High resolution audio(not impressed)

[davedotco]

He didn't come to me and say, "I've got some info for ya, but don't tell anyone who I am as I don't want to be linked to it". This guy has been involved with major studios, and has been doing this for some time, so he's very experienced - he's not a tea boy.

Funny how some throw in about how studios do things differently to home audio, and that there's no BS in the studio, but I quote one and everyone's dubious.

Those that want to listen can. Everyone else can continue as they were.

Are often knee deep in BS, it's just a different type of BS.

The mathematics of digital sampling are known and understood, the whole technology is built on it.

Sampling at 16 bit 44.1 kHz will reconstruct any signal up to 22.05kHz precisely, no ifs or buts but with 100% accuracy, this is a proven fact.

So right from the outset, your expert is wrong.

 

[Al ears]

He didn't come to me and say, "I've got some info for ya, but don't tell anyone who I am as I don't want to be linked to it". This guy has been involved with major studios, and has been doing this for some time, so he's very experienced - he's not a tea boy.

Funny how some throw in about how studios do things differently to home audio, and that there's no BS in the studio, but I quote one and everyone's dubious.

Those that want to listen can. Everyone else can continue as they were.

Involved with major studios in what capacity though? Did he explain how all sampling theory, proven over decades is completely wrong? Because that's what he's saying and unless he can provide the mathematics to show why the accepted theories could be so wrong then people will rightly be skeptical because he's talking nonsense.

Unless of course he's suggesting we can hear frequencies higher than 22.05khz (as that's what CD audio tops out at), in which case you should stop consulting with cats...

Surely that's purely the theoretical maximum, it doesn't mean this sort of level is actually used.

 

[Frank Harvey]

So right from the outset, your expert is wrong.

As you wish.

Oh, and it is my choice not to name him. There was a couple of things he mentioned where he said specifically not to mention him, but I told him I wasn't going to reveal his identity at all.

 

[nopiano]

Maybe I've missed something in the last few exchanges, but I thought there has long been a firm belief that reproducing higher frequencies - for example with super tweeters - can bring some benefits. That doesn't mean - as with so much in audio - that there may not be problems that go along with it (such as picking up unwanted radio signals). Overtones from certain musical instruments may have worthwhile content above 22kHz? These may affect the audible frequencies.

As it isn't 1 April, I read recently (but don't recall where) of some research that suggested we "hear", or perhaps perceive, ultrasonics via our eyeballs - I kid you not. The suggestion was to remove your specs when listening to music, something I've been meaning to experiment with soon. I guess if you have good eyesight, you already perceive everything you can.

On the studio aspect, I thought they usually recorded at higher resolutions to enable more headroom, and give scope for remixing etc. Just as in analogue days they used 30ips tape, even if half that speed would have been ok.

 

[Frank Harvey]

Maybe I've missed something in the last few exchanges, but I thought there has long been a firm belief that reproducing higher frequencies - for example with super tweeters - can bring some benefits. That doesn't mean - as with so much in audio - that there may not be problems that go along with it (such as picking up unwanted radio signals). Overtones from certain musical instruments may have worthwhile content above 22kHz? These may affect the audible frequencies.

They do.

 

[shadders]

He didn't come to me and say, "I've got some info for ya, but don't tell anyone who I am as I don't want to be linked to it". This guy has been involved with major studios, and has been doing this for some time, so he's very experienced - he's not a tea boy.

Funny how some throw in about how studios do things differently to home audio, and that there's no BS in the studio, but I quote one and everyone's dubious.

Those that want to listen can. Everyone else can continue as they were.

Are often knee deep in BS, it's just a different type of BS.

The mathematics of digital sampling are known and understood, the whole technology is built on it.

Sampling at 16 bit 44.1 kHz will reconstruct any signal up to 22.05kHz precisely, no ifs or buts but with 100% accuracy, this is a proven fact.

So right from the outset, your expert is wrong.

Hi,

This not correct. The sampling theory states that the reconstruction has to be an exact replication of the voltage level of each original sample, but this is not the case, as we have quantisation noise/error.

In addition, the perfect reconstruction sample is a sinc function, which many DAC's do not use so there is additional error.

This is simplistically stated.

Regards,

Shadders.

 

[Andrew17321]

The mathematics of digital sampling are known and understood, the whole technology is built on it.

Sampling at 16 bit 44.1 kHz will reconstruct any signal up to 22.05kHz precisely, no ifs or buts but with 100% accuracy, this is a proven fact.

That is true provided the conditions are met precicely. That includes the signal not changing (a single sustained note) and the sampling done exactly (ie not using digital approximations). I am a mathematician!

 

[The_Lhc]

He didn't come to me and say, "I've got some info for ya, but don't tell anyone who I am as I don't want to be linked to it". This guy has been involved with major studios, and has been doing this for some time, so he's very experienced - he's not a tea boy.

Funny how some throw in about how studios do things differently to home audio, and that there's no BS in the studio, but I quote one and everyone's dubious.

Those that want to listen can. Everyone else can continue as they were.

Involved with major studios in what capacity though? Did he explain how all sampling theory, proven over decades is completely wrong? Because that's what he's saying and unless he can provide the mathematics to show why the accepted theories could be so wrong then people will rightly be skeptical because he's talking nonsense.

Unless of course he's suggesting we can hear frequencies higher than 22.05khz (as that's what CD audio tops out at), in which case you should stop consulting with cats...

Surely that's purely the theoretical maximum, it doesn't mean this sort of level is actually used.

I think you're getting mixed up with dynamic range (a function of bit depth not sampling rate). If you set your sample rate to 44.1khz you will capture everything up to 22.05khz.

 

[The_Lhc]

The mathematics of digital sampling are known and understood, the whole technology is built on it.

Sampling at 16 bit 44.1 kHz will reconstruct any signal up to 22.05kHz precisely, no ifs or buts but with 100% accuracy, this is a proven fact.

That is true provided the conditions are met precicely.  That includes the signal not changing (a single sustained note) and the sampling done exactly (ie not using digital approximations).  I am a mathematician!

Oh good, you'll be able to tell us all about Fourier then.

 

[steve_1979]

The mathematics of digital sampling are known and understood, the whole technology is built on it.

Sampling at 16 bit 44.1 kHz will reconstruct any signal up to 22.05kHz precisely, no ifs or buts but with 100% accuracy, this is a proven fact.

That is true provided the conditions are met precicely. That includes the signal not changing (a single sustained note) and the sampling done exactly (ie not using digital approximations). I am a mathematician!

Hi Andrew. Would you mind explaining these two points in a bit more detail please?

p.s. I'm not trolling for the sake of argument. I am genuinely interested to learn more about how and why "the signal not changing" and "the sampling done exactly (not using digital approximations)" can effect Nyquist-Shannon sampling theorem.

I'd also be interested to learn what effect (if any) these two things could have on real world applications of digital audio.

 

[andyjm]

Maybe I've missed something in the last few exchanges, but I thought there has long been a firm belief that reproducing higher frequencies - for example with super tweeters - can bring some benefits.

I think the 'firm belief' is called marketing. How about a link to a study that can demonstrate that humans can perceive frequencies outside their hearing range?

 

[andyjm]

The mathematics of digital sampling are known and understood, the whole technology is built on it.

Sampling at 16 bit 44.1 kHz will reconstruct any signal up to 22.05kHz precisely, no ifs or buts but with 100% accuracy, this is a proven fact.

That is true provided the conditions are met precicely. That includes the signal not changing (a single sustained note) and the sampling done exactly (ie not using digital approximations). I am a mathematician!

Hi Andrew. Would you mind explaining these two points in a bit more detail please?

p.s. I'm not trolling for the sake of argument. I am genuinely interested to learn more about how and why "the signal not changing" and "the sampling done exactly (not using digital approximations)" can effect Nyquist-Shannon sampling theorem.

I'd also be interested to learn what effect (if any) these two things could have on real world applications of digital audio.

At the risk of diving in (and also being called Andrew) - "Nyquist-Shannon sampling theorem" - is a theorem, not a theory. A theory is a belief, generally accepted by the scientific community - in many cases supported by experimental data - but not proved absolutely. The 'big bang theory' is a good example of this, widely believed, lots of evidence to support it, but no definitive proof. A theorem is derived mathematically from a set of basic mathematical principles, and absent an error in the working out, is a definitive proof.

Nyquist-Shannon imposes constraints in order to make the maths work - and remember the theorem proves that the output of a sampled system is exactly the same as the input within these constraints. These constraints have implications - one not unreasonable constraint is that there are no errors introduced by the sampling process, that each sample is an exact measurement of the waveform. Another constraint is that the waveform to be sampled is 'band limited' - has no spectral energy outside of a defined upper frequency. This has non-obvious Fourier implications regarding infinite length samples and invariate signals.

Now unlike the other Andrew, I am not a mathematician, I was an engineer - and in the great tradition of engineering, it doesn't have to be exactly right, it just has to be good enough.

So, your home audio system is violating two fundatmental constraints of Nyquist / Shannon - it uses discrete sample levels introducing quantisation noise, and the original sampling process wasn't perfectly band limited which has aliasing implications.

Mathematically a disaster, engineering-wise it works just fine.

 

[steve_1979]

That's pretty much what I thought (ok guessed) it would be.

I have a pretty good grasp of the concept of N-S theorem but much of the actual maths involved is above me.

 

[nopiano]

Maybe I've missed something in the last few exchanges, but I thought there has long been a firm belief that reproducing higher frequencies - for example with super tweeters - can bring some benefits.

I think the 'firm belief' is called marketing. How about a link to a study that can demonstrate that humans can perceive frequencies outside their hearing range?

Hi Andy, there is an old study in Nature from the 1950s but I've not paid to download the whole extract. Bits I've found do suggest perception beyond the normally accepted upper limit of 20kHz.

Also, the other article has contributors who can hear pest control devices and other items with vhf outputs. Links below.

I realse this is fertile ground for spin and marketing. But ever since I first saw wideband amplifiers accurately reproducing square waves (thank you harman/kardon) I have always believed that retaining the harmonics beyond our nominal upper hearing limit of tones to be a worthwhile aim in hifi. I readily acknowledge it is more intuition than concrete evidence. My choice of equipment invariably follows this belief!

http://recordinghacks.com/articles/the-world-beyond-20khz/
http://www.nature.com/nature/journal/v166/n4222/abs/166571b0.html#References

 

[Andrew17321]

In response to steve_1979 in #185, andyjm in #187 largely answered your question. I would add that the smaller the sampling errors are and the more frequently the sampling happens, relative to the frequencies involved, the closer we get to the theoretical situation, and in theory, the better the system should sound in terms of accuracy.

There is another point to consider: an ear is not simply a frequency analyser. It has an amazing ability to detect transients and to determine the direction from where they are coming. Even very quiet transients in a noisy background. Without that ability our species would probably have been eaten-out many thousands of years ago. Accurate sound reproduction is not a simple business.

Going back to the original question, my music system can deal with 24 bit and higher frequencies, and I have listened to some classical music music at CD quality and at higher quality. I think/imagine there is more involvement and excitement at the higher quality, but not enough to justify the extra cost. I much prefer spending my money on real live classical concerts!

(I am off to look after grandchildren now, so won't be responding for a while. Regards)

 

[CnoEvil]

.Going back to the original question, my music system can deal with 24 bit and higher frequencies, and I have listened to some classical music music at CD quality and at higher quality. I think/imagine there is more involvement and excitement at the higher quality, but not enough to justify the extra cost. I much prefer spending my money on real live classical concerts!

I agree with you....but....Linn have admitted that they often use a different Master for 24 bit, so others may also do so as well. So are we hearing a better Master, or improvements due to 24 bit....or both?

 

[Frank Harvey]

So are we hearing a better Master, or improvements due to 24 bit....or both?

Either way, I don't care - if it sounds better, it is better. I'd rather get on and listen to and enjoy the music rather than get caught up in circular discussions that end up with both sides thinking they're right and no one budging

 

[The_Lhc]

So are we hearing a better Master, or improvements due to 24 bit....or both?

Either way, I don't care - if it sounds better, it is better. I'd rather get on and listen to and enjoy the music

And you're obviously quite happy to be ripped off by a company charging more for the 24-bit version when the different master can be perfectly reproduced in 16-bit. THAT'S the point. If 24-bit audio was the same price as 16-bit nobody would care and nobody would bother having this argument (except Sonos owners) as the consumer would be no worse off regardless of which version they bought. But it isn't the same price, 24-bit is (nearly) always more expensive for no good reason.

 

[CnoEvil]

And you're obviously quite happy to be ripped off by a company charging more for the 24-bit version when the different master can be perfectly reproduced in 16-bit. THAT'S the point. If 24-bit audio was the same price as 16-bit nobody would care and nobody would bother having this argument (except Sonos owners) as the consumer would be no worse off regardless of which version they bought. But it isn't the same price, 24-bit is (nearly) always more expensive for no good reason.

The problem is, if you want the better sounding version, you have to pay more for it....rip off or not.

In the case of Linn, I have a few tracks at different bit rates (as bought from the website), and imo the 24 bit versions sound better, especially the ambient detail. I don't think they are worth the extra though...I got them free when I bought the DS.

 

[The_Lhc]

And you're obviously quite happy to be ripped off by a company charging more for the 24-bit version when the different master can be perfectly reproduced in 16-bit. THAT'S the point. If 24-bit audio was the same price as 16-bit nobody would care and nobody would bother having this argument (except Sonos owners) as the consumer would be no worse off regardless of which version they bought. But it isn't the same price, 24-bit is (nearly) always more expensive for no good reason.

The problem is, if you want the better sounding version, you have to pay more for it....rip off or not.

Errr, yes, that's what I said isn't it?

In the case of Linn, I have a few tracks at different bit rates (as bought from the website), and imo the 24 bit versions sound better, especially the ambient detail. I don't think they are worth the extra though...I got them free when I bought the DS.

I'll hold my hands and say I have bought 24-bit audio in the past, simply because of the off chance a different master was being used, however I've only done that when the price difference was minimal, eg Björk's Vulnicura was only a pound more for 24-bit (and was only 8 quid all in) so it didn't seem worth worrying about. I've never listened to it in 24-bit mind, I don't have anything to play it on. I immediately down converted it to 16/48 for the Sonos and 320kbps mp3 for the phone.

 

[Frank Harvey]

And you're obviously quite happy to be ripped off by a company charging more for the 24-bit version when the different master can be perfectly reproduced in 16-bit. THAT'S the point. If 24-bit audio was the same price as 16-bit nobody would care and nobody would bother having this argument (except Sonos owners) as the consumer would be no worse off regardless of which version they bought. But it isn't the same price, 24-bit is (nearly) always more expensive for no good reason.

Not really. Like you, I won't pay the earth for them. There are two things I won't pay over the odds for - CDs, and 'non physical' downloads.

I don't have a lot of hi-res downloads, but the Peter Gabriel ones came with the vinyl copies, and the ones I have purchased have been only a pound or so more than the CD copies.

 

[Infiniteloop]

And you're obviously quite happy to be ripped off by a company charging more for the 24-bit version when the different master can be perfectly reproduced in 16-bit. THAT'S the point. If 24-bit audio was the same price as 16-bit nobody would care and nobody would bother having this argument (except Sonos owners) as the consumer would be no worse off regardless of which version they bought. But it isn't the same price, 24-bit is (nearly) always more expensive for no good reason.

The problem is, if you want the better sounding version, you have to pay more for it....rip off or not.

In the case of Linn, I have a few tracks at different bit rates (as bought from the website), and imo the 24 bit versions sound better, especially the ambient detail. I don't think they are worth the extra though...I got them free when I bought the DS.

I have downloaded a few 24 bit Albums from Linn and they invariably sound excellent. To my ears, 24 bit sounds 'richer' and more effortless to listen to.

 

[CnoEvil]

Errr, yes, that's what I said isn't it?

Possibly.

IMO. It could be argued that it's not a rip off, if it gives you an improvement that you think is worth the premium. The fact that they might do better on CD, is academic...though I can also see where you're coming from.

I'm just thinking out loud....pay no heed!

 

[CnoEvil]

I have downloaded a few 24 bit Albums from Linn and they invariably sound excellent. To my ears, 24 bit sounds 'richer' and more effortless to listen to.

A really good track to hear the difference is called, "A case of you" by Ian Shaw, from the Album "Drawn to all things", on Linn Records.

It's a very simple recording, with Voice, Piano and Double Bass and allows you to hear the difference going from 320 -> 16 Bit -> 24 Bit.

 

[BigH]

I have downloaded a few 24 bit Albums from Linn and they invariably sound excellent. To my ears, 24 bit sounds 'richer' and more effortless to listen to.

A really good track to hear the difference is called, "A case of you" by Ian Shaw, from the Album "Drawn to all things", on Linn Records.

It's a very simple recording, with Voice, Piano and Double Bass and allows you to hear the difference going from 320 -> 16 Bit -> 24 Bit.

Yes but as said before it may not down to the bits. Someone did a comparison, took a 24 bit Linn track converted it to 16 bit, it sounded and the soundwave were exactly the same, however the Linn 16 bit version was not the same, the soundwave was quite a lot different.