24/192 Music Downloads ...and why they make no sense

[Gerry_the_Merry]

Whither Hi-Rez?

Why downloading 24-bit 193khz music is pointless very interesting article. Waiting for our resident experts to point out why such drivel will generate more audiophools, or to point out if we have been audiofools all along anyway...!



24/192 Music Downloads are Very Silly Indeed

 

[ankaka]

Re: Whither Hi-Rez?

:clapping::clapping::clapping::clapping::clapping:

 

[thatguy]

After 24 bit, someone tries to debunk 192 KHz

I have seen a few posts on this forum on how 24 bit depth doesn't make sense. Well, here is an article about why 192 KHz doesn't get you much either. It was on Slashdot yesterday

24/192 Music Downloads are Very Silly Indeed

The slashdot.org discussion:

Why Distributing Music As 24-bit/192kHz Downloads Is Pointless - Slashdot

Disclaimer: I read the article but did not understand all the arguments presented. Have bookmarked it for later reading.

 

[jls001]

So where does this place DSD sampling rate of 64 times 44.1 kHz = 2822.4 kHz?

 

[msagar]

I assume people are commenting after experiencing these mediums. I personally experienced DSD which recreates original event the way it's recorded. This is one of the reasons why most of the major labels preserving their masters using DSD.

I listen to 24/192 to music. I do not believe in measurements and/or technical jargon, JUST BELIEVE IN WHAT I HEAR. I would say they sound better than 16/44 and 24/96.

 

[msagar]

Re: After 24 bit, someone tries to debunk 192 KHz

I listen to 24/192 to music. I do not believe in measurements and/or technical jargon, JUST BELIEVE IN WHAT I HEAR. I would say they sound better than 16/44 and 24/96.

 

[sound1]

Re: After 24 bit, someone tries to debunk 192 KHz

Very informative article. Thanks for sharing!

 

[msagar]

Re: Whither Hi-Rez?

I listen to 24/192 to music. I do not believe in measurements and/or technical jargon, JUST BELIEVE IN WHAT I HEAR. I would say they sound better than 16/44 and 24/96.

 

[malvai]

Re: After 24 bit, someone tries to debunk 192 KHz

switch to vinyl guys.... you'll stop reading such stuff! lol

 

[Shivam]

Re: After 24 bit, someone tries to debunk 192 KHz

24/96 and 24/192 do sound superior to 16/44.1
In fact recently when my computer was sending my 24/96 files as 16/44.1 to my DAC due to a small computer configuration error, all I needed was a few seconds of music to realize that the quality was not upto the mark !

... and yes vinyl is one of the other options.

 

[Kannan]

A very very good article on this topic

16 vs. 24 bit Audio Recording Demystified

 

[captrajesh]

Duplicate thread

There is another thread which is the first one to post this info.

It would be better if these three threads are merged.

Sent a PM to supermod. Bragging rights wouldn go the Gerry the Merry, rightfully

 

[captrajesh]

I have merged the threads.

Only the 2nd and 3rd threads are merged. The first thread has not been merged.

Thereafter, post numbers 1,2,7,8,12 & 14 would become redundant and could also be deleted.

 

[doors666]

I didnt read the full article as its pretty long, but some of the items sounded like bull to me. Author claims that once you are above nyquist freq, you can reconstruct the full analog waveform without any distortion or smoothing issues. This just beats physics.
Take 50hz for example, you just need 100 samples per second. This is just a simple sinewave, if you take two sample per sinewave, how are you going to reconstruct the original sinewave. How would your dac know that its supposed to construct a sinewave instead of say a sawtooth wave or squarewave? You just got two values!!!
This doesnt mean i agree that 24/192 should be a lot better. Atleast for higher frequencies, theoretically it should be a lot better. for 20khz, 44.1 sounds like really low to me, so 192 should do a lot better. I personally havent compared both on the same album.
I read some guys comparisons, atleast as per him there are hardly any differences. At high volumes, with low volumes parts of the songs there is slightly more distortion in 44.1 due to the high noise floor of 44.1

 

[thatguy]

Gave the article a quick read. For those who didn't read the article, here is a synopsis (my comments start with **). I will stick to the sampling rate part only since the bit depth has been discussed in other threads:

His premises:

1. Human hearing is limited to 20Hz-20KHz. Although empirical, the limits hold good.

** some of the forum members believe they *hear* frequencies above 20KHz. For them the conclusion will be invalid.

2. All signal content under the Nyquist frequency (half the sampling rate) is captured perfectly and completely by sampling.

** Yup. True. Too late to argue about the validity of the sampling theorem.

3. Audio systems have limited frequency range. When fed with signals outside this range, they will produce distortions. Ultrasonic components, when outside the equipment range can produce distortions in the audible range.

* Don't know.

His conclusions:

1. Music sampled at 192KHz will have ultrasonic components that our equipment may not be designed to handle. The output may be inferior to the one at 44.1 or 48KHz.

2. Oversampling a band limited signal (20Hz-20KHz) is good, because it allows the anti-aliasing filters to do their job better. Besides it doesn't have the ultrasonics. After the anti-aliasing, you can down-sample to 44.1 KHz without loss of quality.

** Don't know.


** Come to think of it, the guy isn't unreasonable. He even has some advice for the 'I trust my ears' types: "you can trust your ears. It's brains that are gullible."

 

[sound1]

I didnt read the full article as its pretty long, but some of the items sounded like bull to me. Author claims that once you are above nyquist freq, you can reconstruct the full analog waveform without any distortion or smoothing issues. This just beats physics.
Take 50hz for example, you just need 100 samples per second. This is just a simple sinewave, if you take two sample per sinewave, how are you going to reconstruct the original sinewave. How would your dac know that its supposed to construct a sinewave instead of say a sawtooth wave or squarewave? You just got two values!!!
This doesnt mean i agree that 24/192 should be a lot better. Atleast for higher frequencies, theoretically it should be a lot better. for 20khz, 44.1 sounds like really low to me, so 192 should do a lot better. I personally havent compared both on the same album.
I read some guys comparisons, atleast as per him there are hardly any differences. At high volumes, with low volumes parts of the songs there is slightly more distortion in 44.1 due to the high noise floor of 44.1

AFAIK, this works based on a few assumptions.

1. You have filtered out all the frequencies higher than 50Hz before sampling with 100 samples/second (to prevent aliasing)
2. The filter is an ideal LPF -- meaning it removes everything higher than 50Hz
3. Same ideal LPF is applied to the signal coming out of the DAC (all DACs already have these filters)

A sawtooth or a square wave has infinite number of frequencies (i.e. infinite harmonics of 50Hz fundamental tone). A quick glance over Fourier Series will give you a better idea: Fourier series - Wikipedia, the free encyclopedia.

So under ideal conditions, a 50Hz low pass filter should output a 50Hz sine wave regardless of whether the input is 50Hz sawtooth or 50Hz square wave. It is definitely possible and we have done lab experiments in college to confirm it (okay, sine wave was little distorted). Which means 100 samples are enough to represent a 50Hz signal

Since such steep filters are impossible to implement, we have chosen 44,100, 48,000, and so on (instead of 40,000 samples/second) to prevent aliasing. According to the article, 44.1k or worst case 48k is enough to account for non-ideal filters and higher sample rates are redundant.

Of course all of the above is theory and I am not sure if there are other practical reasons for choosing higher sample rates (eg. mixing, mastering etc. -- same reasons we shoot in RAW instead of jpeg).

 

[kaushik]

point here is we will believe or not ?

for me this is Simple , when some rich guy [will be expensive for sure ,with the speakers capable of doing this somehow ]does setup for 24/192

I will do a guest visit and evaluate...more likely will not impressed as improvement in SQ vs money spent ratio will be unreal.

 

[Gerry_the_Merry]

The same article has set fire to Audio Asylum. Generally, what cant be disputed is that at 16/44.1, you have two options. Go for better equipment, which will certainly lead to sound improvement. Do a better CD Player, DAC etc. Second, go for higher sampling rate and bit depth. With so much music there days available for free, this is an automatic no-brainer. No cost involved in determining if it is better or worse to increase the 16 or the 44.1.

 

[Thad E Ginathom]

This is a really great article. Many thanks.

I'll give it lot more attention when I don't have to recover from a dead hard disk and console myslelf over what's lost from the last backup.

However, just in passing, I wonder why this is so seldom heard in the "audiophile" community:

Phew! That's just great to know that those biger numbers or fattter cables or more expensive components or whatever are not worthwhile!

 

[Gerry_the_Merry]

Doors666: note that all frequencies are sampled at 44.1. It is just that 44.1 is sufficient for upto 20khz. It doesn't mean that sampling rate drops with frequency. So no risk of sawtooths.