[Anderton]...wishing the problems weren't there in the first place doesn't make them go away; raising the sample rate does.

 

[Bitflipper]This implies that aliasing is an inherent and unavoidable byproduct of DSP.

 
No, it implies that if problems exist, wishing they weren't there doesn't make them go away. 
 

Let's look at a practical example, an electric guitar played through a high-gain amp sim. You play a very high note on your guitar, say with a fundamental frequency of 1.3 KHz (an octave above an open high-E string). The amp sim will generate harmonics at 3x, 5x, 7x, 9x, etc. The 15th harmonic is 19500 Hz, still legal at 44.1 KHz. You have to get up to the 17th harmonic before changing the sample rate would deliver any benefit. I didn't do the math, but the level of the 17th harmonic is going to be down more than 90 dB from the fundamental. IOW, inaudible.

 
There's something I'm not understanding. If this is the case, then why would it be necessary to include oversampling and filtering if the problem they're intended to solve is inaudible? When I'm listening to amp sims, if you switch back and forth between high and low sample rates with some (not all by any means) processing, I can tell with 100% accuracy which is which. So something is audible...
 

If Z3ta+ falls into that category, then I maintain that it's a design flaw.

 
So do you consider including oversampling compensation for a design flaw, or simply good practice? If the answer is simply good practice and therefore it must always be done, then that restricts how many instances you can use and balance out CPU consumption. In some cases, there will be no audible difference whether the synth is loafing along at normal speed or oversampling, and many might choose not to oversample as a tradeoff for something like being able to overdub a guitar part with lower latency.
 
So the bottom line is this: if oversampling and filtering are good practices to compensate for deficiencies caused by generating undesirable harmonics that can create problems with lower sampling rates, why is recording at 96kHz inherently negative if it too can compensate for deficiencies caused by generating undesirable harmonics that can create problems with lower sampling rates? Aren't they the same thing, except one is applied as a global preference as opposed to a localized one?
 

Sample rate determines the highest frequency you can process. Period.

 
Doesn't it also influence latency? And I'm still wondering why the imaging on the TH2 reverb is better at 96k than at 44.1. My mind doesn't believe it's possible, but it's an audible difference. Maybe it's not due to a different sample rate per se but something associated with using a different sample rate? This difference existed whether I listened at 96 or sample rate converted it back down to 44.1 again. I have no idea why. Any theories?
 

[Anderton]Saying I need better plug-ins is all well and good, but that's sort of like saying all my plug-ins should be 64-bit, which they should be. But some aren't, so I have to use bridging, which is a flawed technology but serves its purpose.

 

[Bitflipper] Careful, let's not bring bit-depth into the discussion. It's irrelevant to the topic.

 
But it's very relevant to my point, which in this case is not about fidelity but about how wishing that a particular reality existed won't make it reality. In the case of 64-bit Studio One, not having a 64-bit plug-in would have a dramatic impact on fidelity - it would produce no sound, because you wouldn't be able to load it  
 

Now, I do understand and agree with your underlying argument, which is that you're stuck with the tools at hand, and a higher sample rate may improve the performance of some of them. I certainly couldn't afford to replace all of my plugins! So yeah, if I found out that most of my plugins were deficient I'd be seriously bummed and looking for any way to work around their limitations.
 
To put your premise another way, you could rephrase it thus: "if your plugins are substandard, a higher sample rate might or might not help". I'm good with that.

 
Then that's what you should say  What I'm saying that a higher sample rate definitely improves the audible performance of, at least so far, several plug-ins I've tried. In the case of plug-ins that achieve a doubled sample rate through oversampling and therefore double the sample rate internally without having to double the project sample rate, I'd go so far as to say that the vast majority of them sound better when oversampled. Some improvements are subtle, and sometimes the result is no audible improvement, but I do believe the cumulative effect of multiple small improvements can add up.
 
The AIR plug-ins, Native Instruments, WAVES, and a zillion other manufacturers offer oversampling. I don't think their motivation for including oversampling to double (or more) the sample rate is to compensate for being substandard products. But I don't design software, so for all I know maybe they do cut corners, then gloss over the deficiencies by doubling the sampling rate.
 
What I do know is this: I've found a simple way to make projects that rely on extensive use of synthetic processes sound better, so I can't think of any reason not to use it. I've told how I conducted some of the tests to come to this conclusion. Other people can try these tests and determine with their ears whether they hear an improvement. If they do, and I suspect some will, they can choose whether the CPU hit and fidelity improvement are sufficiently enticing to trade off file size and track count. 
 
 
 

Craig to me the 27 is the better sounding clip.
 

The biggest difference between Craig's files is that one drops off sharply above 18 KHz and the other does not. There was also a small level difference. After level-matching them and applying a LPF so that their spectra were more closely matched, the 96->44.1 file sounded more pleasing to my ear and I could not reliably distinguish between the two in an A/B test.
 
Goes to show that it's difficult to isolate a single variable and say with certainty "that's why it sounds better".

I am sidestepping the whole discussion - I am quite happy with the results I get with using 48 k, and I never have to worry about dropouts or otherwise taxing my system, in any way.
 
I will instead continue to concentrate on the quality and strength of a song's writing and playing, and I happily drive a Nissan Cube.  :)
 
Bob Bone

The plot thickens...
 
Actually, both Bitflpper and I might be wrong about 96 making a difference only with signals inside the computer. No less an authority than James A. Moorer wrote a paper that proposed, among other things, that hearing involves not just frequency and amplitude, but time and how it relates to localization when listening with both ears. He claims that most people can distinguish a time delay of 15 microseconds or more when a pulse is put into each ear, and that some people can differentiate delays as low as 3 to 5 microseconds. Given that a sample at 48kHz is about 21 microseconds and 10.5 microseconds at 96kHz, that means the minimum time delay most people can differentiate is actually less than one sample at 48kHz, but more than one sample at 96kHz. 
 
If this is the case, and of course the conclusion is controversial, then recording acoustic sources at 96kHz preserves localization information you don’t capture by recording at a lower sample rate, and which also won’t play back at lower sample rates. This has nothing to do with frequency response, distortion, aliasing, or any of those other characteristics but could explain why some people prefer 96kHz recordings yet are at a loss to explain why, because there’s no obvious audible change they can identify. Instead, they say that it sounds more “open” or “transparent” or [insert cork-sniffing word of your choice - I kind of like "pert, yet unassuming"].
 
I'm not saying this is right, wrong, or whatever...just putting it out there...hmmmmm. Whenever I think this stuff is too off-the-wall, I remember the interview in Guitar Player where Eric Johnson claimed different batteries sounded different in some effects. Of course, he was laughed at--"voltage is voltage, what an idiot." But he was only laughed at by people who didn't know about the poor power supply rejection of older effects, and how the internal impedance of alkaline and carbon-zinc batteries differ. 
 

Grem
Craig to me the 27 is the better sounding clip.

I understand that, a few of the non-oversampled Guitar Rig amps sound "better" to my ears than the oversampled ones. I guess this means that if the world goes to 96. someone will need to create an undersampling button to get "that vintage digital sound so popular in the 2010s"
 
But the point of the comparison is about the accuracy with which the sound represents what the synth was generating. (If I had any synth sound in a track with that many highs, I'd run for the QuadCurve's lowpass filter.) 28 has high frequencies that simply don't exist in 27, and they're not distortion byproducts.
 

Craig, my point of contention is extrapolating a general principle from a specific experience. I have seen no evidence nor heard any plausible explanation that might explain how higher sample rates might generally improve the quality of recordings. Oh, there's been plenty of speculation and the grasping of proverbial straws, but so far nothing science-based.
 
You are correct in saying that oversampling isn't there to gloss over product defects. Quite the contrary, oversampling is a key component of a quality design. Harmonics happen as a natural and unavoidable side-effect of distortion. Oversampling merely assures that they won't exceed Nyquist without resorting to aggressive filtering that might introduce its own audible artifacts.
 
 

bitflipper

Geo524
Speaking of recording at higher sample rates somebody once told me 88.2 is the best to record at simply because it's 44.1 doubled. They said the math that is computed is handled better at this sample rate? I don't know a thing about it. I'm not that technical when it comes to this sort of thing, but it makes sense to me. I record at 24 bit/44.1 but I think I'm going to try recording at the higher sample rates. Why not?

Urban myth, unfortunately. One of those concepts that sounds reasonable but just ain't so.

Actually I doubt you're old enough to know this, but that was true at one point. It's easy to forget that digital audio has been around for a long time, and that audio engines used to be 16-bit or [gasp] even less. When DAT was invented, it was deemed that the math involved in sample converting from 44.1 to 48 was so daunting (1.0884353741496599) that it would discourage digital copying. 
 
Quality sample rate conversion is not trivial, and the accuracy has improved dramatically over the past 30 years. This site is really interesting: 
 
http://src.infinitewave.ca/
 
Check out the difference in sample rate converters between Ableton Live 7 and Ableton Live 9, and that was only a few years' difference...then consider the days when we had 16-bit engines. Back then, conversions from 88.2 to 44.1 did sound better than 96 to 44.1. Fortunately that period didn't last long, but it did exist. 
 
Oh, and if you want to feel good about Sonar, while you're on that site compare it to a bunch of other DAWs. They used 8.5, but I assume the sample rate conversion didn't get any worse in the X-series.

bitflipper
I have seen no evidence nor heard any plausible explanation that might explain how higher sample rates might generally improve the quality of recordings.

 
I think you actually provided a very reasonable explanation - "oversampling is a key component of a quality design. Harmonics happen as a natural and unavoidable side-effect of distortion." Basically, all I'm doing is oversampling, and it's improving the sound quality of my recordings, many of which include saturation and distortion. Whether I should need to do oversampling, whether it should be a default in all plug-ins, or whether you should be able to switch it off if you're running on an old laptop are separate discussions. I have to work with today's set of tools and have no choice in that matter, but do have a choice of sample rates.
 
I'll leave the official explanations to others. Remember, I'm coming at this from the "wrong" direction anyway. I initially ran these experiments so I could have A-B comparisons that showed no audible difference and provide a voice of reason in case I ran into any "IT MUST BE 384KHz!!" zealots on the upcoming New Music Seminar panel. Instead, the 96k recordings not only sounded better, they even sounded better when converted to 44.1. Oooops. So then I had to figure out why. I'm not interested in being right or wrong, I'm interested in the truth.

bitflipper
The biggest difference between Craig's files is that one drops off sharply above 18 KHz and the other does not. There was also a small level difference. After level-matching them and applying a LPF so that their spectra were more closely matched, the 96->44.1 file sounded more pleasing to my ear and I could not reliably distinguish between the two in an A/B test.

It's interesting that you found the 96k file more pleasing to the ear. However the point was not to apply an LPF to match their spectra, but that the one recorded at 96k represented the high frequencies more accurately, even when played back at 44.1kHz. Hey, I just report, it's up to other people to come up with the theories 
 
If you listen closely to the two files on headphones you'll also hear non-harmonic noise in the upper ranges of the 44.1 file but not in the one that was recorded at 96. At first I thought it was because the additional high-frequency content was masking the noise, but I don't think so. The high frequencies are in a different range.