Hope this help you to explain Hi-Res music to your CD friends

[danadam]

I have a hunch that these started with goldensound apparently passing blind ABX on two different DAC filters, and OP probably concludes that NOS sound different (to subjectivists) and thus negating that DACs that measure 0.0000x% THD+N sound the same

It started around here:
https://www.audiosciencereview.com/...g-16-44-1-collection-a-good-idea.53641/page-6
Although his account is removed now and all that's left are quotes in replies from other people.

 

[sunjam]

It's a messed up filter, there are no staircases with the other filters, which is evidence that this filter in particular is the problem(and not CD or whatever narrative your blog tries to sell).
The original article spends a lot of time on that filter and showing how it's a weird filter that does not attenuate out of band frequencies nearly enough. It's because of how little the filter removes above sample rate/2 that we get to see the staircases we should only see without a filter(which is also showed on the other staircases screenshot from, this time, a DAC with no filter).
For straight lines like a momentary DC signal, you need infinite bandwidth. A properly band limited signal does not have that. No matter how you spin that example, the result will remain the same. A correctly filtered signal would never look like that.

It's funny how out of 6 filters, you only looked at the clearly defective "let's recreate vintage crap" one, and invented some story about hires out of it. IDK if it's an attempt to generate traffic or your idea of a fun April fools experiment where you try to propagate the old staircase story to see how gullible people are?

It is not a messed up filter. The NOS filter attempt to faithfully reproduce the digital input (please let me explain later)

Food for thought:

1. What would you expect to see if we feed the same perfect 10kHz perfect sine wave in 768k / 32 bit format Hi-Res format to the same Topping E30 with F5 (NOS filter)?
Would we see a similar stair step output as 44.1k / 16 bit CD format. Would the 10kHz analog "sine" wave output from 768k / 32 bit be "smoother" than the one from CD format?

2. To my understanding, Topping E30 (and similar modern DACs) does have analog filter stage (for all its digital filter settings) but it is not very effective (as shown in the original article) as it is very hard to build a perfect analog filter. This is also the reason why modern DACs use digital filter in addtion to the existing analog filter)

3. Why people still want F5 (NOS filter)? IMHO, the original article provided a very good reason. I suggest you to read the whole article word by word as I did. To be honest, I found his article really "open my eyes". I think he did a very good job for the Hi Fi / audio science world if people really want to know exactly how digital filters works in modern DACs in reality.

He mentioned:

What I’ve found is three main differences produced by the different filter options:

• Different treatment in the high frequency response;
• Differences in the amount and timing of any ringing in impulse response; and
• Differences in the generation of unwanted ultrasonic noise or artefacts.

It is the amount and timing of any ringing in impulse response.
Which of the following filter give you the "best" response for an impulse input signal?

F1:

F2:

F3:

F4:

F5:

F6:

Which one would you pick just based on the impluse response? F1? F2? or F5?
(I'll pick F5, not sure about your answer)

4. If the output from a CD format is like what the "Monty's video" suggested (i.e. the DAC will output "perfect sine wave"), which filter in E30 would you consider as properly built, not broken?

5. If F5 is classified as broken or not properly built, how about the other filters (F1, F2, F3, F4, F6)? Are they also broken (for other reasons)? My answer is Yes, all filters are broken if F5 is considered as broken. Agree?

6. As I stated above, the original article "opened my eyes". Could the article open other people's eyes? Well... I am not sure.

I am sorry to say that for the people who blindly trust "the Monty's video" or blindly listen to the professor in their digital audio / signal processing classes would simply believe that the sampling theory guarantee you will get a perfect 10 kHz sine wave out of a properly built DAC with 44.1 kHz sampling....

I believe people who have a second thought (or who apply their own critical thinking) would come up with a different answer with the help of the original article, i.e. How to pick the best filter setting for your DAC – Addicted To Audio My blog regarding the "Monty's video" is just to highlight what people may miss. Of course, the bottom line is IF you want to know how digital filters work in modern DACs.

(To be honest, I am happy to discuss with you guys. Here is indeed having a group of people who have a different calibre than the ones I met on ASR. Cheers)

Feel free to comment and let me think

 

[sunjam]

It started around here:
https://www.audiosciencereview.com/...g-16-44-1-collection-a-good-idea.53641/page-6
Although his account is removed now and all that's left are quotes in replies from other people.

Thanks for pointing it out. Yes, I used the same graphs to explain to the people there but what I got back was just "bullying".

Too bad that they removed the most important discussion on the thread. It was having 20 pages and now it is down to 15 pages. Their action, IMO, confirms that I did something right.

Feel free to check out my blog for the missing part of that discussion. Cheers.

 

[theveterans]

It is not a messed up filter. The NOS filter attempt to faithfully reproduce the digital input (please let me explain later)

Food for thought:

1. What would you expect to see if we feed the same perfect 10kHz perfect sine wave in 768k / 32 bit format Hi-Res format to the same Topping E30 with F5 (NOS filter)?
Would we see a similar stair step output as 44.1k / 16 bit CD format. Would the 10kHz analog "sine" wave output from 768k / 32 bit be "smoother" than the one from CD format?

2. To my understanding, Topping E30 (and similar modern DACs) does have analog filter stage (for all its digital filter settings) but it is not very effective (as shown in the original article) as it is very hard to build a perfect analog filter. This is also the reason why modern DACs use digital filter in addtion to the existing analog filter)

3. Why people still want F5 (NOS filter)? IMHO, the original article provided a very good reason. I suggest you to read the whole article word by word as I did. To be honest, I found his article really "open my eyes". I think he did a very good job for the Hi Fi / audio science world if people really want to know exactly how digital filters works in modern DACs in reality.

He mentioned:


It is the amount and timing of any ringing in impulse response.
Which of the following filter give you the "best" response for an impulse input signal?

F1:


F2:


F3:


F4:


F5:


F6:


Which one would you pick just based on the impluse response? F1? F2? or F5?
(I'll pick F5, not sure about your answer)

4. If the output from a CD format is like what the "Monty's video" suggested (i.e. the DAC will output "perfect sine wave"), which filter in E30 would you consider as properly built, not broken?

5. If F5 is classified as broken or not properly built, how about the other filters (F1, F2, F3, F4, F6)? Are they also broken (for other reasons)? My answer is Yes, all filters are broken if F5 is considered as broken. Agree?

6. As I stated above, the original article "opened my eyes". Could the article open other people's eyes? Well... I am not sure.

I am sorry to say that for the people who blindly trust "the Monty's video" or blindly listen to the professor in their digital audio / signal processing classes would simply believe that the sampling theory guarantee you will get a perfect 10 kHz sine wave out of a properly built DAC with 44.1 kHz sampling....

I believe people who have a second thought (or who apply their own critical thinking) would come up with a different answer with the help of the original article, i.e. How to pick the best filter setting for your DAC – Addicted To Audio My blog regarding the "Monty's video" is just to highlight what people may miss. Of course, the bottom line is IF you want to know how digital filters work in modern DACs.

(To be honest, I am happy to discuss with you guys. Here is indeed having a group of people who have a different calibre than the ones I met on ASR. Cheers)

Feel free to comment and let me think

As much as you want a perfect digital filter, there's no such thing. However, objectively speaking, the ideal filter would be the one that performs best on the FFT Spectrum with the steepest stopband at Nyquist-Shannon bandlimit because it's the most transparent digital filter available. It's actually called sinc filter. However, as you increase tap length to make the passband flatter and stopband steeper, you also increase latency and thus DAC manufacturers had to compromise with a simpler FIR filter. Then you have to pay a penalty in the impulse response of the FFT signal with lots of pre and post ripples that you provided using the F1 filter (linear phase sharp roll-off). Whether those preringing and post ringing that's not on the original sample is audible or not is debatable. The opposite is to have no oversampling where in this case, you have no attenuation above the Nyquist-Shannon bandlimit and a gentle FR roll-off at human hearing frequencies thus it's NOT transparent to the original signal (so much for that "purist" factor lol). However, you get the best impulse response with nonoversampling and filterless because there's no FIR filter to add those ripples in the pre and post impulse. And the killer is the aliasing introduced in the analog domain where extensive analog filtering above the passband is needed or else you end up with tons of IMD in the audible range.

 

[bigshot]

Audible transparency is the goal, not theoretical perfection. One is easily achievable. One isn't.

 

[Quazar]

It is not a messed up filter. The NOS filter attempt to faithfully reproduce the digital input (please let me explain later)

Food for thought:

1. What would you expect to see if we feed the same perfect 10kHz perfect sine wave in 768k / 32 bit format Hi-Res format to the same Topping E30 with F5 (NOS filter)?
Would we see a similar stair step output as 44.1k / 16 bit CD format. Would the 10kHz analog "sine" wave output from 768k / 32 bit be "smoother" than the one from CD format?

2. To my understanding, Topping E30 (and similar modern DACs) does have analog filter stage (for all its digital filter settings) but it is not very effective (as shown in the original article) as it is very hard to build a perfect analog filter. This is also the reason why modern DACs use digital filter in addtion to the existing analog filter)

3. Why people still want F5 (NOS filter)? IMHO, the original article provided a very good reason. I suggest you to read the whole article word by word as I did. To be honest, I found his article really "open my eyes". I think he did a very good job for the Hi Fi / audio science world if people really want to know exactly how digital filters works in modern DACs in reality.

He mentioned:


It is the amount and timing of any ringing in impulse response.
Which of the following filter give you the "best" response for an impulse input signal?

F1:


F2:


F3:


F4:


F5:


F6:


Which one would you pick just based on the impluse response? F1? F2? or F5?
(I'll pick F5, not sure about your answer)

4. If the output from a CD format is like what the "Monty's video" suggested (i.e. the DAC will output "perfect sine wave"), which filter in E30 would you consider as properly built, not broken?

5. If F5 is classified as broken or not properly built, how about the other filters (F1, F2, F3, F4, F6)? Are they also broken (for other reasons)? My answer is Yes, all filters are broken if F5 is considered as broken. Agree?

6. As I stated above, the original article "opened my eyes". Could the article open other people's eyes? Well... I am not sure.

I am sorry to say that for the people who blindly trust "the Monty's video" or blindly listen to the professor in their digital audio / signal processing classes would simply believe that the sampling theory guarantee you will get a perfect 10 kHz sine wave out of a properly built DAC with 44.1 kHz sampling....

I believe people who have a second thought (or who apply their own critical thinking) would come up with a different answer with the help of the original article, i.e. How to pick the best filter setting for your DAC – Addicted To Audio My blog regarding the "Monty's video" is just to highlight what people may miss. Of course, the bottom line is IF you want to know how digital filters work in modern DACs.

(To be honest, I am happy to discuss with you guys. Here is indeed having a group of people who have a different calibre than the ones I met on ASR. Cheers)

Feel free to comment and let me think

"Which one would you pick just based on the impluse response? F1? F2? or F5?"
Definitely the F1!

 

[Quazar]

@sunjam
The CS43198 is a modern DS DAC. Oversampled by default.
The NOS mode is only a simulation!

 

[bigshot]

@sunjam
The CS43198 is a modern DS DAC. Oversampled by default.
The NOS mode is only a simulation!

Defective by design.

 

[Quazar]

Defective by design.

Yes, Yes...
A real NOS DAC can only be multibit (R2R).

 

[sunjam]

As much as you want a perfect digital filter, there's no such thing. However, objectively speaking, the ideal filter would be the one that performs best on the FFT Spectrum with the steepest stopband at Nyquist-Shannon bandlimit because it's the most transparent digital filter available. It's actually called sinc filter. However, as you increase tap length to make the passband flatter and stopband steeper, you also increase latency and thus DAC manufacturers had to compromise with a simpler FIR filter. Then you have to pay a penalty in the impulse response of the FFT signal with lots of pre and post ripples that you provided using the F1 filter (linear phase sharp roll-off). Whether those preringing and post ringing that's not on the original sample is audible or not is debatable. The opposite is to have no oversampling where in this case, you have no attenuation above the Nyquist-Shannon bandlimit and a gentle FR roll-off at human hearing frequencies thus it's NOT transparent to the original signal (so much for that "purist" factor lol). However, you get the best impulse response with nonoversampling and filterless because there's no FIR filter to add those ripples in the pre and post impulse. And the killer is the aliasing introduced in the analog domain where extensive analog filtering above the passband is needed or else you end up with tons of IMD in the audible range.

Cool, thanks a lot for the reply. I agreed that there is no perfect digital filter.

Food for thought:

What if we try to push the Nyquist frequency to 768kHz (using 1536kHz sampling)?

i.e.
case one: 1536k Hz digitized input of an analog audio input signal, play back with E30, NOS filter
case two: 44.1k Hz digitized input of the same analog audio input signal , play back with E30 and the filter you considered as the best amongst all 6 filters available

For the audio signal output, which output would be closer to the original analog audio input?

 

[castleofargh]

It is not a messed up filter. The NOS filter attempt to faithfully reproduce the digital input (please let me explain later)

Food for thought:

1. What would you expect to see if we feed the same perfect 10kHz perfect sine wave in 768k / 32 bit format Hi-Res format to the same Topping E30 with F5 (NOS filter)?
Would we see a similar stair step output as 44.1k / 16 bit CD format. Would the 10kHz analog "sine" wave output from 768k / 32 bit be "smoother" than the one from CD format?

2. To my understanding, Topping E30 (and similar modern DACs) does have analog filter stage (for all its digital filter settings) but it is not very effective (as shown in the original article) as it is very hard to build a perfect analog filter. This is also the reason why modern DACs use digital filter in addtion to the existing analog filter)

3. Why people still want F5 (NOS filter)? IMHO, the original article provided a very good reason. I suggest you to read the whole article word by word as I did. To be honest, I found his article really "open my eyes". I think he did a very good job for the Hi Fi / audio science world if people really want to know exactly how digital filters works in modern DACs in reality.

He mentioned:


It is the amount and timing of any ringing in impulse response.
Which of the following filter give you the "best" response for an impulse input signal?

F1:


F2:


F3:


F4:


F5:


F6:


Which one would you pick just based on the impluse response? F1? F2? or F5?
(I'll pick F5, not sure about your answer)

4. If the output from a CD format is like what the "Monty's video" suggested (i.e. the DAC will output "perfect sine wave"), which filter in E30 would you consider as properly built, not broken?

5. If F5 is classified as broken or not properly built, how about the other filters (F1, F2, F3, F4, F6)? Are they also broken (for other reasons)? My answer is Yes, all filters are broken if F5 is considered as broken. Agree?

6. As I stated above, the original article "opened my eyes". Could the article open other people's eyes? Well... I am not sure.

I am sorry to say that for the people who blindly trust "the Monty's video" or blindly listen to the professor in their digital audio / signal processing classes would simply believe that the sampling theory guarantee you will get a perfect 10 kHz sine wave out of a properly built DAC with 44.1 kHz sampling....

I believe people who have a second thought (or who apply their own critical thinking) would come up with a different answer with the help of the original article, i.e. How to pick the best filter setting for your DAC – Addicted To Audio My blog regarding the "Monty's video" is just to highlight what people may miss. Of course, the bottom line is IF you want to know how digital filters work in modern DACs.

(To be honest, I am happy to discuss with you guys. Here is indeed having a group of people who have a different calibre than the ones I met on ASR. Cheers)

Feel free to comment and let me think

A Dirac pulse is not music! There is a reason why we often talk about a Dirac pulse as an illegal signal. The concept is infinitely high amplitude at instant t, and nothing before or after. In the real world with sound, that would require the power of teleportation, and even that would only solve half, the appearing or disappearing, not both consecutively at the same place and basically the same time.
Any decent DAC can and will filter the signal in a way that shows no staircases!

An actual filterless NOS DAC, if it could do a perfect job(not physically possible), would give, not a spike but a lonely tower looking signal(like a portion of a square wave).
A real one would show ringing at the base and at the top of that "tower" on both sides. And it would indeed show staircases. It would have low fidelity because obviously those staircases shouldn't be there. They're like very many ultrasonic frequencies that didn't exist in the band limited record.
A NOS DAC with a filter designed by someone with a brain would try to remove as much aliasing as possible by attenuating the ultrasonic content. But doing that well with an analog filter is expensive and still nowhere near perfect. That's why I said that many of those designs have audibly rolled off treble, because they start filtering earlier to achiever a higher attenuation in the ultrasounds.
The only bonus of that design is usually low noise floor. The rest isn't great(fidelity wise) compared to delta sigma and a mix of digital and analog filter.

That F5 filter is indeed not that, the DAC is not a R2R DAC in the first place, and delta sigma just doesn't work without oversampling(it sure wouldn't look like staircases if it did not oversample a sine wave, but it would be ugly).

You're taking that weird filter that does weird things and try to sell a weird story out of it. There is no merit and no truth in doing so.
Fidelity wise, it's a terrible filter, no more, no less.

An audio signal must be band limited when recorded. A band limited Dirac pulse will ring! So the assumption that a straight non ringing pulse response is better for audio fidelity, is false. And that might be the only valid take away from all this. Because the nice audiophile-looking pulse does happen to come out of the signal with the lowest fidelity.

 

[gregorio]

It is not a messed up filter. The NOS filter attempt to faithfully reproduce the digital input (please let me explain later)

These two sentences contradict each other because the whole point of digital audio in the first place is to reproduce/reconstruct the analogue signal, NOT to just output the digital input itself. This is why an anti-image/reconstruction filter is a MANDATED REQUIREMENT of digital audio and why not having a filter (or having a filter which emulates not having a filter) is by definition a broken, “messed up filter”! The second quoted sentence above, along with some of your other statements indicates that you don’t actually know what digital audio is, or what analogue audio is, or that digital audio is NOT analogue audio. In light of your assertions about “critical thinking” and a “good learning technique” that is shocking!!

1. What would you expect to see if we feed the same perfect 10kHz perfect sine wave in 768k / 32 bit format Hi-Res format to the same Topping E30 with F5 (NOS filter)? Would we see a similar stair step output as 44.1k / 16 bit CD format. Would the 10kHz analog "sine" wave output from 768k / 32 bit be "smoother" than the one from CD format?

This is where you’re making the mistake! We would see a stair step output with 768/32 as well and it is therefore broken, any stair step output is broken because again, the whole point of digital audio is to reproduce/reconstruct an analogue signal, NOT to just output the digital signal, hence why an appropriate anti-image/anti-image filter is a requirement. Using no filter (or a deliberately broken filter emulating no filter), the example 10kHz sine wave will be “smoother” with 768/32 than with 44/16 (CD format) but it will be LESS smooth/accurate than 44/16 with appropriate filter!

2. To my understanding, Topping E30 (and similar modern DACs) does have analog filter stage (for all its digital filter settings) but it is not very effective (as shown in the original article) as it is very hard to build a perfect analog filter. This is also the reason why modern DACs use digital filter in addtion to the existing analog filter)

3. Why people still want F5 (NOS filter)?

2. Modern DACs and the DACs in CD players going all the way back to the 1980’s do have a very effective analogue reconstruction filter and it is not very hard to build a (near) perfect analogue filter. So again, your understanding is wrong! What is hard, is to build a very steep analogue filter, which is why even some of the very first generation of CD players oversampled, therefore a digital anti-image filter could be applied which greatly reduced the requirement on the analogue reconstruction filter.
3. Already asked and answered. In short and again, false marketing (inc. Pseudo science)!

Which one would you pick just based on the impluse response? F1? F2? or F5?
(I'll pick F5, not sure about your answer)

4. If the output from a CD format is like what the "Monty's video" suggested (i.e. the DAC will output "perfect sine wave"), which filter in E30 would you consider as properly built, not broken?

5. If F5 is classified as broken or not properly built, how about the other filters (F1, F2, F3, F4, F6)? Are they also broken (for other reasons)? My answer is Yes, all filters are broken if F5 is considered as broken. Agree?

6. As I stated above, the original article "opened my eyes". Could the article open other people's eyes? Well... I am not sure.

I would pick F1.
4. F1, hence why I would pick it.
5. Your answer here is WRONG, it should be “no”, the other filters are not broken, only F5 is broken because it is the only one that does not remove the stair step. The other filter options F2, F3, F4 and F6 are not broken (because they remove the stair steps) but they are not optimal.
6. The original article could indeed “open other people’s eyes” but for some bizarre reason it has closed YOUR eyes, which is even more bizarre as you falsely claim the exact opposite?! The article explains why F5 is broken and why F1 is the best option but you’ve closed your eyes to all of that and chosen the broken one! To continue:

I believe people who have a second thought (or who apply their own critical thinking) would come up with a different answer with the help of the original article, i.e. How to pick the best filter setting for your DAC – Addicted To Audio

Exactly, with the help of that article and applying “their own critical thinking”, people would indeed come up with a different answer (to F5 appearing to be better solely from it’s Dirac impulse response) but even with the help of that article you have still not “come up with a different answer” so from your own words, you have NOT applied your own “critical thinking”!!

Why is this, why do you keep mentioning applying critical thinking but then do not apply it yourself? The most likely reason is either you are just trolling or you have a vested interest in promoting this false narrative/psuedo-science, although it *might* be that you are just very deluded. Eg. You falsely believe your are applying critical thinking, a good learning technique and have had your eyes opened but actually the opposite is true; you are not applying critical thinking, you apparently have no learning technique (let alone a good one) and your eyes are closed to the actual facts/science! That’s why you were censored on ASR and why you will have a rough time here as well, but I suspect you already know that!

G

 

[gregorio]

@castleofargh you posted your response while I was writing mine, I might not have bothered if I’d seen it.

case one: 1536k Hz digitized input of an analog audio input signal, play back with E30, NOS filter
case two: 44.1k Hz digitized input of the same analog audio input signal , play back with E30 and the filter you considered as the best amongst all 6 filters available
For the audio signal output, which output would be closer to the original analog audio input?

If you have a good learning technique and apply critical thinking to the original article (that should “open people’s eyes”), you would be able to work out for yourself that the answer to your question is “case two”!

G

 

[sunjam]

A Dirac pulse is not music! There is a reason why we often talk about a Dirac pulse as an illegal signal. The concept is infinitely high amplitude at instant t, and nothing before or after. In the real world with sound, that would require the power of teleportation, and even that would only solve half, the appearing or disappearing, not both consecutively at the same place and basically the same time.
Any decent DAC can and will filter the signal in a way that shows no staircases!

An actual filterless NOS DAC, if it could do a perfect job(not physically possible), would give, not a spike but a lonely tower looking signal(like a portion of a square wave).
A real one would show ringing at the base and at the top of that "tower" on both sides. And it would indeed show staircases. It would have low fidelity because obviously those staircases shouldn't be there. They're like very many ultrasonic frequencies that didn't exist in the band limited record.
A NOS DAC with a filter designed by someone with a brain would try to remove as much aliasing as possible by attenuating the ultrasonic content. But doing that well with an analog filter is expensive and still nowhere near perfect. That's why I said that many of those designs have audibly rolled off treble, because they start filtering earlier to achiever a higher attenuation in the ultrasounds.
The only bonus of that design is usually low noise floor. The rest isn't great(fidelity wise) compared to delta sigma and a mix of digital and analog filter.

That F5 filter is indeed not that, the DAC is not a R2R DAC in the first place, and delta sigma just doesn't work without oversampling(it sure wouldn't look like staircases if it did not oversample a sine wave, but it would be ugly).

You're taking that weird filter that does weird things and try to sell a weird story out of it. There is no merit and no truth in doing so.
Fidelity wise, it's a terrible filter, no more, no less.

An audio signal must be band limited when recorded. A band limited Dirac pulse will ring! So the assumption that a straight non ringing pulse response is better for audio fidelity, is false. And that might be the only valid take away from all this. Because the nice audiophile-looking pulse does happen to come out of the signal with the lowest fidelity.

1. Agree, a true Dirac pulse is not music. It is an extreme case for theorectical analysis. How about a physical impulse (not Dirac)? I think all the filter analysis from the original article still applies. Correct?

2. Agree, a decent DAC can have filter to "fix" the staircase output. However, a decent DAC can also have filter to "faithfully" reproduct the digital input, i.e. a NOS mode that could produce staircase output, with legitimate reason (as I mentioned for the high bitrate case).

3. "A NOS DAC with a filter designed by someone with a brain would try to remove as much aliasing as possible by attenuating the ultrasonic content. But doing that well with an analog filter is expensive and still nowhere near perfect. That's why I said that many of those designs have audibly rolled off treble, because they start filtering earlier to achiever a higher attenuation in the ultrasounds." <=== agreed, that's why modern DAC use both digital and analog filter.
4. Why we stuck on filterless NOS DAC? Could we focus on E30 with F-5? The real modern DAC we are talking about, i.e. a modern DAC with a digital filter F-5 (that emulate "NOS") and analog filter

5. "You're taking that weird filter that does weird things and try to sell a weird story out of it. There is no merit and no truth in doing so." <=== sorry, I cannot agree. I already let you know the real world use case, i.e. Hi-Res playback with NOS filter vs CD playback wth other filter. If you consider Hi-Res playback is a weird thing to do then we would have a different defintion of "weird things"

6. "An audio signal must be band limited when recorded. A band limited Dirac pulse will ring! So the assumption that a straight non ringing pulse response is better for audio fidelity, is false" <=== agreed. My previous reply with different graphs is to show that why F-5 exists (as it has the least ringing impulse response). I did not say it is the best filter for any music (if I did, that was my mistake). As I highlighted many times, there is a legimate reason for F-5. May be it is not for the general public who play CD-quality music. IMO, there is no perfect filter. It is always a compromise <=== this is the area that the "Monty's video" try to avoid to mention.

 

[sunjam]

These two sentences contradict each other because the whole point of digital audio in the first place is to reproduce/reconstruct the analogue signal, NOT to just output the digital input itself. This is why an anti-image/reconstruction filter is a MANDATED REQUIREMENT of digital audio and why not having a filter (or having a filter which emulates not having a filter) is by definition a broken, “messed up filter”! The second quoted sentence above, along with some of your other statements indicates that you don’t actually know what digital audio is, or what analogue audio is, or that digital audio is NOT analogue audio. In light of your assertions about “critical thinking” and a “good learning technique” that is shocking!!

This is where you’re making the mistake! We would see a stair step output with 768/32 as well and it is therefore broken, any stair step output is broken because again, the whole point of digital audio is to reproduce/reconstruct an analogue signal, NOT to just output the digital signal, hence why an appropriate anti-image/anti-image filter is a requirement. Using no filter (or a deliberately broken filter emulating no filter), the example 10kHz sine wave will be “smoother” with 768/32 than with 44/16 (CD format) but it will be LESS smooth/accurate than 44/16 with appropriate filter!

2. Modern DACs and the DACs in CD players going all the way back to the 1980’s do have a very effective analogue reconstruction filter and it is not very hard to build a (near) perfect analogue filter. So again, your understanding is wrong! What is hard, is to build a very steep analogue filter, which is why even some of the very first generation of CD players oversampled, therefore a digital anti-image filter could be applied which greatly reduced the requirement on the analogue reconstruction filter.
3. Already asked and answered. In short and again, false marketing (inc. Pseudo science)!

I would pick F1.
4. F1, hence why I would pick it.
5. Your answer here is WRONG, it should be “no”, the other filters are not broken, only F5 is broken because it is the only one that does not remove the stair step. The other filter options F2, F3, F4 and F6 are not broken (because they remove the stair steps) but they are not optimal.
6. The original article could indeed “open other people’s eyes” but for some bizarre reason it has closed YOUR eyes, which is even more bizarre as you falsely claim the exact opposite?! The article explains why F5 is broken and why F1 is the best option but you’ve closed your eyes to all of that and chosen the broken one! To continue:

Exactly, with the help of that article and applying “their own critical thinking”, people would indeed come up with a different answer (to F5 appearing to be better solely from it’s Dirac impulse response) but even with the help of that article you have still not “come up with a different answer” so from your own words, you have NOT applied your own “critical thinking”!!

Why is this, why do you keep mentioning applying critical thinking but then do not apply it yourself? The most likely reason is either you are just trolling or you have a vested interest in promoting this false narrative/psuedo-science, although it *might* be that you are just very deluded. Eg. You falsely believe your are applying critical thinking, a good learning technique and have had your eyes opened but actually the opposite is true; you are not applying critical thinking, you apparently have no learning technique (let alone a good one) and your eyes are closed to the actual facts/science! That’s why you were censored on ASR and why you will have a rough time here as well, but I suspect you already know that!

G

Thanks for your reply. It looks to me you may have a differet defintion of "critical thinking". Based on wikipedia:

Are you talking the same thing here?

May I ask you a simple question: "Hi-Res is useless. Do you agree?"