What do you audiophiles think about Pono music?

[M Paquette]

But aren't all DACs in the market now of the 1-bit Sigma/Delta type? And the 16-bit or 24-bit type refers to the bit resolution that is meaningful for the oversampling frequency and the processing (noise-shaping, anti-aliasing and all those details) built into the chip?

Heh. Yup!

No digital audio DACs, not even that insanely expensive Burr-Brown part, are true 16, 24, 32, or 48 bit DACs. They're all one bit DACs feeding sigma-delta modulators, a nifty little circuit built around the humble operational amplifier wired up as an integrator.

The number of bits only sets the maximum dynamic range of the encoded signal, and that is a function of the digital media encoding as an absolute maximum, and the signal processing done in mastering to ensure the digital media's dynamic range limit is not reached. (Clipping. In digital media, you really wouldn't like what happens...)

It's primarily a Marketing Department number. If someone can claim to have 192 KHz 24 bit audio decoding, they can sell more expensive cr*p to The Golden-Eared Ones, the True Audiophiles. You know, the guys who buy a couple of AudioQuest Type 4 ten foot speaker cables for $110.

 

[NW-Bound]

Heh. Yup!

No digital audio DACs, not even that insanely expensive Burr-Brown part, are true 16, 24, 32, or 48 bit DACs. They're all one bit DACs feeding sigma-delta modulators, a nifty little circuit built around the humble operational amplifier wired up as an integrator.

The number of bits only sets the maximum dynamic range of the encoded signal, and that is a function of the digital media encoding as an absolute maximum, and the signal processing done in mastering to ensure the digital media's dynamic range limit is not reached. (Clipping. In digital media, you really wouldn't like what happens...)

But, but, but the 24-bit chip thinggy has a higher oversampling frequency than the pedestrian 16-bit chip, more fancy digital filtering, etc... Surely the performance will be better.

Whether I or someone else can hear the difference is another thing, but the chip with the higher bit resolution IS better.

 

[M Paquette]

But as you say, and the point of my earlier post, regardless of the specific technology, I would not expect a phone to have anywhere near the accuracy in the audio path as a dedicated device. Would any technologist question that?

The entire audio path? Sure. It's a phone. Different purpose.

 

[M Paquette]

But, but, but the 24-bit chip thinggy has a higher oversampling frequency than the pedestrian 16-bit chip, more fancy digital filtering, etc... Surely the performance will be better.

Whether I or someone else can hear the difference is another thing, but the chip with the higher bit resolution IS better.

Paging Doctor Nyquist... Doc Nyquist, please pick up.

Once I'm sampling fast enough to accurately reconstruct the original signal, additional sampling doesn't actually improve things further. The extreme oversampling would just be noise not present in the original signal.

For purposes of feeding a delta-sigma modulator, I'll oversample at a (small) multiple of the target sample rate set by the digital media, specifically to meet the Nyquist limit criteria and feed the noise shaping of the modulator.

Fun audiophile example: Given a 48 KHz bandwidth two channel signal, what frequency can we claim in the marketing literature?

Well, the Nyquist rate for each 48 KHz wide channel would be sampling at 96 KHz. There are two channels, so we are clearly doing two samples at a time, or two streams at 96 KHz. Why, that must be 192 KHz total! Put that on the packaging!

 

[NW-Bound]

The oversampling I talked about was the internal oversampling prior to anti-aliasing the bandwidth down to the audible range. It's the oversampling that gives a chip the 24-bit resolution in contrast with the 16-bit one.

 

[ERD50]

Heck, my Tact amps are 1-bit DACs. Digital in, analog speaker outputs. I love them.

Right, I've got a few of those switching amps at home too. But mine (and my external DAC to drive my Class AB power amps) do not accept an input beyond 16 bit. (more on that later)

...
It's primarily a Marketing Department number. If someone can claim to have 192 KHz 24 bit audio decoding, they can sell more expensive cr*p to The Golden-Eared Ones, the True Audiophiles. You know, the guys who buy a couple of AudioQuest Type 4 ten foot speaker cables for $110.

But, but, but the 24-bit chip thinggy has a higher oversampling frequency than the pedestrian 16-bit chip, more fancy digital filtering, etc... Surely the performance will be better.

Whether I or someone else can hear the difference is another thing, but the chip with the higher bit resolution IS better.

I see this the same as NW-B. As I've said earlier in this (and probably other) threads, I'm not sure I could hear the difference between 16 bit source material and 24 bit source with reasonable respective hardware. But there is a measurable difference, due to technical differences. Or 'better' as NW-B puts it, but I suppose it is questionable if it is truly 'better' if no one can hear the difference. But I'm willing to give some 'golden ear' the benefit of the doubt when the measurements show differences that could at least lead to a plausibility of being detected (unlike many of the 'super cables').

The entire audio path? Sure. It's a phone. Different purpose.

Well that was the context of my post. It was in reply to someone claiming that something like the PONO isn't any better than a phone, since some phones can play FLAC format (I don't know if many/any phone can even decode 24 bit material). That's why I didn't understand what came across as a condescending comment from you:

Audiophiles...

-ERD50

 

[ERD50]

The oversampling I talked about was the internal oversampling prior to anti-aliasing the bandwidth down to the audible range. It's the oversampling that gives a chip the 24-bit resolution in contrast with the 16-bit one.

Right. It seems to me that M Paquette is somehow confusing the rate that the samples are taken/decoded with bit depth of the samples that are taken/decoded?

I can have source material of 16bit sampled at 44.1K, or say 24bit sampled at 44.1K. If my DAC (regardless of the specific technology) can only decode 16 bit inputs 44.1k times per second, it can't have more than 96db dynamic range. If it can decode 24 bits 44.1k times per second, it can (theoretically) provide 144 db dynamic range. Same sample rate.

Internally, those DACS (whether sigma-delta or binary weighted resistor ladder) must be different, and the 24 bit (all else being equal) will have better specs, right (realizing that it won't come near to 144 db, due to other limits)?

-ERD50

 

[springnr]

Honor Roll of well-mastered pop CDs, pretty much nothing after 1990.

...iTunes radio - some hopeful news
Dynamic Range compression: Are The Loudness Wars Over?

Sampling Theory For Digital Audio is a good article on digital sampling rate.

 

[NW-Bound]

No, I do not think M Paquette is confused between the bit depth and the oversampling.

All 1-bit D/A converters have to run with an internal conversion frequency at a multiple frequency of the sampling rate of the incoming digital signal. That was what I talked about.

At the output, after all the antialiasing and filtering, the output sample frequency may be back to 44.1KHz as done in old CD players, or it could be at a higher rate. The latter was what M Paquette talked about when he mentioned Nyquist.

PS. If the output sampling rate is high (oversampling), it makes anti-aliasing filtering an easier job. When the output sampling rate is 44.1KHz, a low-pass (brickwall) filter of 20KHz would need to be of a very high-order design, because the bandwidth approaches the Nyquist rate.

 

[ERD50]

No, I do not think M Paquette is confused between the bit depth and the oversampling.

All 1-bit D/A converters have to run with an internal conversion frequency at a multiple frequency of the sampling rate of the incoming digital signal. That was what I talked about.

At the output, after all the antialiasing and filtering, the output sample frequency may be back to 44.1KHz as done in old CD players, or it could be at a higher rate. The latter was what M Paquette talked about when he mentioned Nyquist.

PS. If the output sampling rate is high (oversampling), it makes anti-aliasing filtering an easier job. When the output sampling rate is 44.1KHz, a low-pass (brickwall) filter of 20KHz would need to be of a very high-order design, because the bandwidth approaches the Nyquist rate.

Right, and it is the oversampling that plays a major role in the S/N ratio of the DAC. Maybe the confusion here is that a 'one-bit' DAC is rated in terms of resolution by the theoretical S/N ratio that its topology can obtain, and that is expressed in 'bits of resolution'?

Some excerpts:

Demystifying Delta-Sigma ADCs - Tutorial - Maxim

Note that the SNR for a 1-bit ADC is 7.78dB (6.02 + 1.76). Each factor-of-4 oversampling increases the SNR by 6dB, and each 6dB increase is equivalent to gaining one bit.


....

If we apply a digital filter to the noise-shaped delta-sigma modulator, it removes more noise than does simple oversampling (Figure 6). This type of modulator (first-order) provides a 9dB improvement in SNR for every doubling of the sampling rate.

My brain hurts - off to bed!

-ERD50

 

[NW-Bound]

By the way, my laptop has D/A and A/D sampling rates up to 192KHz, and bit depth up to 24 bits. It's not that big a deal anymore.

I should try to sample input signals up to 96KHz (1/2 of 192KHz) to see what happens, but when I output a signal, I saw that there was a low-pass filter that rolled off fast beyond 20KHz.

 

[ERD50]

By the way, my laptop has D/A and A/D sampling rates up to 192KHz, and bit depth up to 24 bits. It's not that big a deal anymore.

I should try to sample input signals up to 96KHz (1/2 of 192KHz) to see what happens, but when I output a signal, I saw that there was a low-pass filter that rolled off fast beyond 20KHz.

Interesting, but of course, just because a device has some technical capability, that does not mean it is implemented well.

Like the 'mega-pixel' wars in cameras, the higher mega-pixel ratings didn't always mean higher quality - the actual size of the sensor, noise levels, and a few other specs said more about quality than the pixel count.

And I suspect that is the case in your laptop, and a phone. They won't have optimum ground layouts, or low noise power supplies, or high quality audio paths in general. At 16 bits, we are already talking ~ 96db capabilities, and that is already quite demanding on the other components and implementation to keep noise floor that low. So I doubt that increasing the stated bit rating of the DAC in those cases is meaningful.


Getting back to Nyquist rates, higher sample rates, and that 20kHz filter:

Of course there are debates about all things audio, but if we agree that output beyond 20KHz is not providing anything we can hear, I still think there may be advantages to sampling rates beyond the CD standard of 44.1KHz (which means you can theoretically capture sounds up to 22.05Khz).

It takes a sharp filter to pass 20KHz, but block 22.05KHz (often called a 'brick-wall filter' - you really don't want anything >22.05KHz leaking into the converter, that gets 'folded' back to a low frequency, which is a completely non-natural sound, just UGLY! *NOTE BELOW). So a higher sample rate can reduce the demands on that filter.

Now I don't know if a gentler filter is cheaper in practice, but if it is, that means that they can afford to spend more on other components, or provide more silicon real-estate to those other areas. And a gentler filter can have less phase shift in the audio band - again, I'm not sure that is an audible thing or not, but I would not rule these factors out.

So I can at least entertain the idea that a 48KHz, or even 96KHz could provide some benefit if there were also changes in the filter to take advantage of that rate. But I'm very, very, very skeptical that 128KHz is going to do anything that anyone could hear beyond what 96KHz could/might provide, 96KHz would already really lighten the demands on that filter, compared with 44.1Khz which already sounds very good.

*NOTE on Nyquist for less technical types: An easy visual parallel is the well known effect of seeing a wagon wheel appear to go backwards, or slow, or stop in a motion picture. The frames of film are 'sampling' the wheel at discreet times, just like an ADC 'samples' sound at discreet times. If the object moves fast compared to the sample rate, you see some slower, false difference frequency. Same in audio - if a 22KHz signal gets into a 44.1KHz converter, it will record a 50Hz signal. You will have a bass note appear out of nowhere!

-ERD50

 

[NW-Bound]

Interesting, but of course, just because a device has some technical capability, that does not mean it is implemented well.

Like the 'mega-pixel' wars in cameras, the higher mega-pixel ratings didn't always mean higher quality - the actual size of the sensor, noise levels, and a few other specs said more about quality than the pixel count.

And I suspect that is the case in your laptop, and a phone. They won't have optimum ground layouts, or low noise power supplies, or high quality audio paths in general. At 16 bits, we are already talking ~ 96db capabilities, and that is already quite demanding on the other components and implementation to keep noise floor that low. So I doubt that increasing the stated bit rating of the DAC in those cases is meaningful...

I suspect that you could be right that having the D/A-A/D chip sitting next to a bunch of fast switching digital chips is going to cause some noise to couple into the analog path. My point was that the chip itself was not that expensive anymore.

But this makes me wonder. Darn, now I'll just have to find out what the audio hardware inside my laptop is capable of. I have not been able to hear much difference between this and that, but I can "see" with scopes, FFT, and graphs.

 

[Mr. Paul]

Honor Roll of well-mastered pop CDs, pretty much nothing after 1990.

Lots of great sounding CDs there ...

 

[ERD50]

Honor Roll of well-mastered pop CDs, pretty much nothing after 1990.

...iTunes radio - some hopeful news
Dynamic Range compression: Are The Loudness Wars Over?

Sampling Theory For Digital Audio is a good article on digital sampling rate.

Thanks, some very interesting info in those links.

The last one covers some of what I talked about regarding diminishing returns in higher sample rates, and gave some specific advice (I bolded):

Nyquist pointed out that the sampling rate needs only to exceed twice the signal bandwidth. What is the audio bandwidth? Research shows that musical instruments may produce energy above 20 KHz, but there is little sound energy at above 40KHz. Most microphones do not pick up sound at much over 20KHz. Human hearing rarely exceeds 20KHz, and certainly does not reach 40KHz. The above suggests that 88.2 or 96KHz would be overkill. In fact all the objections regarding audio sampling at 44.1KHz, (including the arguments relating to pre ringing of an FIR filter) are long gone by increasing sampling to about 60KHz.

He says the downside of going beyond 60KHz (in addition to higher storage requirements and higher power consumption in the device), is that higher frequencies will actually reduce the accuracy in those converters - it's generally easier to get better accuracy at lower frequencies. 60KHz is apparently in the 'sweet spot'.



The iTunes article brings up some fascinating points that might be 'heresy' to some self-proclaimed 'audiophiles' (emph mine, with some minor edits for clarity):

Mastered for iTunes is an initiative to get higher quality conversions to AAC than ever before, by supplying 24-bit masters to Apple for encoding.

... generally, 256 kbps AAC sounds as good as or better than 320 kbps mp3 ...and 320 kbps mp3 is just a bit below CD quality. However, when the 256 kbps (AAC) are made from the 24 bit master, everything turns around, and many AAC masters sound better than the CD because the AAC preserves more of the depth and space from the 24-bit original than the 16-bit CD.

Ah-hah! 'Compressed' sound that is better than CD (but the key is to start with a master higher than CD quality)! I remember thinking about this years ago, but didn't know if it would pan out in real life. But I recently read an article about photo compression, and it makes sense there at least.

That photo article pointed out that for a given output file size, you get better image quality by starting with a relatively high resolution (large file size) and using relatively high compression, compared to a lower resolution and mild compression. IOW, the compression algorithms are pretty good and they do their intended job - the 'compressed' image is better than a non-compressed at the same file size. But you need that higher rez image at the start.

Apparently, the same is true of audio. Pretty neat. So if a 256 kbps from a 24 bit master can (apparently not 'always') sound better than CD quality, I would think that something like 512 kbps from a 24 bit master would always have more info than CD quality - and still be smaller than a CD quality FLAC (~ 720 kbps file size). Win-Win!

I still might want to have access to that master though, so that I could convert to other formats w/o generational artifacts from re-compressing, but that might not be a practical issue if that source compressed file is already higher than CD Q.

-ERD50

 

[NW-Bound]

Honor Roll of well-mastered pop CDs, pretty much nothing after 1990.

...iTunes radio - some hopeful news
Dynamic Range compression: Are The Loudness Wars Over?

Sampling Theory For Digital Audio is a good article on digital sampling rate.

The list of CDs is interesting. I will be sure to check some out, now that I am trying to move beyond the rank of casual listeners.

About the article on Digital Audio, it is good. But, but, but speaking of Nyquist sampling theorem, there's something that still bothers me when it applies to audio signals and its universally accepted bandwidth of 20KHz. Let me try to explain further.

If the signal is truly bandwidth limited at 20KHz as commonly accepted as the human hearing limit, then there should be no problem with the CD sampling rate of 44.1KHz. However, a signal that has a fast varying amplitude contains higher harmonics than indicated by its fundamental component. An example is given on page 7 of the article as a burst of a 3KHz sine wave, which is not properly reproduced from the sampled data points. An abrupt change in amplitude means that the signal has higher BW than seen in its fundamental components.

So, my questions are the following.

1) Is 20KHz adequate to describe the BW of the sounds generated by known instruments? Do the crash of the cymbals, and the sound of a percussion music piece require a higher sampling rate to capture their sharp attacks, even though their fundamentals are below 20KHz?

2) And if the above is true, and the waveforms are not accurately reproduced because harmonics above 20KHz are missing, can the listener tell even though he cannot hear above 20KHz? Can he tell that the fast-attack waveform has been "smeared"?

I ask the above questions as a layman, and would think that these have already been researched and answered. Please enlighten me. Note that these questions are somewhat academic for me, as my hearing is subpar.

 

[M Paquette]

Please remember that there are more losses in the path from performance to digital recording than data compression, and that data compression may be lossless!

When a recording master is being made, losses are deliberately introduced from shaping the frequency range from each microphone or pickup, adjusting levels to alter (generally reduce) the dynamic range of the performance, and in the case of a fair amount of pop music, by adding deliberate distortion, performing signal processing, and mixing multiple previously recorded performances together in novel ways. Bandpass filtering is done to meet whatever limits are specified for the master recording.

(And then there are all the wacky RIAA and similar pre-emphasis/de-emphasis/equalization curves, and the variations whipped up by CD mastering shops to 'improve' the sound. Maybe someone was worried about the laser light wearing out the pits...)

This may all be done on a digital or analog board, each with it's own advantages, drawbacks, and artifacts.

The digitized master might be compressed with one of a variety of lossless methods, like ALAC, ATRAC, HD-AAC, or one of the MPEG-4 lossless coders. Lossless compression allows for the entire digitized master to be recovered with bit-for-bit accuracy, without compression artifacts.

There are also lossy compression methods, like good old MP3 encoding. These filter the digitized master recording in various ways to produce something that is more easily compressed, at the expense of introduced artifacts and lost information compared to the original digitized master recording.

The point being that playing with compression only touches one spot in the long, tortured path audio takes from the singer's lips to your ears...

 

[Birdie Num Nums]

From the updated Pono Music Kickstarter site (currently ~$4M above its original $800K goal, with still 23 days yet to go!)

 

[NW-Bound]

....
When a recording master is being made, losses are deliberately introduced from shaping the frequency range from each microphone or pickup, adjusting levels to alter (generally reduce) the dynamic range of the performance, and in the case of a fair amount of pop music, by adding deliberate distortion, performing signal processing, and mixing multiple previously recorded performances together in novel ways. Bandpass filtering is done to meet whatever limits are specified for the master recording...

But would musicians like Neil Young who go the extra steps beyond the conventional ways be able to get us some better "true-to-life" music? Again, my hearing would not let me judge, but let's say if I do an objective measurement of his recordings, would I "see" something different?

By the way, I recently saw that the Studio L890 speakers by JBL have a tweeter dedicated to reproducing sounds above 20KHz. Above 20KHz! Are they kidding? What is that all about? JBL has been in business for a long time, and I would hesitate to call them a fool or a scammer.

 

[springnr]

Monty's Demo Pages
I like this video.... it uses test equipment I'm familiar with from back in the day. ;>)

Xiph.Org Video Presentations: Digital Show & Tell

 

[M Paquette]

1411-9216 Kbps? Finally, a recording format that preserves the Medium Wave and Short Wave content of audio recordings!

"It was like a heavy veil lifted off of my ears!"

Yah. Right.

I dunno about the rest of you, but I'm a 60 year old male exposed to weapons noise, turbines, and a nasty decompression accident, so my hearing tops out at 8 KHz, and has some dead spots below that.

I'd love to meet anyone human who can tell whether or not 1000 KHz tones are present in a proper double-blind test. Heck, I'd be impressed with detecting 100 KHz tones...

This is a Very Silly Music Spectrum.

 

[ERD50]

...
When a recording master is being made, losses are deliberately introduced from shaping the frequency range from each microphone or pickup, ....

The point being that playing with compression only touches one spot in the long, tortured path audio takes from the singer's lips to your ears...

Absolutely. And it seems the most 'realistic', 'natural' sounding recordings I have were recorded with the simplest of techniques, usually just two mics and minimal processing (and some great sounding ones were low budget affairs - often, just keeping the engineer out of the way is the best thing for the sound).

Sure, the path will always affect the signal to some degree, we can't repeal the laws of physics, but it really seems to me that with moderately high end equipment these days, one can easily tell a 'natural/minimalist' recording from a processed one. So the resolution (of all types, not just 'bits') seems to be there to a fairly high degree, or it would all smear together.

Along those lines, I recall seeing/hearing an opera singer on TV (Andrea Bocelli), and he sounded fantastic. So I got his CD out of the library, and I couldn't listen to it! The voice was so processed, it sounded all 'tizzy' to me. So fake! What a terrible thing to do to what I believe is a very good voice. But the TV fidelity was so poor, that I either could not hear that 'tizzy-ness' and/or his voice was not so processed on the TV show.

...

About the article on Digital Audio, it is good. But, but, but speaking of Nyquist sampling theorem, there's something that still bothers me when it applies to audio signals and its universally accepted bandwidth of 20KHz. Let me try to explain further. ...

So, my questions are the following.

1) Is 20KHz adequate to describe the BW of the sounds generated by known instruments? Do the crash of the cymbals, and the sound of a percussion music piece require a higher sampling rate to capture their sharp attacks, even though their fundamentals are below 20KHz?

2) And if the above is true, and the waveforms are not accurately reproduced because harmonics above 20KHz are missing, can the listener tell even though he cannot hear above 20KHz? Can he tell that the fast-attack waveform has been "smeared"?

I ask the above questions as a layman, and would think that these have already been researched and answered. Please enlighten me. Note that these questions are somewhat academic for me, as my hearing is subpar.

Well, our dear old friend Fourier tells us that an impulse such as that can be defined by its component sine waves. So I guess this gets circular - if we (you and I at least) cannot hear a steady-state sinewave > 20KHz, and those impulses are made up of shorter bursts of > 20KHz, then what would lead us to think we could hear a burst of > 20KHz, but not a steady state? So I lean towards thinking that it does not matter.

But I'm hesitant to completely rule it out, the ear/brain is a complex beast. I know some will claim that these higher freqs have some effect on us, one explanation being that our non-linear ear/brains create IM products between say a 27KHz tone and a 13KHz tone, creating a 14KHz difference tone that we 'hear'. Now if this is occurring between our nerve cells and our brain, it isn't something that we can measure (AFAIK), so who knows? But it seems a test could be set up - gate that 27KHz tone on/off and do we hear a difference in the 13KHz tone?

But it has made me curious, and maybe I'll try digging up some recent studies (or finish my taxes!). But the pragmatic side of me says if this effect exists, it must be very, very subtle - or if we were suddenly subjected to it, would we jump up and say "Yes! That is what is missing in a recording!"? Could be?

From the updated Pono Music Kickstarter site (currently ~$4M above its original $800K goal, with still 23 days yet to go!)

Interesting, but it is misleading to refer to those bars as 'quality'. They represent how many bits of info are stored. It is very difficult to say how much each increase in bit population correlates with 'quality', and there certainly diminishing returns. Plotting that on a log scale would likely be a better (but still crude) representation.

But would musicians like Neil Young who go the extra steps beyond the conventional ways be able to get us some better "true-to-life" music? Again, my hearing would not let me judge, but let's say if I do an objective measurement of his recordings, would I "see" something different?

By the way, I recently saw that the Studio L890 speakers by JBL have a tweeter dedicated to reproducing sounds above 20KHz. Above 20KHz! Are they kidding? What is that all about? JBL has been in business for a long time, and I would hesitate to call them a fool or a scammer.

Hard to say. I could make the case that reproducing sounds above 20KHz helps to assure a flat response and minimal phase shift up to/through 20KHz? It might not matter, it may just be that they found it wasn't all that hard to get the tweeter to go that high (after all, there are ultra-sonic transducers), so why not gain some bragging rights? I'll cut them a bit of slack, and say it was a bit of creative marketing rather than 'foolish' or 'scamming'.

I should shut up and go listen to some music!

-ERD50

 

[ERD50]

1411-9216 Kbps? Finally, a recording format that preserves the Medium Wave and Short Wave content of audio recordings!

"It was like a heavy veil lifted off of my ears!"

Yah. Right. ...

Isn't that a very flawed analogy?

If my ADC has 16 bit resolution and a 48KHz sample rate, that has nothing to do with recording frequencies in the 16*48/2 Khz range (384 KHz). That ADC can still only record signals up to <24KHz.

It has nothing at all to do with whether I can hear a 100KHz tone, it's not being recorded, it has to do with using > 64,000 bits to represent a 1 KHz tone (or any tone in the input filter's range). And (all else being equal), an ADC with 16 bit resolution will record that 1 KHz tone more accurately than an ADC with 8 bit resolution. Even if that 8-bit ADC has a 96 KHz sample rate, and the 16 bit has a 48 KHz sample rate.

That doesn't mean that I think these very high bit populations are meaningful, but we need to keep the comparisons useful.

This is a Very Silly Music Spectrum.

Let's not fight 'silly' with 'silly'.

-ERD50

 

[springnr]

So, my questions are the following.

1) Is 20KHz adequate to describe the BW of the sounds generated by known instruments? Do the crash of the cymbals, and the sound of a percussion music piece require a higher sampling rate to capture their sharp attacks, even though their fundamentals are below 20KHz?

2) And if the above is true, and the waveforms are not accurately reproduced because harmonics above 20KHz are missing, can the listener tell even though he cannot hear above 20KHz? Can he tell that the fast-attack waveform has been "smeared"?

Note that these questions are somewhat academic for me, as my hearing is subpar.

As an INTP, heavy on the I part, I know it sorta defeats the give an take of a forum, so sorry about posting links vs direct answers. Others have already disseminated the info quite well, definitely more verbose than I at any rate.

Interesting info on hearing in:

Gentlemen, meet your ears

 

[springnr]

It's primarily a Marketing Department number. If someone can claim to have 192 KHz 24 bit audio decoding, they can sell more expensive cr*p to The Golden-Eared Ones, the True Audiophiles. You know, the guys who buy a couple of AudioQuest Type 4 ten foot speaker cables for $110.

Ok Ok, if you can't afford 5K for a measly cable here an there, we will sell ya the parts to make your own for ~ 1K. Happy? ;>)

Bulk Buy Sale Bonanza