Hi-resolution versus 16 bit 44.1 Khz

[andyr]

14 minutes ago, March Audio said:

If there is some technical point you don't believe or understand in the video, just say and we can discuss.

 

I've already said, Alan ... I'm not interested in watching long videos when I can speed-read text; so I'm afraid I won't be watching the vid.

 

[MLXXX]

45 minutes ago, andyr said:

Re. this so-called 'digital myth' Alan ... it seems to me that a similar concept is the sound of 45rpm LPs vs. the standard 33rpm LPs.

 

It may seem that way intuitively, but the effect on audible sound quality between a sampling rate of 44.1kHz and 48kHz  is more like the difference between running a light bulb on 60Hz mains rather than 50Hz mains, i.e. no noticeable difference.

[March Audio]

1 hour ago, andyr said:

 

I've already said, Alan ... I'm not interested in watching long videos when I can speed-read text; so I'm afraid I won't be watching the vid.

 

Ignorance is bliss eh Andy.

 

It's hardly long.

 

So let's get this straight.  You were criticising the content of a video you haven't even watched.

 

Wow.

Edited January 18, 2023 by March Audio

[March Audio]

13 hours ago, andyr said:

 

Re. this so-called 'digital myth' Alan ... it seems to me that a similar concept is the sound of 45rpm LPs vs. the standard 33rpm LPs.

 

I happen to have an LP which has just one track on each side - one side to be played at 33.333rpm ... the other at 45rpm.  It is quite obvious that the SQ of the 45rpm side is better than the that of the 33rpm side.

 

To me ... 45rpm means "a higher sampling rate".

 

Sure, I have lots of other LPs where I have both a 33rpm copy and a 45rpm copy - but if I said with these, the 45rpm version sounded better ... you may well retort that it might be a different recording / mastering.  Whereas this one LP is the same version - simply with 2 different play speeds.

 

 

Again an intuitive error. The two are not comparable in the way you have.

 

On an LP the same mechanical information is spread over a longer linear track at higher rpm.  Easier to cut, easier to read.  Also 12" 45 rpm records are mastered differently.  Wider groove spacing, greater depth etc. Literally less mechanical limitations.

 

With sampling, as I already explained, we already have all the information contained within our hearing bandwidth at a sample rate of 44kHz.  Going faster only captures higher frequency signals.  It does not capture the signals under 22kHz "better".

 

Honestly, watch the video.  You will learn a lot.

Edited January 19, 2023 by March Audio

[davewantsmoore]

17 hours ago, stereo coffee said:

Certainly starting, but comparison between a hi res player with good power supplies and a good plain 16 bit player

If you do this test .... and you take the "high res" file, and you resample it carefully to 16/44, and play it on the "16bit player"... and they players are both very good quality.   They will sound the same.

 

17 hours ago, Dingbat said:

I think the idea behind sampling rates higher than 20kHz is not that it's beyond human hearing, it's the interpolation in the digital signal.

This is somewhat moot.

 

To do high quality interpolation does not in any way require that we either feed DACs with a high rate input (because it can be oversampled internally) or that we distribute audio in high rates.

 

17 hours ago, Dingbat said:

It might give some intangible benefit

It certainly can depending on the DAC

[davewantsmoore]

11 hours ago, March Audio said:

It is a common misconception; that we can miss important audible signal "jiggles" in between the sample points. 

 

Yes.   If there is any "jiggles" (this is an official technical term now) .... they are the higher frequencies, which are purposefully not being stored.

 

11 hours ago, March Audio said:

Therefore by sampling faster we can capture them and improve audio quality. Unfortunately, whilst that seems intuitively correct, it's not.

Yes... if we sample faster we can capture the higher frequencies.  

[bob_m_54]

12 hours ago, andyr said:

To me ... 45rpm means "a higher sampling rate".

By definition, analog signals don't have a sample rate.

[andyr]

7 minutes ago, bob_m_54 said:

By definition, analog signals don't have a sample rate.

 

Apologies, Bob - I thought you were an electronics tech ... not an English teacher!  

 

What I should've said was "To me ... 45rpm vs 33rpm is the analogue equivalent of "a higher sampling rate".

 

[davewantsmoore]

12 hours ago, andyr said:

 

Re. this so-called 'digital myth' Alan ... it seems to me that a similar concept is the sound of 45rpm LPs vs. the standard 33rpm LPs.

 

I happen to have an LP which has just one track on each side - one side to be played at 33.333rpm ... the other at 45rpm.  It is quite obvious that the SQ of the 45rpm side is better than the that of the 33rpm side.

 

To me ... 45rpm means "a higher sampling rate".

 

Higher sampling rate (in digital audio) means that the audio contains (or can contain) higher frequencies.

 

It doesn't mean higher sound quality, or higher precision, or higher anything else.

 

I think you should reconsider watching the video and trying to understand it.... if you care about understanding digital audio.

[davewantsmoore]

13 minutes ago, andyr said:

What I should've said was "To me ... 45rpm vs 33rpm is the analogue equivalent of "a higher sampling rate".

So you are saying that 45 rpm records (can) contain higher frequencies than 33 rpm records ?!

 

Edited January 19, 2023 by davewantsmoore

[andyr]

4 minutes ago, davewantsmoore said:

So you are saying that 45 rpm records (can) contain higher frequencies than 33 rpm records ?!

 

I don't think I said that at all, Dave?

 

What I'm saying is that when you have the same track cut to be played back at 45rpm and also 33rpm ... the faster rotational speed makes it sound better.

 

Which is similar to what I understand the devotees of high-res digital playback say - the higher resolution sounds better.  (But as I am a newbie, in terms of digital sound ... I will understand if you say that is actually not what devotees of high-res digital playback say.)

 

[davewantsmoore]

11 minutes ago, andyr said:

I don't think I said that at all, Dave?

Only you can tell me what you mean (I'm not trying to put word in your mouth, heh).

 

.... high(er) sampling rates allow for high(er) frequencies to be stored.   Nothing more, nothing less.

So, I figured you meant that, but for record players.

 

11 minutes ago, andyr said:

What I'm saying is that when you have the same track cut to be played back at 45rpm and also 33rpm ... the faster rotational speed makes it sound better.

Sure, I've heard a lot of people say that.... it's the analogy to sampling rate, which is problematic.

 

11 minutes ago, andyr said:

Which is similar to what I understand the devotees of high-res digital playback say - the higher resolution sounds better.  (But as I am a newbie, in terms of digital sound ... I will understand if you say that is actually not what devotees of high-res digital playback say.)

There are a lot of people who say that.   They are wrong.

 

Sure, the thing which they are listening to, might sound better than some other thing ..... but their "reason why" (the higher sampling rate) is incorrect.

[andyr]

3 minutes ago, davewantsmoore said:

Sure, the thing which they are listening to, might sound better than some other thing ..... but their "reason why" (the higher sampling rate) is incorrect.

 

Understood.  

 

[Ars Paart]

It could be that with specific DAC implementations the higher (than CD) res could sound better or at least different. For instance, there could be no oversampling implemented which may cause the reconstruction filters to crud the audible band which may be audible.  But that is down to poorly designed/built DACs rather than anything inherently superior with higher resolution digital audio. 

 

*Brain not working properly. Of course NOS dacs don't have digital filters.

Edited January 19, 2023 by Ars Paart

[bob_m_54]

9 hours ago, andyr said:

 

Apologies, Bob - I thought you were an electronics tech ... not an English teacher!  

 

What I should've said was "To me ... 45rpm vs 33rpm is the analogue equivalent of "a higher sampling rate".

 

Yeah.. nah.. . There is no analogy you can make between the two formats really, that I can think of.

 

 

[andyr]

1 hour ago, bob_m_54 said:

Yeah.. nah.. . There is no analogy you can make between the two formats really, that I can think of.

 

OK, so why does faster speed sound better?

 

 

[sir sanders zingmore]

31 minutes ago, andyr said:

 

OK, so why does faster speed sound better?

 

 

It's a great question but the subject for a separate thread perhaps?

[acg]

36 minutes ago, andyr said:

 

OK, so why does faster speed sound better?

 

 

 

Probably something to do with tracking...hills and hollows are about one-third longer presenting an easier task for the stylus.

[davewantsmoore]

57 minutes ago, andyr said:

OK, so why does faster speed sound better?

Well, it isn't anything to do with higher sampling rates in digital audio.... I can't tell you that.   

[Cloth Ears]

12 hours ago, andyr said:

 

OK, so why does faster speed sound better?

 

 

 

11 hours ago, acg said:

 

Probably something to do with tracking...hills and hollows are about one-third longer presenting an easier task for the stylus.

Probably for the same reason that a faster tape speed provides more information when recording - so 30ips has more info than 15ips which has more info than 7.5ips. There's simply more information there (whether it's better or not depends on how it's recorded and what's done with it afterwards).

[andyr]

12 minutes ago, Cloth Ears said:

Probably for the same reason that a faster tape speed provides more information when recording - so 30ips has more info than 15ips which has more info than 7.5ips. There's simply more information there (whether it's better or not depends on how it's recorded and what's done with it afterwards).

 

Zackly!  

 

Which is why I can't help thinking that the "more information" that's available from higher sampling rates in a digital recording ... similarly delivers a benefit.

 

[March Audio]

1 hour ago, andyr said:

 

Zackly!  

 

Which is why I can't help thinking that the "more information" that's available from higher sampling rates in a digital recording ... similarly delivers a benefit.

 

 

If you watch the video you will understand why that's not the case.

The situation in mechanical systems is not analogous to digital sampling.

 

Or perhaps as you say you prefer reading you can read up on sampling theorem.

 

https://en.m.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem

 

The Nyquist–Shannon sampling theorem is a theorem in the field of signal processing which serves as a fundamental bridge between continuous-time signals and discrete-time signals. It establishes a sufficient condition for a sample rate that permits a discrete sequence of samples to capture   *all the information*   from a continuous-time signal of finite bandwidth.

 

The sampling theorem introduces the concept of a sample rate that is sufficient for perfect fidelity for the class of functions that are band-limited to a given bandwidth, such that no actual information is lost in the sampling process.

Edited January 20, 2023 by March Audio

[Cloth Ears]

1 hour ago, andyr said:

 

Zackly!  

 

Which is why I can't help thinking that the "more information" that's available from higher sampling rates in a digital recording ... similarly delivers a benefit.

 

Andy, I was specifically referring to analog systems, where 'more is better'. Unfortunately, this doesn't apply (or should I say "mostly doesn't apply"?) in the digital realm, no matter how much it would seem to be "obvious". AFAIK, higher sampling rate allows for proper 'transcription'(?) of the frequency (so 192kHz rate allows 96kHz frequency to be properly transcribed) and higher bit-rate allows for more program level (i.e. ~100dB for 24-bit, less that 90dB for 16-bit). But whether those actually provide a benefit (except on paper) is probably a whole other debate...

 

28 minutes ago, March Audio said:

 

If you watch the video you will understand why that's not the case.

The situation in mechanical systems is not analogous to digital sampling.

 

Or perhaps as you say you prefer reading you can read up on sampling theorem.

 

https://en.m.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem

 

The Nyquist–Shannon sampling theorem is a theorem in the field of signal processing which serves as a fundamental bridge between continuous-time signals and discrete-time signals. It establishes a sufficient condition for a sample rate that permits a discrete sequence of samples to capture   *all the information*   from a continuous-time signal of finite bandwidth.

And it is still a theorem, but it has withstood a reasonable test of time and attempts to debunk. So a pretty good working model...

[March Audio]

1 minute ago, Cloth Ears said:

Andy, I was specifically referring to analog systems, where 'more is better'. Unfortunately, this doesn't apply (or should I say "mostly doesn't apply"?) in the digital realm, no matter how much it would seem to be "obvious". AFAIK, higher sampling rate allows for proper 'transcription'(?) of the frequency (so 192kHz rate allows 96kHz frequency to be properly transcribed) and higher bit-rate allows for more program level (i.e. ~100dB for 24-bit, less that 90dB for 16-bit). But whether those actually provide a benefit (except on paper) is probably a whole other debate...

 

And it is still a theorem, but it has withstood a reasonable test of time and attempts to debunk. So a pretty good working model...

 

The theorem was proposed prior to the ability to actually perform digital sampling, but has subsequently proven correct.  It's been in practical use since the 60s.  As you say, no one has debunked it yet.

[aussievintage]

18 minutes ago, Cloth Ears said:

And it is still a theorem, but it has withstood a reasonable test of time and attempts to debunk. So a pretty good working model...

 

It's funny the negative notions that go with the word "theorem".  In this instance, it just means something that has been demonstrated to be true.

Edited January 20, 2023 by aussievintage