[gregorio]

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

Maybe we do have a “different definition of critical thinking”. My definition is as you have quoted from Wikipedia, however, yours is different because you have not formed a rational judgement according to the information/facts (for example, the original article you linked to) you’ve actually formed an irrational judgement that’s opposite to the information you linked to!

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

No, I do not agree. Hi-Res has attributes that are useful in certain applications and therefore by definition, it is not useless. The only useful attribute it has as far as consumer audio reproduction is concerned is in audiophile marketing, so it’s not useless to audiophile marketers but is useless to consumers or arguably worse than useless as it can be used to mislead them.

G

 

[sunjam]

Maybe we do have a “different definition of critical thinking”. My definition is as you have quoted from Wikipedia, however, yours is different because you have not formed a rational judgement according to the information/facts (for example, the original article you linked to) you’ve actually formed an irrational judgement that’s opposite to the information you linked to!

No, I do not agree. Hi-Res has attributes that are useful in certain applications and therefore by definition, it is not useless. The only useful attribute it has as far as consumer audio reproduction is concerned is in audiophile marketing, so it’s not useless to audiophile marketers but is useless to consumers or arguably worse than useless as it can be used to mislead them.

G

Cool, thanks for your reply.

Please sorry for my poor communication skill, I am confused about the red part in your reply.

1. Do you mean "Hi-Res is useless to consumers (other than those are classified as audiophile)?
2. What do you mean by misleading them? Could you give some concrete examples?
3. What are the difference between audiophile and non-audiophile in your mind?

Cheers.

 

[gregorio]

I am confused about the red part in your reply.
1. Do you mean "Hi-Res is useless to consumers (other than those are classified as audiophile)?
2. What do you mean by misleading them? Could you give some concrete examples?
3. What are the difference between audiophile and non-audiophile in your mind?

1. No, audiophiles are also consumers of course. So, I meant hi-res is useless to both audiophiles and other consumers.
2. The very name “Hi-res” indicates “high resolution” but within the audible range it has the same resolution as standard resolution (44/16) and therefore is obviously misleading. There are other examples but none more obvious than that!
3. An audiophile is literally someone who is particularly enthusiastic about high fidelity audio reproduction. However, these days it’s a bit difficult to know because some of the most ardent/hardcore audiophiles actually seem particularly enthusiastic about low fidelity audio reproduction (typically due to being mislead/deluded by audiophile marketing). Hence why there is some demand for NOS DACs amongst these mislead hardcore audiophiles.

BTW, just out of curiosity; as your definition of “critical thinking” is obviously not the one you quoted from Wikipedia, what is your definition?

G

 

[sunjam]

1. No, audiophiles are also consumers of course. So, I meant hi-res is useless to both audiophiles and other consumers.
2. The very name “Hi-res” indicates “high resolution” but within the audible range it has the same resolution as standard resolution (44/16) and therefore is obviously misleading. There are other examples but none more obvious than that!
3. An audiophile is literally someone who is particularly enthusiastic about high fidelity audio reproduction. However, these days it’s a bit difficult to know because some of the most ardent/hardcore audiophiles actually seem particularly enthusiastic about low fidelity audio reproduction (typically due to being mislead/deluded by audiophile marketing). Hence why there is some demand for NOS DACs amongst these mislead hardcore audiophiles.

BTW, just out of curiosity; as your definition of “critical thinking” is obviously not the one you quoted from Wikipedia, what is your definition?

G

Thanks again for your reply.

Sorry for my poor communication skill again, in order to improve my understanding, please let me define something first:

original audio signal <=== Let's call it A_original
digitized input from A_original <=== Lets' call it D. i.e. D=ADC(A_original) where ADC( ) is the ADC process
reconstructed audio signal from a DAC <=== Let's call it A_from_DAC, i.e. A_from_DAC=DAC(D) whre DAC( ) is the DAC process

high fidelity audio audiophiles <=== people who wants output A_from_DAC to be very close to A_original. The closer the better. Ideally, A_from_DAC = A_original
<=== Let's call this group of people as HFA

low fidelity audio audiophiles <=== people who wants output_A_from_DAC sounds good to their ears.
<=== Let's call this group of people as LFA

A generic customer are the rest of the people <== Let's call this group of people as G

Let's be specific about Hi-Res. Let's use DXD (352.8kHz / 24bit) as an example for our discussion.


Now, go back to your reply. You said:

Hi-Res has attributes that are useful in certain applications and therefore by definition, it is not useless

Could you share which applications that you are referring to?

No, audiophiles are also consumers of course. So, I meant hi-res is useless to both audiophiles and other consumers.

It means Hi-Res (DXD) is useless to G and HFA. Correct?

The only useful attribute it has as far as consumer audio reproduction is concerned is in audiophile marketing, so it’s not useless to audiophile marketers

Ok, got it. You mean Hi-Res is only useful for the Hi-Res seller

but is useless to consumers or

You mean Hi-Res is useless to G. Correct?

arguably worse than useless as it can be used to mislead them.

You mean DXD is worse than CD. Correct?

The very name “Hi-res” indicates “high resolution” but within the audible range it has the same resolution as standard resolution (44/16)

In the ADC process, you would not consider using 352,800 sample points per second (aka DXD) for sampling an audio signal as higher resolution than using 44,100 sample points per second (aka CD)?

p.s.: My critical thinking is same as the one defined in wiki

 

[gregorio]

Could you share which applications that you are referring to?

Almost no applications if you’re going to define hi-res as 384/24, maybe it has some scientific applications and when recording certain animals which produce ultrasound (maybe for echo location). However, hi-res is typically defined as any sample rate higher than 44.1kHz or 48kHz and more than 16bit, so 96/24 for example. Using this definition of hi-res there are more applications, 24bits allows for considerably more headroom when recording audio and is therefore very useful, although it’s useless for playback because almost no headroom is required. A sample rate of 96kHz has little application in audio with two exceptions: Sound design, where it is sometimes necessary to pitch a recording down an octave or more and in localised processing (plugins) where audio freqs higher than 24kHz are sometimes necessary, for example when emulating certain vintage compressors or limiters. Higher than 96kHz has no applications in audio content creation, except for marketing of course.

I am confused here. Do you mean Hi-Res (DXD) is useful to LFA? or HFA? Which type of audiophile you are referring to?
You mean Hi-Res is useless to G. Correct?
You mean DXD is worse than CD. Correct?

DXD is useless to everyone involved in audio, both the content creators (audio engineers) and consumers (G, LFA and HFA). It’s only useful as a marketing gimmick, hi-res can be falsely marketed as high resolution and distributors can charge more for it because some consumers who believe that false marketing will pay more for it.

In the ADC process, you would not consider using 352,800 sample points per second (aka DXD) for sampling an audio signal as higher resolution than using 44,100 sample points per second (aka CD)?

No, I would not “consider using 352,800 sample points a second” in the ADC process for several reasons, not least because it’s not possible! My first professional ADC used a 2.8MHz sampling rate for the initial ADC process but that was over 30 years ago, modern ADCs use 11.2MHz or 22.4MHz, I don’t know of any pro ADCs in the last 30 years that have used only a 352.8KHz sample rate. However, that is only for the initial ADC process, the secondary process (“decimation”) I would only consider 44/24, 48/24 or for sound design where pitch shifting is anticipated then 96/24. I would never consider 192KHz or higher sample rates unless a client specifically requested it (for marketing purposes), because it is NOT higher resolution, it is exactly the same resolution just with an unnecessarily wider bandwidth.

p.s.: My critical thinking is same as the one defined in wiki

Then why do you keep mentioning the application of critical thinking but then make false assertions which demonstrate you are NOT applying critical thinking?

For example, your questions I’ve answered above indicate you do not really understand sample rates, bit depths, the ADC process or resolution but the application of critical thinking (as defined by Wikipedia) dictates that you must “analyse the facts”, just making up assertions based on ignorance, on not understanding/analysing the facts is therefore NOT critical thinking!

G

 

[sunjam]

Almost no applications if you’re going to define hi-res as 384/24, maybe it has some scientific applications and when recording certain animals which produce ultrasound (maybe for echo location). However, hi-res is typically defined as any sample rate higher than 44.1kHz or 48kHz and more than 16bit, so 96/24 for example. Using this definition of hi-res there are more applications, 24bits allows for considerably more headroom when recording audio and is therefore very useful, although it’s useless for playback because almost no headroom is required. A sample rate of 96kHz has little application in audio with two exceptions: Sound design, where it is sometimes necessary to pitch a recording down an octave or more and in localised processing (plugins) where audio freqs higher than 24kHz are sometimes necessary, for example when emulating certain vintage compressors or limiters. Higher than 96kHz has no applications in audio content creation, except for marketing of course.

DXD is useless to everyone involved in audio, both the content creators (audio engineers) and consumers (G, LFA and HFA). It’s only useful as a marketing gimmick, hi-res can be falsely marketed as high resolution and distributors can charge more for it because some consumers who believe that false marketing will pay more for it.

No, I would not “consider using 352,800 sample points a second” in the ADC process for several reasons, not least because it’s not possible! My first professional ADC used a 2.8MHz sampling rate for the initial ADC process but that was over 30 years ago, modern ADCs use 11.2MHz or 22.4MHz, I don’t know of any pro ADCs in the last 30 years that have used only a 352.8KHz sample rate. However, that is only for the initial ADC process, the secondary process (“decimation”) I would only consider 44/24, 48/24 or for sound design where pitch shifting is anticipated then 96/24. I would never consider 192KHz or higher sample rates unless a client specifically requested it (for marketing purposes), because it is NOT higher resolution, it is exactly the same resolution just with an unnecessarily wider bandwidth.

Then why do you keep mentioning the application of critical thinking but then make false assertions which demonstrate you are NOT applying critical thinking?

For example, your questions I’ve answered above indicate you do not really understand sample rates, bit depths, the ADC process or resolution but the application of critical thinking (as defined by Wikipedia) dictates that you must “analyse the facts”, just making up assertions based on ignorance, on not understanding/analysing the facts is therefore NOT critical thinking!

G

Cools, thanks for your reply. You really provide a lot of points and examples. Thank you again.

Based on your reply, looks like that, from your understanding, 44.1kHz is perfect for audio signal (up to 22kHz). So oversampling, like Hi-Res (DXD) provides no addtional benefits for recording audio signal. Correct?

Do you agree HiRes (96/24) would have less quantization noise than 44/16?

For the 2.8Mhz, 11.2MHz or 22.4MHz sampling rate in the ADC process that you mentioned above, may I ask which format they use for recording? Is it same as DXD recording format but just with a higher bitrate?

For the 2.8Mhz, 11.2MHz or 22.4MHz sampling rate in the ADC process you mentioned above, what is the bitdepth they use?

p.s. For the critical thinking part, it looks to me that we have different definition. Feel free to use any description/name to describe my thinking style. I would not argue with you. If you ask me what style I am using, I would answer "critical thinking". For this, you don't have to agree. Feel free to disagree.

 

[theveterans]

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?

Case one: on NOS filter, you won't be able to hear the ultrasonics but definitely much better IMD and THD performance with oversampling since now the aliasing is pushed far, far away from passband, impulse still the same (depends on jitter performance as well in addition). FR should be flatter and almost indistinguishable to FIR filter applied to 44.1 KHz input and no way in hell to pass DBT ABX test with confidence.
Case two: impulse response of F1 (linear phase sharp roll-off) will result pre and post ringing, and again pre and post ringing are debatable (see GoldenSound passing ABX with different DAC filter), but in 99.9% of cases, indistinguishable to NOS with 1536KHz sampling

You're telling me though that the case for Hi-Res is valid ONLY if you're using NOS filter and oversampling outside of DAC such as HQPlayer. NOS on a 44.1 KHz music is audibly different than OS using a decent headphones or speakers though if you want that answer as well, however, as you age, you lose that ability to hear the higher frequencies as well

Just to add: competent DACs oversample by default though. NOS on 44.1 KHz is a fun sound and can tame those bright 80's music especially recordings from "The Cure" or Billy Idol

 

[sunjam]

Case one: on NOS filter, you won't be able to hear the ultrasonics but definitely much better IMD and THD performance with oversampling since now the aliasing is pushed far, far away from passband, impulse still the same (depends on jitter performance as well in addition). FR should be flatter and almost indistinguishable to FIR filter applied to 44.1 KHz input and no way in hell to pass DBT ABX test with confidence.
Case two: impulse response of F1 (linear phase sharp roll-off) will result pre and post ringing, and again pre and post ringing are debatable (see GoldenSound passing ABX with different DAC filter), but in 99.9% of cases, indistinguishable to NOS with 1536KHz sampling

You're telling me though that the case for Hi-Res is valid ONLY if you're using NOS filter and oversampling outside of DAC such as HQPlayer. NOS on a 44.1 KHz music is audibly different than OS using a decent headphones or speakers though if you want that answer as well, however, as you age, you lose that ability to hear the higher frequencies as well

Just to add: competent DACs oversample by default though. NOS on 44.1 KHz is a fun sound and can tame those bright 80's music especially recordings from "The Cure" or Billy Idol

Cool.

No, I am not saying that Hi-Res is valid ONLY as you described.

If you look at my definitions in my previous reply, I would say

Given the same A_original, then

ADC_HiRes (A_original) => D_Hi_Res
ADC_CD(A_original) => D_CD

DAC(D_HiRes) => A_from_DAC_HiRes
DAC(D_CD) => A_from_DAC_CD

For any given A_original,
A_from_DAC_HiRes would be closer to A_original than A_from_DAC_CD (if both are using the "best" filter) <=== statement 1

Using DAC_F-1( ) or DAC_F-5( ) is your option

In short,
there is legitimate reason for F-5 (that would reconstruct stair-step output form for D_CD_1kHz), and
A_from_DAC_CD is not perfect (same as A_from_DAC_HiRes) (as the Monty's video suggested that A_from_DAC_CD is perfect)

Please let me know if the above is not clear.

Yes, pretty much all modern DAC has internal OS (as most are using DSM). That's why people would not be able to see stair step easily even for 44.1/16. However, this doesn't mean the "NOS mode" is broken or useless.

For people who do external OS would like to have NOS mode as they don't want any addtional DSP from the DAC.

 

[gregorio]

Based on your reply, looks like that, from your understanding, 44.1kHz is perfect for audio signal (up to 22kHz).

No. Based on your reply it looks like you are again not applying critical thinking! I did not state or imply 44.1kHz is perfect up to 22kHz. It would depend on the application and the bit depth and is not perfect but it can exceed the abilities of human hearing.

Do you agree HiRes (96/24) would have less quantization noise than 44/16?

It would have less inaudible quantisation noise but as the quantisation noise of both 96/24 and 44/16 is inaudible then obviously 96/24 provides no audible benefit for playback at reasonable listening levels.

So oversampling, like Hi-Res (DXD) provides no addtional benefits for recording audio signal. Correct?

No that is not correct, I don’t know where you got that from? And, the oversampling process in ADCs is NOT like DXD. Again, ADCs have all been oversampling for well over 30 years, it does provide a benefit because it relaxes the requirement of the analogue anti-alias filter. You don’t seem to know how ADCs work, that the initial sampling rate is not the recorded audio signal.

For the 2.8Mhz, 11.2MHz or 22.4MHz sampling rate in the ADC process that you mentioned above, may I ask which format they use for recording? Is it same as DXD recording format but just with a higher bitrate?

2.8MHz is no longer available, the last 64x oversampling ADC I can remember was about 20 or more years ago, although it was used as a 1bit recording format (for SACD). The more modern initial sampling rates (EG. 11.2MHz) are not a recording format, they are not even output by the ADC, those rates are only for initial conversion to digital, thereafter the digital signal is “decimated” to the required format (48/24, 44/16, 352/24 or whatever) before it is output. The initial sample rate is not the same as DXD, it uses a much higher sample rate and far fewer bits.

For the 2.8Mhz, 11.2MHz or 22.4MHz sampling rate in the ADC process you mentioned above, what is the bitdepth they use?

I believe for that 31 year old 64x oversampling ADC it was 4bit but it might have been 3bit. Today with say 11.2MHz sampling I would guess it’s probably around 6bits but that’s not something the manufacturers usually make public as far as I’m aware.

If you ask me what style I am using, I would answer "critical thinking".

Obviously you would and indeed you have, numerous times. What is bizarre though is that you state you’re using the Wikipedia definition but don’t, you’re using some completely different definition and falsely stating you’re following the Wiki one, as demonstrated/explained in my previous post. Is it really possible you just don’t realise you’re making a false statement and somehow truly believe you’re following the wiki definition? I’m not sure it is possible.

G

 

[sunjam]

No. Based on your reply it looks like you are again not applying critical thinking! I did not state or imply 44.1kHz is perfect up to 22kHz. It would depend on the application and the bit depth and is not perfect but it can exceed the abilities of human hearing.

It would have less inaudible quantisation noise but as the quantisation noise of both 96/24 and 44/16 is inaudible then obviously 96/24 provides no audible benefit for playback at reasonable listening levels.

No that is not correct, I don’t know where you got that from? And, the oversampling process in ADCs is NOT like DXD. Again, ADCs have all been oversampling for well over 30 years, it does provide a benefit because it relaxes the requirement of the analogue anti-alias filter. You don’t seem to know how ADCs work, that the initial sampling rate is not the recorded audio signal.

2.8MHz is no longer available, the last 64x oversampling ADC I can remember was about 20 or more years ago, although it was used as a 1bit recording format (for SACD). The more modern initial sampling rates (EG. 11.2MHz) are not a recording format, they are not even output by the ADC, those rates are only for initial conversion to digital, thereafter the digital signal is “decimated” to the required format (48/24, 44/16, 352/24 or whatever) before it is output. The initial sample rate is not the same as DXD, it uses a much higher sample rate and far fewer bits.

I believe for that 31 year old 64x oversampling ADC it was 4bit but it might have been 3bit. Today with say 11.2MHz sampling I would guess it’s probably around 6bits but that’s not something the manufacturers usually make public as far as I’m aware.

Obviously you would and indeed you have, numerous times. What is bizarre though is that you state you’re using the Wikipedia definition but don’t, you’re using some completely different definition and falsely stating you’re following the Wiki one, as demonstrated/explained in my previous post. Is it really possible you just don’t realise you’re making a false statement and somehow truly believe you’re following the wiki definition? I’m not sure it is possible.

G

'2.8MHz is no longer available, the last 64x oversampling ADC' <== this is known as DSD64
11.2Mhz is DSD256 (256x oversampling) <=== this is the format of the music I am listening now.

Question 1: Do you consider DSD256 a Hi-Res format?
Question 2: If yes, do you think DSD256 would help to reconstruct better audio signal from the "decimated" 44/16 format?

If you answer Yes to both questions, then it means you agree Hi-Res is not useless for reconstructing audio signal

 

[gregorio]

In short,
there is legitimate reason for F-5 (that would reconstruct stair-step output form for D_CD_1kHz), and
A_from_DAC_CD is not perfect (same as A_from_DAC_HiRes) (as the Monty's video suggested that A_from_DAC_CD is perfect)

Please let me know if the above is not clear.

No, your post was not really very clear but it appears to be based on marketing BS and not the actual facts. Certainly the “In short” section I’ve quoted is WRONG! Just outputting the sample values from a CD is absolutely wrong and therefore there is NO legitimate reason for the F5 option. You don’t seem to understand the difference between analogue and digital, you cannot just output digital, it has to be converted and the analogue signal reconstructed! How can you not know this basic fact?

2.8MHz is no longer available, the last 64x oversampling ADC' <== this is known as DSD64

No it is not! What definition of “critical thinking” are you using that allows you to simply make-up falsehoods and assert them as fact, it most definitely is NOT the definition you quoted from Wikipedia!!

Unfortunately, despite your claim of “a good learning technique”, you’re not even demonstrating a poor learning technique, you’re just repeating the same falsehoods you started with and have therefore demonstrated absolutely no learning technique whatsoever! Maybe it’s the same as your “critical thinking”, you claim to do it, provide a reliable definition of it but what you actually demonstrate is pretty much exact opposite! What rational conclusion do you expect we can draw from this?

G

 

[sander99]

If you answer Yes to both questions, then it means you agree Hi-Res is not useless for reconstructing audio signal

Now we can come full circle:
Hi-res is not the correct word. But closely related to your question: oversampling plus digital filtering is very usefull for reconstructing the audio signal (because a simpler analog filter can be used), and that is why NOS DACs are obsolete!
44.1/16 is audibly transparent (audibly perfect) so it is good enough as a distribution and playback format, and the easiest way to convert it to analog in an audibly transparent way is by using oversampling plus digital filtering plus a final simple analog filter.

 

[sunjam]

Now we can come full circle:
Hi-res is not the correct word. But closely related to your question: oversampling plus digital filtering is very usefull for reconstructing the audio signal (because a simpler analog filter can be used), and that is why NOS DACs are obsolete!
44.1/16 is audibly transparent (audibly perfect) so it is good enough as a distribution and playback format, and the easiest way to convert it to analog in an audibly transparent way is by using oversampling plus digital filtering plus a final simple analog filter.

Sorry for my ignorance, feel free to think the way you like as I cannot understand what you are talking about (as you cannot understand what I am talking about).
Enjoy

 

[gregorio]

Sorry for my ignorance

Mmmm, I don’t think you are. If you were sorry for your ignorance you would try to apply a “learning technique” to do something about it but you don’t, you just keep repeating the same falsehoods.

I cannot understand what you are talking about

Really? It was about as simple as it can get.

as you cannot understand what I am talking about

Ah but we do understand what you are talking about, it’s hardly original, it’s the same thing that marketers, shills and trolls have been talking about for well over a decade, as well as the occasional genuine person who’s just been suckered by the marketing. Claiming to be applying “critical thinking”, a “good learning technique” and to be anti-pseudoscience but then actually demonstrating the exact opposite would tend to indicate that we’re not dealing with the latter.

G

 

[sunjam]

Mmmm, I don’t think you are. If you were sorry for your ignorance you would try to apply a “learning technique” to do something about it but you don’t, you just keep repeating the same falsehoods.

Really? It was about as simple as it can get.

Ah but we do understand what you are talking about, it’s hardly original, it’s the same thing that marketers, shills and trolls have been talking about for well over a decade, as well as the occasional genuine person who’s just been suckered by the marketing. Claiming to be applying “critical thinking”, a “good learning technique” and to be anti-pseudoscience but then actually demonstrating the exact opposite would tend to indicate that we’re not dealing with the latter.

G

Feel free to enjoy what's in your mind. Cheers.

By the way, feel free not to answer the following question:

Question 1: Do you consider DSD256 a Hi-Res format?
Question 2: If yes, do you think DSD256 would help to reconstruct better audio signal from the "decimated" 44/16 format?