What The Matrix can teach us about “resolution” in digital audio

[o0OBillO0o]

@Kris.. I'm diggin' I saw that too.

[o0OBillO0o]

The the wiki article suggests as bit depth increases, there are more levels to represent the amplitude of a signal- i.e 16bit depth is about 65,536 levels.

However a generalization is that many recordings do not swing greater than 30dB either way (about 11bits) in amplitude. The question I'm asking is what is there to gain if recorded music is not using the other 5bits?

Bit depth speaks to how low the noise floor is.

Resolution is not what you think it is here. It's been redefined to something like: "A large amount of information per second." Unlike a video display's Definition of high resolution - A large amount of information per square inch on a display screen or printed form. Measured in dots per inch (dpi), the more dpi, the higher the resolution and quality. Screens are in the 70-400 dpi range whereas printers are in the 300-1200 dpi range.

I'm at my limit for the rest of the understanding of digital music, but sampling may help understand the other half.

http://en.wikipedia.org/wiki/Samplin...nal_processing)

[djwhog]

Originally Posted by Dom_P

What's actually kind of funny is that old laserdiscs were analog. And contained analog video as well as analog and digital audio sound tracks.

In hind sight maybe that was the future.

As I understand it, the pits and lands on a disc don't need to be constant size as the are on a digital CD (in this case used to the define ones and zeros). An analog signal can actually be encoded on the disc using variable sized pits and lands arranged in linear sequence. A laser can then pick then pickup that analog signal with no D/A conversion required. Pretty much like a turntable, but without the needle and associated wear, dust, static, etc. issues.

EDIT: a little further research on Wikipedia turned up this.


"An earlier analog optical disc recorded in 1935 for Licht-Tone Orgel (sampling organ)"

Yes and NO PC stupid slow MPEG compression. Yep was ahead of it's time and could be way better than anything we have now if tweaked a little.

[o0OBillO0o]

@Kris.. Basically, bit-depth is how far the noise is below the maximum signal, sample rate is the bandwidth from lowest frequency to the highest.

The author suggests that all you need is reduced noise and decent bandwidth, recorded at the right levels (i.e. No clipping), and the musician(s). That's "high resolution."

That's why lots of recordings sound great no matter the format. Although "noise" is purported to be bad, it's part of the art form, and noise is imparted by many objects in the signal chain. Sometimes on purpose. I.e Tubes, EQ, cables, and even speakers.

[House de Kris]

Believe me, o0OBillO0o, I'm not being argumentative here and picking a fight. My intent is to truly be part of the "Friendly Forum for High End Audio." I'm just following the suggestion of the OP, "proceed to debate and air out that horse."

Originally Posted by o0OBillO0o

However a generalization is that many recordings do not swing greater than 30dB either way (about 11bits) in amplitude. The question I'm asking is what is there to gain if recorded music is not using the other 5bits?

This is a good question, and comes up a lot in various forums in various incarnations. Put another way, your question is asking, "why have the noise of the carrier be any less than the minimum signal to be recorded?" To counter, I ask, "why is having the noise of the carrier being at the same level as the lowest amplitude expected to record acceptable?" If the goal is to satisfy the masses for just 'good enough,' then I suppose we are done. But, if the pursuit of audio is to obtain the highest fidelity possible, ANYTHING that can be done to reduce the detrimental effects of the medium should be applauded, in my opinion.

But, let's think about this a bit further. This dynamic range of music, which you just described as being generalized to 60dB does not merely consist of amplitude modulated sine waves. It is made up of individual sounds of instruments combined together. Each instrument has its own set of harmonics, which give each instrument its own unique sound. These harmonics are of an amplitude much less than the fundamental (typically). So, when the peak amplitude of some instrument's sound is -60dBFS, the energy of the harmonics are even further down. Now, it is true that we can hear below the noise floor. It is typically acknowledged we can hear 10 to 15dB below the noise floor. But, why this limitation? Why couldn't we hear 70dB, or more, below the noise floor? At some point the noise is going to dominate and obliterate the sound we desire to hear. It becomes a chore for our brain to decipher the signal of intent when swamped with noise. Hearing below the noise floor is possible, but it does take effort. Why not lower the noise and lighten the effort required of our brain to make sense of what confronts the ear? More bits than what is barely necessary is not a waste, in my opinion.

This was a long winded explanation that those "other 5bits" are indeed used, and important.

Heck, let's take this even further. I can remember back to when mechanically based carriers for the distribution and storage of music was king (what we call analog now [rekkids and tape]). The term resolution was used back then, when not a 'bit' was within sight. People talked of how a superior high-end system could resolve finer details and nuances of music. These high resolution systems were a joy to listen to, and were outside the reach of what John Q. Public could afford. Audiophiles with lower resolution systems would save up their bucks and upgrade their systems to enhance their system's resolution to hear a more life-like presentation of their beloved music. The path to these higher resolution systems was to obtain equipment that lowered the amount of undesirable signals contaminating the music signal. These undesirable elements are noise and distortion. Amazingly, we were able to discuss how well a system resolves the sound of a violin, and now Ian tells us there is no resolution. This is the issue I have with the linked article in the OP, which I thought was the point of this thread.

Originally Posted by o0OBillO0o

Bit depth speaks to how low the noise floor is.

This is true. Just as the noise level defines how well a system resolves the music signal of interest. I would then modify your statement to:

Bit depth defines the noise floor which defines the resolution.

[o0OBillO0o]

Kris... quick response. Good stuff! Keep it coming.

[House de Kris]

Originally Posted by o0OBillO0o

The author suggests that all you need is reduced noise and decent bandwidth, recorded at the right levels (i.e. No clipping), and the musician(s). That's "high resolution."

I'm not so sure that that is what the author is suggesting. How can he suggest that is "high resolution" when he repeatedly states there is no resolution. If there is no resolution, there is no high or low of it (resolution).

Originally Posted by o0OBillO0o

That's why lots of recordings sound great no matter the format. Although "noise" is purported to be bad, it's part of the art form, and noise is imparted by many objects in the signal chain. Sometimes on purpose. I.e Tubes, EQ, cables, and even speakers.

One of the great things about forums is that people of all different beliefs can discuss things. We don't all need to believe the same to have a conversation. Contrary to your beliefs, I don't feel that noise is part of the art of music. I can listen to live music, and noise of the venue is part of the package, that's a given. But, noise of the venue is not necessarily part of the art. When listening to a symphony and a large truck drives by the hall, I don't consider that part of the art. I consider that an unfortunate accident. Likewise, when that performance is recorded, any additional noise added to the performance is not part of the art. It, again, is an unfortunate accident. Noisey mic preamps, noisey random particles on tape, surface noise in grooves, quantization noise in digitization, noise in amplifiers driving speakers, noise from my neighbor's lawn mower when listening all come after the fact of creating the art and not part of the the art of music at all - in my opinion.

Oh, and all of these noises after the fact reduce the resolution.

[o0OBillO0o]

@Kris.. What's the most likely course of action? What's the most dangerous?

Most likely, the industry will continue to advertise and differentiate their product as high resolution. While low noise and high bandwidth are positive attributes worth marketing, both factors do not tell us enough information to make well informed decision.

Most dangerous, (and that's a relative term, this is audio after all) the industry will keep marketing larger bit depths and higher sample rates and further obfuscate where they get their recordings. Or at the other end advertise and all you can eat subscriptions of every song on the planet with a lackluster versions of the recording that is compressed to good enough levels for most to accept.

What you should be after is what you feel is the best* recording. If qualifying "best recording to you" involves format, bit depth, sample rate, or even bit rate, then so be it.


* I do believe there is a general consensus of a group of well recorded material of the same song/album that can be considered the best. There isn't a one rules them all "best" recording.

[o0OBillO0o]

Another one..

Why (Almost) Everything You Thought You Knew About Bit Depth Is Probably Wrong | SonicScoop - Creative, Technical & Business Connections For NYC

[House de Kris]

Originally Posted by o0OBillO0o

@Kris.. What's the most likely course of action? What's the most dangerous?

Most likely, the industry will continue to advertise and differentiate their product as high resolution. While low noise and high bandwidth are positive attributes worth marketing, both factors do not tell us enough information to make well informed decision.

Most dangerous, (and that's a relative term, this is audio after all) the industry will keep marketing larger bit depths and higher sample rates and further obfuscate where they get their recordings. Or at the other end advertise and all you can eat subscriptions of every song on the planet with a lackluster versions of the recording that is compressed to good enough levels for most to accept.

What you should be after is what you feel is the best* recording. If qualifying "best recording to you" involves format, bit depth, sample rate, or even bit rate, then so be it.


* I do believe there is a general consensus of a group of well recorded material of the same song/album that can be considered the best. There isn't a one rules them all "best" recording.

It seems to me that you are now changing the intent of the thread from discussing merits of the carrier (an objective discussion) to what sounds best to the listener (a subjective discussion). I believe these to be two independent and unrelated topics. I'm not sure anyone has suggested that what sounds best to someone can be determined by the number of bits or the number of samples per second.

I believe the industry will continue to offer new options as time goes on. Just as I believe the buying public has the right to ignore such options, or to embrace them. No one's got a gun to their head here.