Any science to explain the “weight” or “3D depth” of hardware audio vs software that some people claim?

[ghettosynth]

...

[Unaspected]

All of which can be emulated.

Absolutely.

It can all be emulated but the overlord will always be the CPU limit. So corners have to be cut as the trade off between subjective quality and objective efficiency is balanced. Then it is how they are emulated or whether the developer feels they need to be emulated at all.

Certainly, one could address inductance in a general sense but there is no point in tackling the immensely complicated subject of electromagnetism. Then you still have all these components interacting in sympathy with each other - with potentially very complex relationships. This makes it better to generalise, approximate and address the state of portions of circuits as a whole.

For example, if the result is system noise, then why not simply provide this via sample playback or a procedural method that uses 2nd generation DSP. Though it is only more recently that plugin developers have started to include noise options.

First, I would separate "weight" and "3D depth" - they suggest different properties. For me, "weight" suggests some kind of saturation at play; whilst "3D depth" suggests instability in pass bands - ripple and phase - and modulation around certain frequencies. In both cases we would probably be looking at frequency dependant situations.

Of course "weight" and "3D depth" are subjective terms so I doubt there are any papers written that address any science here.

[GRUMP]

Quite the opposite.

I have to insist on what I said, althought I´d say that modern SW will frequently even sound better - simply because you have other options to shape the sound. We´re running in that old discussion again, but I´d recommend to consider the listening environment and conditions as crucial in this case.

When I activate ReHead or an other correction SW the problem areas in the highend disappear or fall back. You can´t hear no difference over monitors or in a mix. But with AKGs on and medium volume - sound solo - my ears are still capable of identifying several differences.

The Virus is a quiet extreme excample. It´s absolutely clear and present. You shouldn´t compare a toybox with its character and not every HW synth had/has this pristine sound. But I compare a lot (samples) and although I´d frequently not vote vor the HW in the end I have learned a lot about several differences and weaknesses of SW. In the frequency / balance domain as well as in 3D. Width and depth.

To make a long story short: n0body will ever hear that difference in a track via smartphone. But as I am completely focussed on sound design I feel that my hearing abilities have improves a lot over the last years

[BONES]

If that's true, you are kidding yourself. For a start, you are comparing a samples so the signal chain and recording device will play a role in the final sound. And the fact that you have to work so hard to notice any difference at all means that those differences are completely irrelevant. If you can't tell whether an artist has used a MiniMoog or The Legend in a completed song, and by your own admission you would not, then any differences between the two are totally irrelevant.

I think a lot of the reason people think their hardware sounds better is that the signal chain they use is better than that for their software. But if you patch your hardware through your computer interface, just like your softsynths, I think you'll find that those synths you were sure sounded better, suddenly sound a lot more like your softsynths.

It can all be emulated but the overlord will always be the CPU limit. So corners have to be cut as the trade off between subjective quality and objective efficiency is balanced.

That may have been a consideration/limit in 2005 but Moore's Law took care of it many years ago. When a developer is able to model the difference in sound created by using US AC power or EU AC power, as Synapse have done with The Legend, then you can be pretty sure the rest of it is as close to perfect as anyone could possibly want or need.

Certainly, one could address inductance in a general sense but there is no point in tackling the immensely complicated subject of electromagnetism.

Don't be absurd. We are not talking about vacuum tubes and discrete transistors soldered to a circuit board, we are talking about Integrated circuits, chips that look like a CPU or a piece of RAM and fit into a socket on a circuit board. This, for example, is what the analogue guts of an Uno Synth look like -

For example, if the result is system noise, then why not simply provide this via sample playback or a procedural method that uses 2nd generation DSP.

Because noise is undesirable and the engineers who created those old synths would have busted a gut trying to get rid of it. They'd be laughing their arses off that any idiot would want it put back in as a sample.

Though it is only more recently that plugin developers have started to include noise options.

I think that's because they have only recently realised just how completely stupid certain segments of the market really are. Studio One has "vintage" options, that add cross-talk and noise into their mixer. I've owned half-a-dozen mixers over the years and each time I have upgraded it was to get rid of cross-talk and noise. Now there are crazy people out there who want to put that shit back in! It's utter insanity.

First, I would separate "weight" and "3D depth" - they suggest different properties. For me, "weight" suggests some kind of saturation at play

Which is precisely the problem with idiotic terms like "weight". To me, weight is 384 saw oscillators in a single ArcSyn patch, something no ancient piece of garbage could hope to match. Saturation provides grit and presence, not weight, and can be easily added to any sound.

whilst "3D depth" suggests instability in pass bands - ripple and phase - and modulation around certain frequencies. In both cases we would probably be looking at frequency dependant situations.

To me, 3D depth is a Dolby Surround set-up, nothing you can get from any ancient synth with a single output channel.

[Richard_Synapse]

If that's true, you are kidding yourself. For a start, you are comparing a samples so the signal chain and recording device will play a role in the final sound. And the fact that you have to work so hard to notice any difference at all means that those differences are completely irrelevant. If you can't tell whether an artist has used a MiniMoog or The Legend in a completed song, and by your own admission you would not, then any differences between the two are totally irrelevant.

And quite a few of those artists have already bought our plugins, making The Legend in fact more likely, or Dune 3 for the EDM crowd. This is another important point- the music you hear today is largely digital/software-based, people have moved on long ago.

That doesn't mean us developers cannot learn from vintage/analog hardware, it is of course fascinating. But it's definitely irrelevant for professional musicians, from a practical perspective.

Richard

[DaveClark]

Hi all,

1) In reading this thread, I see that perpetual3 is repeatedly told that he (I assume “he”) must provide proof of his claims about differences between left and right channels resulting in a more 3-D type of sound, etc. These responses appear to be based on confusion between the burden of proof in a debate taken on when it starts and the responsibility of proving any further claims as the debate proceeds.

Based on the numerous well-known tricks for enhancing the 3-D nature of a sound by introducing differences between left and right channels, perpetual3 is, in my opinion, clearly entitled to the presumption in a debate of the subject. Not only that, but it’s a natural presumption, not an artificial one due to convention. We have two ears, we hear in stereo because of differences, left and right channels, etc. Hence, and contrary to the reaction of some posters here, no response or justification is actually required until a reasonable counter-argument is advanced. This hasn’t happened. See (*) below.

2) As perpetual3 indicated, with two-channel hardware it’s almost impossible to create a truly monophonic sound due to differences between hardware components. With software, it’s the opposite. Thanks to digital technologies, channels can be and often are processed completely identically.

Expanding a bit: With both users and developers often not taking time to create differences between channels, there should be a statistical bias in software as a whole towards monophonic sounds for monophonic sources. With two-channel hardware, there should be a bias away from monophonic sound for the same sources. Similarly, there should be a bias towards preserving the stereophonic sources exactly as they are for software processing and a bias towards increasing the stereophonic characteristics for stereophonic sources in the case of two-channel hardware.

This seems to be a plausible explanation of why “some people” may actually hear a difference as the original post stated and with which others have concurred.

--------------------

(*) David Zarefsky. Argumentation: The Study of Effective Reasoning. The Great Courses lecture series. Now in second edition. Zarefsky has also recently written a book on the subject (Cambridge University Press, 2019). The presumption is the position that would prevail, lacking controversy. In American jurisprudence, for example, presumption is with the defendant in criminal cases and with the respondent in civil cases. In science, presumption is with the consensus of opinion of recognized experts in the given field. The burden of proof rests with those who take positions that oppose the presumption. A burden of rejoinder is not triggered by a request for support for the presumption.

[mystran]

The presumption is the position that would prevail, lacking controversy. In American jurisprudence, for example, presumption is with the defendant in criminal cases and with the respondent in civil cases. In science, presumption is with the consensus of opinion of recognized experts in the given field. The burden of proof rests with those who take positions that oppose the presumption. A burden of rejoinder is not triggered by a request for support for the presumption.

This is where you go on the wrong path. You see, in criminal justice you are trying to make a binary decision: guilty or not guilty. In science, we have a ternary choice: true, false or unknown. In science, the presumption is "unknown" until proven otherwise. In science, there is nothing wrong with presenting a hypothesis and then admitting that you don't know if it's actually correct, but in science you don't make claims one way or another (at least not without clearly labelling them as hypothesis) unless you have something to back it up with.

I want to further clarify that I actually do believe that the presented hypothesis is probably correct to some extent. I do suspect there's more to it, probably something to do with not just inter-aural but also temporal variation (since entirely monophonic sounds can seemingly have "depth" to them as well), but I do believe inter-aural differences might contribute. However, the keyword here is believe because without hard evidence one way or another it basically amounts to a claim that a god exists, which has absolutely nothing to do with science.

I also find the suggestion to "just Google" highly offensive, because I've personally spent far too much time doing literature search in order to try to find some actual studies on the subject without finding much substance. This obviously does not mean that such literature does not exists, but it sure as hell doesn't trivially come up with a simple Google search.

[ghettosynth]

...

[BONES]

I'd suggest the presented hypothesis is irrelevant to the discussion, in that all of these things can be modelled and, therefore, are not exclusive to the domain of analogue or of hardware. I made a really simple plugin in SynthEdit that simply inverted the waveform in one channel, which has the effect on a stereo signal of cancelling out in the centre of the stereo field but preserving the signal at the extreme left and right edges. The effect it produced usually made it sound like the channel in question was coming from way outside the physical location of the speakers but it was a purely digital effect. It was actually quite freaky how wide it made the stereo sound with something like a ping-pong delay.

[whyterabbyt]

Hi all,

1) In reading this thread, I see that perpetual3 is repeatedly told that he (I assume “he”) must provide proof of his claims about differences between left and right channels resulting in a more 3-D type of sound, etc. These responses appear to be based on confusion between the burden of proof in a debate taken on when it starts and the responsibility of proving any further claims as the debate proceeds.

False premise. This is 'DSP and Plugin Development', a technical forum, focussed on peers discussing DSP development and sharing knowledge. When someone asserts there's a method of doing something, requesting that the source of that information be passed on isnt about 'burden of proof.' Its about respecting the information sharing this forum was basically created for.

The downright refusal to do anything except wriggle around trying to avoid passing on actual information is utterly at odds with the spirit of this specific forum.

[Zaphod (giancarlo)]

Whoever has solved a problem that makes him attractive in the market, will hardly share his work completely, because of market rules. If you know how to make the best lemonade and that is why you sell a lot of lemonade, you will probably share a lot, but you would not give the precise and complete recipe to everyone. This forum, on the other hand, is useful to solve parts of a problem, to solve problems that sometimes block a development and to share basic philosophies.

[whyterabbyt]

Lemonade buyer turned up in lemonade seller forum insisting its really easy to find publically-disseminated scientific information on the veracity of a proposed method of improving lemonade production, but weaselled out on pointing to that information.

[quikquak]

Lemonade buyer turned up in lemonade seller forum insisting its really easy to find publicly-disseminated scientific information on the veracity of a proposed method of improving lemonade production, but weaselled out on pointing to that information.

I know I'm going to quote this somewhere today.

[kamalmanzukie]

why am i reading this thread, every thing about it is pissing me off. what is this 'weight' supposed to be?

is it undefined? yes? then it doesn't exist

no? measure it and figure it out. but posting questions about it in these vague terms without even establishing what any of it is even supposed to be... what are you people doing with your life


i skip to page 10 and was almost certainly it had descended into farce, funny guy ironic spitballing, but i was wrong. tragic

[Zaphod (giancarlo)]

Lemonade buyer turned up in lemonade seller forum insisting its really easy to find publically-disseminated scientific information on the veracity of a proposed method of improving lemonade production, but weaselled out on pointing to that information.

Sure a user can decide if one lemonade is better than another, but often technically they have no idea about the production process.
Lemonade sellers, in turn, judge everything by the production process that they now consider an established standard and used by everyone equally, because they have bought all the standard and best tools on the market that are usually used in production.
A bit like the mixing engineer: after a few years he can no longer hear a song as a normal user, but he breaks it down at the first listen and judges it with a "it has a bass sample that everyone has already used and the mixing is not so good". But maybe that song for the user of the street is fantastic and revolutionary because the melody line is beautiful and different from the usual.
I think the truth is in the middle. It is undisputed that some lemonades are perceived as better and superior to the sum of the elements. If one goes to break it down, it's still lemon, water and sugar, and at limit some acidity corrector. So someone tries to describe some pleasant lemonades as "fruity", and the sales people do not understand the language. Should more lemon be added? Do we need to change the sugar? Because they have machines to calibrate, and all they want to understand is whether they need to change the recipe and how, or whether the user is inventing unscientific terminology to describe the same old recipe.
So one day someone is looking for a way to translate the word "fruity" into a recipe. And it fails, because "fruity" in the context of a production process means nothing.
Someone timidly points out that small details enhance what lemonade is, a squeeze of fruit.

When the man in the street tries to describe a sound as 3d, he is actually comparing the sensation to having an extra dimension. Taken from any source, the extra dimension can be anything. For example, time. If a sound varies in time in a musical way, then that new dimension is born that he cannot describe in words. For the developer, perhaps, it is a trivial amplitude modulation. A variable harmonic distortion. Or an unpredictable phase shift in stereo content. The additional dimension we are talking about in this context is almost certainly time, but you can understand that it is such a general description that it is difficult to understand exactly how everyone should interpret it.
And the solution is simple: some lemonades simply seem to have a better balance, and some will say they are fruitier than others.