[GreenBow]

I keep wanting to buy CDs, but am put off. HD-audio will come soon and I will have to buy them all again.
 
Right now I am sat with Hawkwind selected in Amazon. I am trying to decide if I want to pay £15 for Sonic Attack on CD. That is a high price. Only there are un-missable tracks on it that I can not do without.
 
Yet sometime soon I hope HD-audio will become mainstream. However then I will be faced with having to replace a 44KHz CD recording with maybe a 192KHz HD-copy.
 
CD players are becoming more obsolete by the day, and some manufacturers simply don't even make them anymore. It's no-man's land at the moment.

 

 

[hogger129]

  I keep wanting to buy CDs, but am put off. HD-audio will come soon and I will have to buy them all again.
 
Right now I am sat with Hawkwind selected in Amazon. I am trying to decide if I want to pay £15 for Sonic Attack on CD. That is a high price. Only there are un-missable tracks on it that I can not do without.
 
Yet sometime soon I hope HD-audio will become mainstream. However then I will be faced with having to replace a 44KHz CD recording with maybe a 192KHz HD-copy.
 
CD players are becoming more obsolete by the day, and some manufacturers simply don't even make them anymore. It's no-man's land at the moment.

 

 
There is a simple way to see if high-resolution music is worth it or not.  Buy an album that was recorded and produced in 24-bits.  Downsample it properly with something like Voxengo r8brain, which is free btw.  Listen to them.  See if you can pick out which is the CD-quality one and which is the original file.  That should lay the issue to rest and tell you if high-resolution music is worth it for you.
 
I am probably in the minority, but CD is the best bang-for-the-buck right now.  You can get good-sounding ones, you just need to do your research on the best masterings before buying.
 
Another option is buying it on vinyl.  If it's a new band or newer release where the CD is mastered poorly, odds are that the vinyl is not mastered poorly.  Then just record it into your computer and burn it onto a CD.

 

[Music Alchemist]

"HD" audio is a scam. Red Book (16-bit / 44.1 kHz) is the highest resolution human ears benefit from. The only reason some hi-res downloads sound different is because they were derived from a different master. When you convert them to Red Book, they sound the same. So if it's a better master than the CD, it might be worth the money. If it's the same master and you already own the CD, you'd more or less be wasting your money. Unfortunately, the digital music stores normally don't disclose details on the master that is used.
 
Anyway, read this to get a grasp of the science behind all this: https://xiph.org/~xiphmont/demo/neil-young.html

 

[GreenBow]

Quote:

   
There is a simple way to see if high-resolution music is worth it or not.  Buy an album that was recorded and produced in 24-bits.  Downsample it properly with something like Voxengo r8brain, which is free btw.  Listen to them.  See if you can pick out which is the CD-quality one and which is the original file.  That should lay the issue to rest and tell you if high-resolution music is worth it for you.
 
I am probably in the minority, but CD is the best bang-for-the-buck right now.  You can get good-sounding ones, you just need to do your research on the best masterings before buying.
 
Another option is buying it on vinyl.  If it's a new band or newer release where the CD is mastered poorly, odds are that the vinyl is not mastered poorly.  Then just record it into your computer and burn it onto a CD.

Yes, that's a point that I have not bought any HD-audio. At the current price I don't intend to. However I came across one website which allows us to stream some clips of HD-audio examples. I will look for more with the intention of finding some with CD resolution to compare with.
 
 

  "HD" audio is a scam. Red Book (16-bit / 44.1 kHz) is the highest resolution human ears benefit from. The only reason some hi-res downloads sound different is because they were derived from a different master. When you convert them to Red Book, they sound the same. So if it's a better master than the CD, it might be worth the money. If it's the same master and you already own the CD, you'd more or less be wasting your money. Unfortunately, the digital music stores normally don't disclose details on the master that is used.
 
Anyway, read this to get a grasp of the science behind all this: https://xiph.org/~xiphmont/demo/neil-young.html

 
OK. I have started reading the link, but wow it's long. I will finish it though, even though I already understand the principles of HD-audio.
 
However this example/quote from it makes no sense, being where I have read to so far. (Under the Spectrophiles heading.)
"These near-IR LEDs emit from the visible boundry to at most 20% beyond the visible frequency limit. 192kHz audio extends to 400% of the audible limit. Lest I be accused of comparing apples and oranges, auditory and visual perception drop off similarly toward the edges."
 
It's a pointless analogy. The initial extra data we are looking for in HD-audio, is not 20KHz plus, i.e. above the hearing range. What we want is the detail of the hearing range spectrum. I have added why CD quality misses some detail below.
 
 
This diagram from the page makes no sense either. They have made a caption which says the signal recover the exact blus shape on playback.
 

The reason this is a poor example is exactly the reason why HD-audio is said to work. OK, if you have the patience I will explain. I have counted 32 samples on this single sine wave, and it would re-create a good example of the original wave. However and big but here. This only works for very low frequency sound waves. The reason is this. A sound wave of 10KHz will have 4.4 samples on it at sampling frequency 44.1KHz. Simple isn't it.
 
The higher up the sound frequency spectrum you go, then the less sample per sine wave. At 20KHz you get 2.2 samples per sine wave. OK you can argue most of us do not hear 20KHz. However some of us do, and young people do. Sound at 15KHz we will only get three samples.
 
The problem with such low sample rate is you miss data. E.g. with only 2.2 sample per sine wave. If the sample fall at the 0' and 180' of the sine wave, then you hear nothing. (0' and 180' being the start and mid point of the sine wave.)
 
At 10KHz and only 4.4 samples per wave, the sampl points could fall at roughly 45', 135', 215', and 315'. It means that you would hear the frequency of the sound, but at only half amplitude. This is why CD audio misses data.
 
We can argue there isn't a vast amount of sound data and instruments that play 10KHz and above. However there is a lot of information. It is precisely the high frequency stuff that gives instruments their individual sounds. It's the harmonics amplitudes that produce and instrument's timbre. I am not sure if we need that on playback because the harmonics have already constructed the sound. (If you know what I mean.) However you can not write off sound data above 10KHz.

 

[Music Alchemist]

 
https://en.wikipedia.org/wiki/44,100_Hz
 
44.1 kHz files are designed to play all the frequencies we can hear. (The average human hearing range is 20 Hz to 20 kHz.) Any frequencies above that are inaudible. 16-bit files have more than enough dynamic range to handle all recordings. 24-bit just adds more dynamic range, but there is no benefit from this, and no audible benefit to hi-res audio files. Just do the experiment I mentioned: convert a 24-bit / whatever kHz file to 16-bit / 44.1 kHz (using a program like dBpoweramp) and listen to both files. They sound exactly the same, unless there is a technical problem in your system.

 

[GreenBow]

Wow, you didn't take the time to understand what I wrote. Then come back with another thread for me to read. I don't need to read Wikipedia on this thanks. If you had understood what I put you would have noticed that.
 
With regard your quote about 44.1KHz being enough to play all frequencies we can hear. Well so is MP3, isn't it? Even 64kbps is. Where is your argument?
 
You must think I have no idea what i am talking about. Maybe you think Hawkwind are just a dum band or something. (Their early stuff is potentially some of the best rock music ever made. I haven't heard much of their later stuff.)

 

[RRod]

   
The reason this is a poor example is exactly the reason why HD-audio is said to work. OK, if you have the patience I will explain. I have counted 32 samples on this single sine wave, and it would re-create a good example of the original wave. However and big but here. This only works for very low frequency sound waves. The reason is this. A sound wave of 10KHz will have 4.4 samples on it at sampling frequency 44.1KHz. Simple isn't it.
 
The higher up the sound frequency spectrum you go, then the less sample per sine wave. At 20KHz you get 2.2 samples per sine wave. OK you can argue most of us do not hear 20KHz. However some of us do, and young people do. Sound at 15KHz we will only get three samples.
 
The problem with such low sample rate is you miss data. E.g. with only 2.2 sample per sine wave. If the sample fall at the 0' and 180' of the sine wave, then you hear nothing. (0' and 180' being the start and mid point of the sine wave.)
 
At 10KHz and only 4.4 samples per wave, the sampl points could fall at roughly 45', 135', 215', and 315'. It means that you would hear the frequency of the sound, but at only half amplitude. This is why CD audio misses data.
 
We can argue there isn't a vast amount of sound data and instruments that play 10KHz and above. However there is a lot of information. It is precisely the high frequency stuff that gives instruments their individual sounds. It's the harmonics amplitudes that produce and instrument's timbre. I am not sure if we need that on playback because the harmonics have already constructed the sound. (If you know what I mean.) However you can not write off sound data above 10KHz.

 
Some points:
.Certain people do hear at 20k, but they don't hear it at the same minimal amplitude at which they could hear, say, a 1k tone
.Ignoring quantization, the only time you would "miss" data is at the Nyquist frequency. There you could indeed line up with the zero crossings. For all other frequencies you will get some mismatch with the zero crossings and indeed not miss the waveform.
.You can generate test tones and verify that a full scale 10kHz tone is not half amplitude

 

[Music Alchemist]

  Wow, you didn't take the time to understand what I wrote. Then come back with another thread for me to read. I don't need to read Wikipedia on this thanks. If you had understood what I put you would have noticed that.
 
With regard your quote about 44.1KHz being enough to play all frequencies we can hear. Well so is MP3, isn't it? Even 64kbps is. Where is your argument?
 
You must think I have no idea what i am talking about. Maybe you think Hawkwind are just a dum band or something. (Their early stuff is potentially some of the best rock music ever made. I haven't heard much of their later stuff.)

 
What you wrote is irrelevant. I clearly explained why it is physically impossible for there to be an audible difference between 16-bit / 44.1 kHz and 24-bit / whatever kHz. (I meant in the context of lossless, though. If you convert to 64 kbps MP3, then of course it's going to sound worse.) Just do the conversion and listen for yourself if you don't believe me.

 

[GreenBow]

   
What you wrote is irrelevant. I clearly explained why it is physically impossible for there to be an audible difference between 16-bit / 44.1 kHz and 24-bit / whatever kHz. (I meant in the context of lossless, though. If you convert to 64 kbps MP3, then of course it's going to sound worse.) Just do the conversion and listen for yourself if you don't believe me.

I said 64kbps will play all the audio frequency, which is what you are claiming about 44.1KHz.
 
I did not argue that it would not sound worse. If I didn't think lower res audio sounded lower quality, why would I be wanting HD-audio. Headphone Supremus you may be but you have no idea of the science behind this. I know exactly what I am talking about.
 
You are either not interested in what I am saying. Or you are trolling me and I will not reply any more. I have patiently explained it to you. Either you are trolling or you don't have the inclination to read or understand my post. I won't reply to you anymore, not until I hear a something different from you.
 
In fact in the link you gave me, the next diagram supports what I said.
 

 
Here you can see where amplitude data is missed. The peaks and troughs of the signal are missed because the sampling rate is not high enough. Exactly what I said.
 
However this is not a discussion about whether high res-audio works. It's about when is it coming mainstream or affordable, like I said in first post. I think it;s still a premium price because quality portable HD-audio players are expensive. Plus people are still content to download MP3 because it's compact. CD players are going out of fashion because CD sales are falling and downloading is is more popular. Many high quality manufactureres don't even make CD players anymore, and stopped years ago. Retailers of HD-audio are taking advantage by charging vast amounts.
 
End.
 
 

   
Some points:
.Certain people do hear at 20k, but they don't hear it at the same minimal amplitude at which they could hear, say, a 1k tone
.Ignoring quantization, the only time you would "miss" data is at the Nyquist frequency. There you could indeed line up with the zero crossings. For all other frequencies you will get some mismatch with the zero crossings and indeed not miss the waveform.
.You can generate test tones and verify that a full scale 10kHz tone is not half amplitude

There are zillions of places you would miss data. It happens anywhere that the peak of the signal does not match the exact point a sample is taken. Please don't take offence. However I don't know how to explain it any more simply that I have already.

 

[Music Alchemist]

  I said 64kbps will play all the audio frequency, which is what you are claiming about 44.1KHz.
 
I did not argue that it would not sound worse. If I didn't think lower res audio sounded lower quality, why would I be wanting HD-audio. Headphone Supremus you may be but you have no idea of the science behind this. I know exactly what I am talking about.
 
You are either not interested in what I am saying. Or you are trolling me and I will not reply any more. I have patiently explained it to you. Either you are trolling or you don't have the inclination to read or understand my post. I won't reply to you anymore, not until I hear a something different from you.
 
In fact in the link you gave me, the next diagram supports what I said.
 
Here you can see where data is missed. The peaks and troughs of the signal are missed because the sampling rate is not high enough. Exactly what I said.
 
However this is not a discussion about whether high res-audio works. It's about when is it coming.
 
End.

There are zillions of places you would miss data. It happens anywhere that the peak of the signal does not match the exact point a sample is taken. Please don't take offence. However I don't know how to explain it any more simply that I have already.

 
I'm just telling it like it is. This isn't about the data; it's about what we can hear. Under controlled conditions, human ears can't distinguish between 256 kbps AAC and lossless Red Book, much less lossless Red Book and hi-res. If you understand that the data is irrelevant when we can't hear any of the extra data, then we are on the same page. You shouldn't care about when "HD" audio is "coming" (it has already come; it just isn't as popular as other stuff), because it has zero audible benefit over Red Book. What you should care about is the recording and mastering quality. If a hi-res download is derived from a different master, it has a good chance of sounding better than the CD version. However, if it's the same master, it will sound the same as the CD. I'm not saying that hi-res downloads aren't worth buying; only when they are the same master as a version you already paid for in the past.
 
Anyway...here's a link you should like:
 
http://www.head-fi.org/a/list-of-lossless-and-high-res-music-flac-alac-aiff-dsd-dxd-etc-download-sites

 

[RRod]

  Here you can see where data is missed. The peaks and troughs of the signal are missed because the sampling rate is not high enough. Exactly what I said.
 
However this is not a discussion about whether high res-audio works. It's about when is it coming.
 
 

There are zillions of places you would miss data. It happens anywhere that the peak of the signal does not match the exact point a sample is taken. Please don't take offence. However I don't know how to explain it any more simply that I have already.

 
The data ultimately have to be reconstructed by the DAC. You can easily test the kind of thing that happens by interpolating a 44.1 signal to a very high rate: the "missed" peaks will reappear. See also this.

 

[GreenBow]

   
The data ultimately have to be reconstructed by the DAC. You can easily test the kind of thing that happens by interpolating a 44.1 signal to a very high rate: the "missed" peaks will reappear. See also this.

That reminds me of the upsampling software that is appearing now. The Meridian Explorer 2 has it. (I have the original Meridian Explorer.) The Sony portable music players have it. They say they can reproduce music at a higher frequency than it was sampled. However this plays into my debate. Why do it if HD-audio doesn't work?
 
To me technically HD-audio explains itself. However I have googled about upsampling, but am struggling to even find a paper about how it works.
 
Anyway going back to square one. The easiest way I can exaplin of how higher sampling rates work is this. The more samples you have, the closer you get to analogue audio. I know there's a high chance that will cause another argument. However if people understood any of what I have written, it makes sense. If you imagine an infinite sampling rate - you would get analogue out.
 
All that needs to come out of a DAC chip is sine wave. The analogue circuits make the music tonality. Therefor the more accurate sampled the sine wave, the more accurate the reproduction. In other words the more samples you take, the more accurate the reproduction.
 
I should add though that I don't dislike CD quality. I am not jumping around annoyed that I am not getting full-fat audio. I am just dissapointed that HD is not mre accessable at a more relevant price. I think it's still a premium price because quality portable HD-audio players are expensive. Plus people are still content to download MP3 because it's compact. CD players are going out of fashion because CD sales are falling and downloading is is more popular. Many high quality manufactureres don't even make CD players anymore, and stopped years ago. Retailers of HD-audio are simply taking advantage by charging vast amounts.
 
We also need to consider why we had CD in the first place. It was because that was all the data we could fit on a disc. There was 48K at the time CD was introduced, but it made too much data. HD-disc players were attempted to be introduced years later, but no one had the music, so they failed. It is only in the download age that higher than CD-quality can really work.

 

[RRod]

The "stairstep" logic (which is what you seem to be pushing) is faulty, as has already been pointed out. I'll leave you with this.

 

[Music Alchemist]

  Anyway going back to square one. The easiest way I can exaplin of how higher sampling rates work is this. The more samples you have, the closer you get to analogue audio. I know there's a high chance that will cause another argument. However if people understood any of what I have written, it makes sense. If you imagine an infinite sampling rate - you would get analogue out.

All that needs to come out of a DAC chip is sine wave. The analogue circuits make the music tonality. Therefor the more accurate sampled the sine wave, the more accurate the reproduction. In other words the more samples you take, the more accurate the reproduction.

 
Doesn't matter how much closer you get in terms of data when there isn't an audible difference in the first place. Here, I'll prove it to you. Download and install the free trial of dBpoweramp. Choose a 24-bit file. Right-click it and click Convert To. Next to Converting To, select either WAV or AIFF. Next to Uncompressed, select 16 bit and 44.1 kHz. Click Convert. Now listen to both files. They sound exactly the same.

 

[GreenBow]

  The "stairstep" logic (which is what you seem to be pushing) is faulty, as has already been pointed out. I'll leave you with this.

That nearly had me.
It says: Nyquist Sampling Theory: A sampled waveforms contains ALL the information without any distortions, when the sampling rate exceeds twice the highest frequency contained by the
sampled waveform'
 
However it is ambiguous, and this is why. It states it contains all the info without distortions. The reason it is ambiguous is because it is referring to the necessary sample rate to eliminate alaising. Alaising occus when the sampling frequency is less than 2x the audio frequency. It doesn't cover amplitude misrepresentation. Like I said, if the sample points fall on the 0' and 180' of the sine wave, then the amplitude is lost. How anyone can not understand that, is beyond me.
 
This explains it. http://whatis.techtarget.com/definition/Nyquist-Theorem
According to the Nyquist Theorem, the sampling rate must be at least 2f_max, or twice the highest analog frequency component. This undesirable condition is a form of distortion called aliasing.
 
This also states that more than Nyquist is preferable.  https://www.cs.cf.ac.uk/Dave/Multimedia/node149.html
(This is an academic website.)
 
The Alaising paragraph on this page explains it quite well. https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem
 
However you have said that the step function as you called it is faulty. This is not correct. The Nyquist Theorum only gaurantees the frequency component of the signal. There is no step function out of a DAC. It is a sine wave and smooth, but potentiall missing amplitude components. That is what causes lower quality music to differ from higher. Direction, spatial, and musical cues come from the higher end components of music. In retrospect, no directional information is given from the frequency of low frequency sound. That is why it is not relevant where you place a sub unit. (I theorise this is why they say HD-audio has better positioning, and clarity. However I have not read as such, and it's only my guess.)
 
However I am exhausted with this for today. I have practically written an A'level paper length on it with theory from BSc.
 
 

Originally Posted by Music Alchemist /img/forum/go_quote.gif
Anyway...here's a link you should like:
 
http://www.head-fi.org/a/list-of-lossless-and-high-res-music-flac-alac-aiff-dsd-dxd-etc-download-sites

 
Thank you, that's a good link. I found something similar on the Sony website. I am going to add it to favourites.