Bit resolution & Sample Rate explanation

[gypsyhick]

Howdy,

I'm posting a quick explaination for the digital jargon that one constantly hears: Bit Rate Resolution & Sample Rates. This post is intended for all who want a better understanding of what these terms actually mean. I recently read up on these definitions to be more informed in the misleading marketing world and to better understand the more esoteric posts on this site. All audiophiles and techies, please tweak this post or add some more insight.

Digital Audio is measured in two ways: Bit Rate Resolution and Sample Rate.

Sample Rate is the number of samples per second taken of the audio. 1Hz = 1 sample per second. RedBook CD (RBCD) is 44.1kHz = 44,100 samples per second.

Bit Rate Resolution is the digital word length of each sample. This determines how precise each sample is. Think of it this way. You can explain last night's concert more precise using 100 words rather than using 20 words. RBCD bit rate resolution is 16bit which gives a 65,535 word length. If a certain sample takes more than 65,535 words to describe than the sample will lose some information which results in distortion. The precision difference with bit rate is noticeable when you realize that most internet audio streams are at 8bit and digital telephones are at 4bit. Background noise is associated with lower bit rates.

So......If one second of audio is analogous to the view of your backyard, the RBCD digital representation of your backyard would be akin to looking at it through a screen door with 44,100 holes in the screen and 65,535 words or colors within each screen hole.

Hope this helps

GH

[RobotCzar]

You explanation is not helpful because you fail to relate the two variables to what we hear.

What are the audible ramifications of sampling frequency and sample size?

The words "resolution" and "precision" are not useful in this context. (How does something sound more precise?)

Those who do not know how sample size and frequency relate to sound should not be offering up opinions in this forum (or rather they should not waste our time by offering up their opinions.) Such people should go read a primer on digital audio (by somebody who never wrote for Stereophile).

[gypsyhick]

I believe you have misinterpreted the purpose of the post.

I wanted to explain what bit rate resolution and sample rates mean. The post was not intended to editorialize the "audio ramifications" of these measurements. There is no need. These forums already contain numerous subjective and objective views on digital sound quality. But what I did not find easily was a clear explanation of what these terms mean. So, I looked it up and posted it for those that browse this forum that are not as technically refined.

I am not going to argue semantics. If you like to, go to a philosophy forum.

Those who are supercilious should not be offering judgements in this forum (or rather they should not waste our time judging others.) Such people should go into politics.

You explanation is not helpful because you fail to relate the two variables to what we hear.

What are the audible ramifications of sampling frequency and sample size?

The words "resolution" and "precision" are not useful in this context. (How does something sound more precise?)

Those who do not know how sample size and frequency relate to sound should not be offering up opinions in this forum (or rather they should not waste our time by offering up their opinions.) Such people should go read a primer on digital audio (by somebody who never wrote for Stereophile).

[woodman]

Not helpful enough gypsy! As a wise man once said ... "paper will hold still for ANYTHING! It follows that the same thing holds true for the internet, of course. What's happened to you my friend is that you went looking for the definitions to some technical terms on the internet ... a place where any "info" that's posted should be suspect, since the web is a place where anyone can post anything they choose, and present it as fact for anyone that stumbles across it.

In the case at hand, you stumbled across someone's attempt to define things that they themselves have but a rudimentary understanding of. To wit:

Digital Audio is measured in two ways: Bit Rate Resolution and Sample Rate.

Almost, but no cigar awarded. The word "resolution" here is misplaced, since Bitrate is what creates "resolution". "Sample Rate" is a correct specification, as is "Bitrate" for Digital audio.

Sample Rate is the number of samples per second taken of the audio. 1Hz = 1 sample per second. RedBook CD (RBCD) is 44.1kHz = 44,100 samples per second.

Got that one right!

Bit Rate Resolution is the digital word length of each sample. This determines how precise each sample is. Think of it this way. You can explain last night's concert more precise using 100 words rather than using 20 words. RBCD bit rate resolution is 16bit which gives a 65,535 word length. If a certain sample takes more than 65,535 words to describe than the sample will lose some information which results in distortion. The precision difference with bit rate is noticeable when you realize that most internet audio streams are at 8bit and digital telephones are at 4bit.

The wheels tend to come off the buggy here. It starts out right (except for the inclusion of the word "resolution" again) ... then falls flat on its face by mixing up and equating "bitrate" and "sample rate" when trying to explain A-D conversion precision and accuracy. The "bitrate" is indeed 16 bits, which is also the "word length". The 65,535 figure is not "words" but "values" (of signal voltage).at the time of the "sample" ... thus the "resolution" of the RB-CD format. Lastly:

So......If one second of audio is analogous to the view of your backyard, the RBCD digital representation of your backyard would be akin to looking at it through a screen door with 44,100 holes in the screen and 65,535 words or colors within each screen hole.

This analogy doesn't serve the purpose very well at all ... just muddies the water rather than clarifying it ... sorry. I admire your desire to help others, but in order to do so, you need to either understand the technoligies yourself, or you need to find more accurate sources of "information" to quote from.

[IsmaVA]

Insterad of just pointing the errors, I will try to clarify. I am sure this is not perfect and further corrections will probably follow.

Sampling rate is how many times per second is the original signal in "measured". A 1Hz sampling rate means that the signal was "measured" once every second. For the redbook standard of 44.1Khz it means that the value of the signal is determined 44,100 times every second. A good analogy is a movie versus a comic book; with a comic book you get to know what happens at each square, but not in between. With a movie you get much more of the action.

sample resolution is about the precision (in its scientific meaning) of each one of those measurements taken every second. A 2-bit resolution means that the diference between the maximum value and the lowest value is divided in 4 equal portions. So, if the maximun value is 3 and the minimun value is 0, a signal value of 2.2 would be registred as a 2, just like a signal value of 1.8. Since 2.2 is definitively not 1.8 an error is introduced by this lack of resolution. For a 16-bit resolution the error is much more smaller.
Comapring low sample resolution to high sample resolution is like comparing a common ruler with a micrometer. While the ruler might tell you that something measures 2.6 centimeters, a micrometer can tell you that its actual length is closer to 2.573 cm.

Increasing the sample rate improves the transient response, or the capacity to match changes in the original signal.

Increasing the resolution improves the quality of each sample, by being able to more accurately represent the value of the signal at each sample instant.