produce. Furthermore, the higher the level of this high/ultrasonic content, the more IMD will be produced. Not having an anti-image filter in the DAC therefore provides a “double whammy”, not only ultrasonic content (“images”) which should not be there but at high levels and therefore the probability of audible IMD is high. And of course, the human ear will not “filter out this junk [IMD]” because it’s within the audible spectrum.

In addition, not oversampling or having an anti-image/reconstruction filter results in a frequency roll-off starting around 2kHz, which of course is also well within the audible spectrum, can also result in timing error and typically more HF noise. The result is that Non-OverSampling (NOS) filterless DACs can be relatively easy to differentiate under controlled listening testing and are effectively a broken/faulty design. In true audiophile tradition this can be falsely marketed as “better”, “more musical” or other superlatives and some audiophiles will happily pay a hefty premium for the privilege of owning such a broken/faulty DAC!

G

 

[gregorio]

So I stated:

You could just accept it and move on, or you could argue on a scientific/factual basis if my refutation were incorrect. Unfortunately you do neither and instead argue irrelevant and personal ad hominems.

And your very next response is:

Has anyone else ever commented on the way you communicate?

That’s just classic, you couldn’t have proven my point more perfectly!! No doubt you’ll just keep doing it ad infinitum now.

G

 

[71 dB]

In addition, not oversampling or having an anti-image/reconstruction filter results in a frequency roll-off starting around 2kHz,...

G

Are you sure you wrote that correctly?

 

[AnalogEuphoria]

That answered my questions, just seems like a load of misinformation and psuedo science then.