End of the rainbow or am I just short-sighted?

DLC86 said:
I think one of the next steps in Fas (r)evolution will be to make these algorhythms work at higher sample rates in the next hardware platforms. More and more hi-res music is coming out these days and at some point in the near future 96khz processing will be considered a standard and a must. Probably it already is a must in professional environments.
Click to expand...
Higher sampling rates don't give you more "resolution." Sampling rate determines the maximum frequency you can record—nothing more.

48 KHz is more than fast enough to capture audio with complete accuracy, up to 20 KHz. Only babies and small children can hear 20 KHz anyway. Faster sampling, at best, contributes no audible difference. At worst, it adds to the noise floor. And it always increases resource usage in processing devices.
 
Rex said:
Higher sampling rates don't give you more "resolution." Sampling rate determines the maximum frequency you can record—nothing more.

48 KHz is more than fast enough to capture audio with complete accuracy, up to 20 KHz. Only babies and small children can hear 20 KHz anyway. Faster sampling, at best, contributes no audible difference. At worst, it adds to the noise floor. And it always increases resource usage in processing devices.
Click to expand...
48khz is enough to not have aliasing in the audible region but it's not "complete accuracy".
For example, take a 12khz sine wave, there will be 4 samples to represent a full cycle and (assuming the wave is perfectly synced with the clock) these will be respectively at 0, upper wave apex, 0 again, and lower apex. Now in this situation, how can you distinguish a sine wave from a triangular wave? They would be represented by the same 4 bits, the only way to make them discernable is to double the sampling rate.
If human ear is able to hear these differences at 12khz is another task though, and probably subjective. I can definetely hear the difference between 48khz and 96khz recordings but I don't know if I'd be able to discern it in a "guitar only" track.
 
Last edited:
DLC86 said:
48khz is enough to not have aliasing in the audible region but it's not "complete accuracy".
For example, take a 12khz sine wave, there will be 4 samples to represent a full cycle and (assuming the wave is perfectly synced with the clock) these will be respectively at 0, upper wave apex, 0 again, and lower apex. Now in this situation, how can you distinguish a sine wave from a triangular wave? They would be represented by the same 4 bits, the only way to make them discernable is to double the sampling rate.
If human ear is able to hear these differences at 12khz is another task though, and probably subjective. I can definetely hear the difference between 48khz and 96khz recordings but I don't know if I'd be able to discern it in a "guitar only" track.
Click to expand...
For the record, four samples is not the same as "4 bits." Each sample includes the full bit depth.

Now let's look at that 12 KHz triangle wave. It's made up of a 12 KHz fundamental plus a series of harmonics. The lowest harmonic of 12 KHz occurs at 24KHz. That's outside the audible frequency range. In other words, the human ear is incapable of hearing the difference between a 12KHz sine wave and a 12 KHz triangle wave. They both sound identical.


Edit: Looking at it another way, a 12 KHz sine wave and a 12 KHz triangle wave both have equal content in the audible spectrum. The differences between them only exist above 20 KHz—outside the audible spectrum.
 
Last edited:
Rex said:
For the record, four samples is not the same as "4 bits." Each sample includes the full bit depth.

Now let's look at that 12 KHz triangle wave. It's made up of a 12 KHz fundamental plus a series of harmonics. The lowest harmonic of 12 KHz occurs at 24KHz. That's outside the audible frequency range. In other words, the human ear is incapable of hearing the difference between a 12KHz sine wave and a 12 KHz triangle wave. They both sound identical.


Edit: Looking at it another way, a 12 KHz sine wave and a 12 KHz triangle wave both have equal content in the audible spectrum. The differences between them only exist above 20 KHz—outside the audible spectrum.
Click to expand...
Just one correction the lowest harmonic would be the 3rd... Triangle waves like square waves are odd harmonics only... Which is the fifth an octave up. 36k.
If you want a test wave that has an even order harmonic at 24 k you'd use a sawtooth wave
 
Ed DeGenaro said:
Just one correction the lowest harmonic would be the 3rd... Triangle waves like square waves are odd harmonics only... Which is the fifth an octave up. 36k.
If you want a test wave that has an even order harmonic at 24 k you'd use a sawtooth wave
Click to expand...
...and you are correct.
 
1. Amp in room sound/emulation
2. Ir modeling. Not tons of irs but build your own within axe
3. Amp modeling. Including power amps. Build your own perfect amp/choice. And as I understand power amps are baked in to the amps by cliffs choice.

E.g. Triaxis has vht power amp baked in not a mesa power amp.
 
Rex said:
For the record, four samples is not the same as "4 bits." Each sample includes the full bit depth.

Now let's look at that 12 KHz triangle wave. It's made up of a 12 KHz fundamental plus a series of harmonics. The lowest harmonic of 12 KHz occurs at 24KHz. That's outside the audible frequency range. In other words, the human ear is incapable of hearing the difference between a 12KHz sine wave and a 12 KHz triangle wave. They both sound identical.


Edit: Looking at it another way, a 12 KHz sine wave and a 12 KHz triangle wave both have equal content in the audible spectrum. The differences between them only exist above 20 KHz—outside the audible spectrum.
Click to expand...
Yeah sorry, I meant the 4 samples would be represented by the same sequence of bits.

Anyway, the way you've put it in terms of harmonics led me to think you're right.
But what if for example that 12KHz is out of sync with the clock by 45°? In that case a sine and a triangle wave would be represented differently: at 45° the sine wave amplitude is 0.707x of its max value while the triangular is 0.5, and a square wave would be 1.
Or do anti-aliasing filters also take account for that?

And how is it possible that I (and many others) clearly hear a difference between 48KHz and 96KHz recordings? And I wonder why 96KHz (or even higher) music and bluray movies are sold at this point, there should be some difference..
 
Last edited:
And now that I think about it, I think there could be more when it comes to modelers' sample rate. A heavily distorted guitar signal can have lots of overtones outside the audible range and these could affect for example how much current a tube draws in an amplifier because the signal it receives would contain more energy compared to the same signal filtered at 20KHz, effectively changing how the amp reacts. I think this is the reason why the amp block in the axe fx uses oversampling.
It could be a noticeable improvement if this will be extended to the fx blocks too (drives in particular).

EDIT: in the while I found these 2 interesting posts by Cliff:

http://forum.fractalaudio.com/threa...pdif-output-other-than-48k.45984/#post-600999

http://forum.fractalaudio.com/threa...equencies-than-48khz.76515/page-2#post-965949
 
Last edited:
DLC86 said:
But what if for example that 12KHz is out of sync with the clock by 45°? In that case a sine and a triangle wave would be represented differently: at 45° the sine wave amplitude is 0.707x of its max value while the triangular is 0.5, and a square wave would be 1.
Or do anti-aliasing filters also take account for that?
Click to expand...
If you filter the signal to eliminate all content above half the sampling frequency, the resulting output waveform will be identical to the input waveform, unless you've intentionally processed it to be otherwise.


DLC86 said:
And how is it possible that I (and many others) clearly hear a difference between 48KHz and 96KHz recordings?
Click to expand...
Time for a double-blind test with everything identical but the sampling rate. :)
 
  • Like
Reactions: Rex
You must log in or register to reply here.
Back
Top Bottom