Bit depth & Sample rates to 16 bit CD

[Full_Tilt]

Ive just started to get my head around some of these issues mainly bit rate sample rate the so called volume war during mix downs / mastering etc .
The research Ive done so far has opened up a pandora's box in terms of how little I thought I knew .
Am I correct in thinking that 24 bit at 44100 is more important than say 16 bit at 48000 sample rate ?
It seems like higher on both is obviously better but the sampling only benefits after improvements in bit depth......I suppose it boils down to the detail being there (bit rate) when the snap shot is taken (sample rate) .
Which leads me to the title of my post when you consider the benefits of burning CD's from a higher rez source recording.
Ive just got all these graphs and diagrams in my head but not seen what actually happens in graph form when this happens so as to see that difference between music recorded at cd quality put to CD and music recorded at higher rez put to CD .
Im not doubting there is some benefit but it would be nice to see it in black and white...... That graph im still waiting to see.

 

[colonel kurtz]

Am I correct in thinking that 24 bit at 44100 is more important than say 16 bit at 48000 sample rate ?

i don't know what you mean by "important"...but yes, i would say that a 24/44.1 file is going to be of higher sonic quality than if the same material was tracked at 16/48...or even 16/96 for that matter

 

[Full_Tilt]

Well in as far as "higher sonic quality" might be seen as important I suppose.

 

[TheDarkProject]

Am I correct in thinking that 24 bit at 44100 is more important than say 16 bit at 48000 sample rate ?

In pure mathematical terms, you're storing about 38% more data per second in 44/24 than you are at 48/16, which means a higher fidelity recording, which in turn permits higher fidelity playback.

In musical/psychoacoustic terms, here too the 24/44 format is likely to be better because the extra information you get from moving from 16 to 24 bit (higher signal to noise ratio) is generally more audible than the extra information you get from moving from 44100Hz to 48000Hz (about 2000Hz extra on the top end, mostly above the limit of human hearing).

On top of that, reducing bit-depth is less prone to aliasing than downsampling. You can chop off the least significant 8 bits with rarely any significant change to the frequency or amplitude, but dropping 39 out of every 480 samples to get from 48KHz to 44.1KHz is quite likely to cause some artefacts, so you have to interpolate between them. That in turn means that most of the other 441 out of 480 samples have to shift a bit to try and maintain the original curve... but you no longer have the original analogue curve, just a 24-bit digital approximation of it, so you end up losing information a second time when your newly interpolated sample gets quantised back to 24 bits. Blah blah blah you get the idea. Obviously good digital workstations have complex interpolation and dithering algorithms to make the most of this and reduce audible artefacts but the rule is that once the information is gone, there's no getting it back.

Obviously, if you're working with DVD-audio, then you might have 96KHz sound to deal with, and 48KHz upscales to 96KHz far better than 44.1KHz does, for similar reasons - all your existing samples can remain unaltered and you just have to interpolate to find the new intermediate samples. Simpler ratios make for less damaging digital conversions.

It seems like higher on both is obviously better but the sampling only benefits after improvements in bit depth...

If you had an application that involved recording fairly loud and pure sounds at frequencies of 15KHz or above, then you'd appreciate the extra sample rate more than higher bit depth. It just depends on what information you're capturing.

 

[abt]

One is the number of samples taken in a second (the sample rate) the other is the amount of data in each on of those samples (the bit rate). Multiply the sample rate by the bit rate and you'll get the total number of data bits per second, the higher the value the higher the fidelity.

Edit: Okay, this explains it better than I do http://manual.audacityteam.org/index.php?title=Digital_Audio

 

[Full_Tilt]

In pure mathematical terms, you're storing about 38% more data per second in 44/24 than you are at 48/16, which means a higher fidelity recording, which in turn permits higher fidelity playback.

In musical/psychoacoustic terms, here too the 24/44 format is likely to be better because the extra information you get from moving from 16 to 24 bit (higher signal to noise ratio) is generally more audible than the extra information you get from moving from 44100Hz to 48000Hz (about 2000Hz extra on the top end, mostly above the limit of human hearing).

On top of that, reducing bit-depth is less prone to aliasing than downsampling. You can chop off the least significant 8 bits with rarely any significant change to the frequency or amplitude, but dropping 39 out of every 480 samples to get from 48KHz to 44.1KHz is quite likely to cause some artefacts, so you have to interpolate between them. That in turn means that most of the other 441 out of 480 samples have to shift a bit to try and maintain the original curve... but you no longer have the original analogue curve, just a 24-bit digital approximation of it, so you end up losing information a second time when your newly interpolated sample gets quantised back to 24 bits. Blah blah blah you get the idea. Obviously good digital workstations have complex interpolation and dithering algorithms to make the most of this and reduce audible artefacts but the rule is that once the information is gone, there's no getting it back.

Obviously, if you're working with DVD-audio, then you might have 96KHz sound to deal with, and 48KHz upscales to 96KHz far better than 44.1KHz does, for similar reasons - all your existing samples can remain unaltered and you just have to interpolate to find the new intermediate samples. Simpler ratios make for less damaging digital conversions.


If you had an application that involved recording fairly loud and pure sounds at frequencies of 15KHz or above, then you'd appreciate the extra sample rate more than higher bit depth. It just depends on what information you're capturing.

Thanks TheDarkProject
So as far as CD burning & playback goes: is there a limit to what sampling rate can be captured If so what would that be ? would this also depend on the CD player ?
(Im not talking about the human ear im just learning about CD quality at this point) Sorry if this sounds ignorant .

 

[egan.]

Cd audio format is 16bit 44.1 always. If you're burning as data you can do whatever you want.

 

[TheDarkProject]

Yep. The reason you might capture at higher rates even when you know it's going to end up as 44/16 on a cd in the end is because you want as much information as possible - resampling issues aside - available at each processing stage. Almost any processing you do on digital audio loses information. The goal is to have enough information in the first place that by the time it comes to render it down to 44/16, you still have more than you need.