Vinyl v. digital curiosity

Seems reasonable. The difference between 24/16 will only be the noise-floor. While with different smpling rates the files are not perfectly the same.

 

Once it's burned to disc there's no difference in hi res vs the standard 16bit/48KHz. The only advantage to high bit high rez is at the time of capture during recording and post processing.

It's been posted several times over the years but it bares repeating:



You'll have to jump to "Bit Depth" at around 8:40

 

If we are just talking sampling rate Im pretty sure Ive seen that being tested multiple times to see if anyone can hear a difference. Like MQA CDs. Never seen one where it was audible.
Has there been one?

 

This guy doesn't use his ears.

 

I typically find it amusing when people on forums refer to the sample depth as in 16-bit vs. 24 bit preferring a 24 bit sample due to it being "higher resolution".

Whereupon I attempt to explain to them that the bit depth does not have anything to do with the "resolution" which is what the sample rate determines.

A standard Redbook 16 bit CD is capable of 96 dB of dynamic range. 24 bits can add another 48 dB for a total of 144 dB of dynamic range. Which is handy to have in a professional recording studio and being used by professional recording engineers.

Of course, in the real world, "The entire dynamic range of some types of music is sometimes less than 12dB. The recordings with the largest dynamic range tend to be symphony orchestra recordings but even these virtually never have a dynamic range greater than about 60dB".

So a 16 bit depth on a standard CD can handle about 50° more dynamic range that even a symphony is capable of producing.

 

Not quite sure what and if you're disputing what's been pointed out in Monty's xiph.org video.

Simply put bit depth is about precision at capture stage especially when A/D converters filter noise using that precision so as not to destroy detail. Once the primo filtered data makes it into the DAW further edits can be applied at high bit and sample rates without much further degradation. Higher sample/bit rates at playback on any consumer audio system can't be heard. It's just thinner slices of the same data that was on the source that gets played at blinding speeds so it's unnecessary overkill.

As a test zoom in on a CD standard waveform in Audacity at point node view and scribble on a couple of peaks and valleys then play it back selecting a broader selection of sine waves lets say 10 seconds of play. Do you hear anything? I don't. So what's the point of adding more samples on playback data that has already been processed to perfection at high bit/sample rate? It's too refined to hear at play back speed. The data is already there. More samples/bits aren't going to make a difference.

This also applies in digital imaging on cameras with 12-14 bit A/D converters. The extra bits above 8 are used to filter low end noise in shadows for shooting Raw (data straight off the digital sensor). On jpegs the data is processed with incamera software on the same A/D converted filtered data. Once the Raw image is loaded in the Raw converter at 16bits those extra bits are used for precise editing to map to an 8 bit video preview to prevent posterization in gradations.

 

You don't know what's causing what you're hearing because you don't know when noise has been filtered so precise that it makes the overall audio sound different. There's no way for you to prove what is affecting what you are hearing.

High bit/sampling data is overkill on playback. Anyone who says they can tell when noise reduction software as been applied to a finished CD would automatically think that because hiss and any other artifact that plays constantly over the entire performance acts like a vale. Some vales can be soft, cripsy or waxy but how would you know what's causing this sound difference unless someone told you it's due to high rez data and not noise reduction.

 

Well the Redbook is 16/44.1. So a CD isn´t high res; that is correct.

 

Not aware that I disagree.

 

"This quandry -- LP vs. CD, SACD, or DVD-Audio -- emerges from the fact that even today's state-of-the-art CD playback doesn't approach the sound quality offered by a good LP playback system. At the very highest level of music production, there's not even a debate: LP is musically superior to CD.

I must qualify that last statement: a high-quality, properly set-up LP playback system, playing a record in good condition, will sound better than any CD. If, however, the record is played on a low-quality or poorly set up system, the CD will usually offer higher sonic performance. The general public's perception that CD is vastly superior to LP (remember the "Perfect Sound Forever" marketing campaign for CD?) is perpetuated because very few listeners have heard high-quality LP playback. When done right, LP playback has an openness, transparency, dynamic expression, and musicality not matched by CD. There's just a fundamental musical rightness to a pure analog source (one that has never been digitized) that seems to better convey the music's expression.

This isn't to say that LPs are perfect. They suffer from a variety of distortions such as mistracking, ticks and pops, speed instability, surface noise, cartridge frequency-response variations, inner-groove distortion, wear, and susceptibility to damage. But for many listeners, these problems are less musically objectionable than the distortion imposed by digitally encoding and decoding and audio signal. Some listeners can hear past the LP's flaws and enjoy the medium's overall musicality. Other listeners can't stand the ritual of handling and cleaning records -- not to mention keeping the turntable properly adjusted -- and think CD is just fine. I think of it this way: LP's distortions are apparent, but separate from the music; digital's distortions are woven into the music's fabric. Consequently, analog's distortions are easier to ignore. If you're inclined to think CD is without fault, and you've never heard a properly played LP, give yourself a treat and visit a specialty audio dealer with a high-end turntable. Listen to what vinyl can do before you write off the possibility of owning a high-quality turntable."​

Excerpt taken from "The Complete Guide To High-End Audio" by Robert Harley.

 

Here's another excerpt from "The Complete Guide To High-End Audio" where he discusses why 44.1kHz sampling and 16-bit quantization simply aren't enough.

"Although the causes of the superior sound quality delivered by increasing the sampling rate are debatable, the benefits of the increase in word length from 16 to 18, 20 and even 24-bits are not in dispute. As I mentioned earlier, word length is the number of bits used to encode the audio signal's amplitude at each sample. Assigning a number -- called a word -- to represent the audio signal's amplitude is called quantization. The word length determines the system's resolution, dynamic range, distortion, and signal-to-noise ratio. We also call the word length resolution, as in the phrase "20-bit resolution." The greater the number of bits per word, the higher the resolution.

Think of a digital clock with only two digits; we could discern only between the hours, with no resolution of minutes. If we add a third digit, we can know the time with ten-minute precision (an error of plus or minus five minutes). Adding a fourth digit allows us to know the time with one-minute accuracy. A fifth digit to indicate tens of seconds further increases the clock's precision. Similarly, the more bits in a digital-audio word, the finer the precision with which we can encode the audio signal's amplitude.

The CD format uses 16-bit quantization, which provides 98dB of dynamic range. Although this number may sound high in theory, it's not enough in practice. Consider an engineer setting the recording levels for a symphony recording. He must adjust the signal driving the analog-to-digital converter so that at the end of the fourth movement when everyone is playing full-bore, the levels never exceed full-scale digital -- the maximum amplitude that can be encoded. At full-scale, all 16 bits are used. Now think about a quiet flute passage in the slow movement; the low signal level is encoded with perhaps four or five bits, which offers very low resolution. And if the recording engineer leaves 6dB of "head room" to account for unexpectedly loud peaks, he's effectively throwing away one full bit of resolution.

Thus the resolution of a digital audio signal isn't defined by the maximum number of bits available, but by the number of bits being used at any given moment. The lower the signal level, the lower the precision with which the signal is encoded."​

He also says:

"When the compact disc was developed in the late 1970's, 44.1kHz sampling and 16-bit quantization were chosen to achieve 74 minutes of playing time on a 120mm disc. The D/A converters of the day were limited to 14-bit resolution. Some argued at the time that a sampling rate of 32kHz with 14-bit quantization were sufficient for CD, but fortunately, the standard was established at 44.1kHz and 16 bits. These specifications are, however, not sufficient to encode all the musical information humans can hear."​

 

OK, I forgot standard CD is 44.100 cycles per second sampling rate. High rez is a relative term. But you flash in my ear a pulse sound cycling 44,000 times a second I'm not going to hear a drop in 100 cycles. 60Hz per second is what electrons cycle at on AC circuits. It's that hum you hear when a transformer blows out. The display picture you view on your computer is flashing a complete picture at 60 cycles per second continuously and yet there's no flicker.

 

Im pretty sure I dont agree. Although I cant say Ive heard the absolute best vinyl setup forged in the heart of a dying star. I wonder if he thinks that a digitally sourced record sounds best on vinyl too.

Then why are the differences never heard in tests?

 

missan said: ↑
Seems reasonable. The difference between 24/16 will only be the noise-floor. While with different smpling rates the files are not perfectly the same.

Well I wrote this, which You responded to. And I have no idea why You did.

 

So that author is disputing Monty's points.

Can you prove with an A/B sound comparison that not all musical information can be reproduced with current CD standards? I'ld like to hear it.

 

It would probably help folks understand your points if you communicated with more than quick one liners for such a complex subject.

 

Monty 'listens' to what ever the numbers on a screen tell him to hear.

He's basically stating that there is missing information in current CD standards due to the inherent nature of creating a digital "sample" of an analog waveform that's limited to only 16 bits.

 

I first heard the differences simply at high end audio stores. Not sure what you're referring to. The difference between CD and vinyl is like night and day assuming it's a decent pressing.

 

He's not making it clear if CD standard resolution is enough for encoding analog data as opposed to playback. I've already stated it's important to capture/encode analog data at high bit/sample rates.

Playback is determined by the DAC because once the data has been downsampled to standard CD resolution a high bit DAC on playback can't add more detail.

 

And you can't prove where in the complex chain of variables starting from capture to mastering to whatever DAC is being used for playback on top of other stuff in audio systems and what it does to sound to attribute to high resolution.

 

Are you talking about tracking in the studio? If so, when tracking digitally they'll often track at 24/96 or higher. However all of this is downsampled to 16/44.1 once it's burned to CD, losing a ton of information at this stage. I believe he says in the book that 24/96 files are vastly superior for playback than CD. I'll see if I can find it.

Agreed.

 

Well it was You that was responding, not folks, and I still don´t know why.

 

I quoted a specific part of your post, regarding the sampling not being enough for what we humans can hear in resolution or detail. Im asking where the people are who have shown that they can hear a difference between normal and super high resolution audio.

 

But does this author say what this lost ton of information sounds like? A/D converters do the same with internal filtering and a lot of that ton of data one would not want to listen to or consider as real information. Not all downsampling software to CD standard does it the same.

You should see the Image Size...downsampling dialog box in Photoshop and all the options provided to get different/best results on final image output. It's bewildering and impressive what these programmers come up with.

 

He's saying that it isn't sufficient for our own innate capability to hear. In other words, 16/44.1 doesn't encompass the full spectrum of sound for what we're able to sense. Also it isn't enough to capture a true representation of recorded music.