Analogue discussion

[zerocrossing]

No, there aren't. There are good reasons to keep sampling rate slightly above 44.1, but not even 96 makes much sense. The reason why "pros" use high sampling rate is mostly ignorance and marketing. Which is kind of whole point of Vojtech's posts in this thread, only he extends it to "analog gear" as well

I think you are misunderstanding the point of high sample rates in this context. It's not about the signal that's being output as much as it is for the calculation rate that's happening in the engine. If you don't believe me, spark up something like Reaktor that allows for a sample rate that's different than your project. Make sure the sample rate is the same as your project, which should be 44.1 or 48 kHz. Take an ens. Like Monark and use the 3rd to modulate the filter cutoff in anything but "low." Make sure the mod wheel graphic is up. Now, crank up the sample rate to 96 kHz and listen. There's a clear difference because you're forcing Reaktor to look at the modulations per sample at double the rate, and there is a clear difference. NI recommends running it at 96 kHz for this exact reason.

[steve2KVR]

...we cool!?.../s~

[zerocrossing]

I'm a fan of Melda... The future, etc. I'm also a fan of using what works and what feels right. Sometimes that's Mpowersynth, sometimes it's not. Like I mention in the last post, audio rate modulation is something that works better at high sample rates, (that can happen internally as it does in a plug in like Diva.) but it happens even better in the realm of analog electronics.

For instance, I'm trying to possibly further trim down my studio. I like to do as much as possible "in the box" as they say. I've got a Spectral Audio Neptune 2. It's basically a Swiss Minimoog in terms of it's architecture . So, naturally I think, "Well how does it stack up to Monark or Minimax for simple audio rate modulation?" The answer is, better. Pretty significantly better even when running at 96 kHz. The patch is super simple. One osc, filter half way, FM the cutoff. The Neptune just oozes rich harmonics where the software emulations sound a bit dull. My ATC-X does a good job too. Actually, the one synth that makes me think I could ditch the Neptune 2 is a hardware synth, but it's a digital VA. John Bowen's ProTone for the Scope system. Not quite as creamy sounding as the Neptune, but perhaps close enough, and in some ways it's a more confident sound that would work better in a mix. Feedback type sounds also don't do as well in software. Monark is the best I've come across, but it pales in comparison to what I can do with my cheap Bass Station 2.

So I think digital can do a lot of what analog does, but to get the audio rate modulation or feedback to sound really great, we're going to have to get to a point where we either can devote more processing power to it or figure out some tricks. Now, for your basic tones, sure there's lots of software that can do that, and plug ins like Mpowersynth can, of course, run rings around what the old Neptune 2, or pretty much any analog synth can do, but if you're looking for that meaty creamy analog sound, you're still best off with an analog synth. The Neptune 2, Bass Station 2 and ATCs can be had pretty cheap too, so it's not like you have to drop $5k for one. They go really nice with a synth like Mpowersynth.

[steve2KVR]

...yup...like the man said "love the one your're with".../s~

[Burillo]

I think you are misunderstanding the point of high sample rates in this context. It's not about the signal that's being output as much as it is for the calculation rate that's happening in the engine. If you don't believe me, spark up something like Reaktor that allows for a sample rate that's different than your project. Make sure the sample rate is the same as your project, which should be 44.1 or 48 kHz. Take an ens. Like Monark and use the 3rd to modulate the filter cutoff in anything but "low." Make sure the mod wheel graphic is up. Now, crank up the sample rate to 96 kHz and listen. There's a clear difference because you're forcing Reaktor to look at the modulations per sample at double the rate, and there is a clear difference. NI recommends running it at 96 kHz for this exact reason.

i know about the concept of "oversmapling", but the original point was in the context of "CD sound being not enough" and that "higher sampling rate is like HD television", both of which are complete bullshit. processing at higher sampling rates makes sense and more often than not even desired, especially for non-linear processes. recording and listening doesn't.

[IncarnateX]

To me, most of this analogue-digital shite can be reduced to psychology. We are mostly discussing the nature of stimulation in terms of objective facts such as the instability and randomness of sound as if it doesn't mean a thing how the brain processes these stimuli on basis of training and cultural context.

Thing is that the brain is not necessarily found of instability from a hard-wired perspective because it is highly specialized to search for stable patterns in the environment. We know that the brain searches for patterns, which makes sense from an evolutionary perspective since it is crucial for animals to recognize what kind of creatures you can eat and which ones who like to eat you. On the other hand we also know that the brain gets habituated to constant stimulation to an extent, where it stops reacting on it (e.g. the ticking of a watch –now you hear it and now you don’t). Thus there is no evidence that the brain “values” or “prefer” either stabile patterns or constant variations over one another. Sheer survival is highly dependent of the recognition of both aspects of stimulation.

The study of both visual and auditory illusions show that the brain actually smooths out or enhance differences according to which patterns of stimulation it is trained to attend to or ignore. Check the studies of Gestalt Psychology https://en.wikipedia.org/wiki/Gestalt_psychology and not at least the studies of auditory illusions https://en.wikipedia.org/wiki/Auditory_illusion. This is not to say that we constantly hallucinate in our daily living but one consequence is that it is very easy to fool the brain into hearing something that is not given in the discrete parts of stimulation. A (funny) illustration of that is the discussion of “backmasking”, https://en.wikipedia.org/wiki/Backmasking which among other things is something that Rock bands have been accused of to induce subliminal Satanic messages to people. Here is an example of “Stairways to Heaven” played backwards:



First close your eyes and just listen. What you hear is probably some unintelligible mumbling. Now listen to it again and read the text while hearing it and suddenly the Satanic words are obvious to the brain.
You can try this with Justin Bieber as well, that naughty little demon:



Now I could post some classic social psychological experiments on the effects of cultural preferences, group pressure and such to further convince you how easily the brain is affected by the beliefs of its environment, but I am going to spare you because I think you know this already. Such experiments usually confirm our daily knowledge of such things and may only surprise in their demonstrations of how severe this effect can be.

At this point, it would be no crime from a psychological point of view to conceive the analog-digital phenomenon as sensitive to these facts as with any other kind of perceptual stimulation. Thus in the end, no matter what features of stimulation you can point to, there will be no direct links to how it is processed and whether we like it or not….though the cultural context can be indicative for a large amount of people.

Then there also is the question of multisensory integration https://en.wikipedia.org/wiki/Multisensory_integration. In daily living the concept of e.g. an analog synth versus a digital synth cannot be reduced to listening to highly compressed mp3s matching tiny snippets of sound. Actually the look, the feel, yes maybe even the smell of the synth, play a part in your brains evaluation of its “sound”. Add to this your knowledge about which kind of synth it is, since this can direct your attention to specific aspects that you wouldn’t notice without this knowledge, e.g. searching for the variations (analog) or artificialness (digital) of the sound at an unnaturally detailed level.

And now I will say something that you may not expect from all this, namely that even given these circumstances, I still like the sound of analog synths a bit better than that of digital! BUT…the difference between me and many other seems to be that I do not make any illusions about this being caused by some analogue laws of stimulation affecting my value system in a 1:1 relationship. I know it is due to my upbringing with these synths, the knowledge that the synth is analog and not digital, and probably the cultural hype there is about analog synths as well. But in the end, I don’t give a f.u.c.k. about that as long as my brain somehow enables me to love the sound. As far as that goes, ignorance is bliss to me

Cheers

[Tagirijus]

Well, I guess all those researchers at the Audio Engineering Society and people like David Lavry can learn a thing or two from you! And all those people with expensive high end converters that run them at 96/24 or greater. and forget sites like HD Tracks that sell 96/24 and 192/24 digital media - they are also barking up the wrong tree according to you. So much stupidity out there! Vintage King should be going out of business soon. And forget hacks like Jack Puig and all those other studios crammed full of analogue gear and high data rate gear. Wrong wrong wrong. Yep, a copy Tracktion and some Melda plugins running at 44/16 are all that is needed to produce professional music records.

Sorry posting this without a source of the information, but I had a similar discussion in another forum. Some people posted a study which revealed that higher sample rates (I think it was something like 96 kHz), are not audible. I made some tests by myself and I could not hear a difference, but ...

If you don't believe me, spark up something like Reaktor that allows for a sample rate that's different than your project. Make sure the sample rate is the same as your project, which should be 44.1 or 48 kHz. Take an ens. Like Monark and use the 3rd to modulate the filter cutoff in anything but "low." Make sure the mod wheel graphic is up. Now, crank up the sample rate to 96 kHz and listen.

... I will definitely check this one out though!


Back to the other quote:
The thing which (at least at this moment) makes me believe that higher numbers and stuff are just for marketing, is my own experience in music and sound design production. I've always managed to get the sound which made me really happy, while not using 96 kHz for example. Of course: my experience cannot stand for the "real truth" (whatever this might be ...), but it tells me that I (personally) do not need analogue or 96 kHz etc. No question, maybe the day will come and I'll see what exactly makes the difference, who knows. But at the moment I can make everything I have in mind and therefore somehow think that the music industry is like german politics: ruled by lobby, haha. (... . .. . ... actually the fact about german politics is sad )

[MeldaProduction]

Batty: Almost agree with you . But I absolutely don't agree about "analog being the way we perceive sound". I already explained, it's just a different representation of the very inaccurate approximation of sound pressure.

plexxuss: I never said 44k/16 . Personally I do everything in 48k/24. There have been several papers defining that everything above 48k is an overkill and can even do more harm than good. Hard to say, I tried myself and I just cannot hear a difference between 48k and 96k. The digital processing is a different question. But in most cases it doesn't matter. And when it does, you can always use upsampling on the plugins.
For the record, 192k is known to degrade audio quality in many cases, because it's just too hard to get an accurate clock signal for frequencies that high. And even if you can make it work perfectly, I'm not aware of any study that would prove any positive difference. Note that the fact that some high end engineers use it doesn't mean a single thing! These are only just people, and even worse, these are expected to have the most expensive stuff out there. Not mentioning they are used as walking commercial for obvious reasons .
I think you should follow your instinct . If it tells you 192k is good and you really feel it, you can hear the difference (please test it and stay unbiased), then go for it. But note that when testing, you only need to play the same thing in 192k, 96k and 48k, no additional processing except for downsampling, otherwise your test wouldn't prove anything.

Grizzellda: Damn sorry . It's like my name, so hard to spell . How many Vojteks, Vojechs and such are there .
Hehe but CDs suck... I love classic music and with all the dynamics it gets really bad in the silent parts, which I like the most .
As for warming up - why not!!! I never said we shouldn't saturate and stuff. That's what we learned from analog! But it doesn't mean we should spend our lives trying to recreate the ancient stuff (aka vintage ). I just want to take what I like from that .
And now I want a hug from you!

Zerocrossing: Yep, it needs to feel right. And there are things that doesn't work that well in digital indeed! It's the nonlinear operations - extreme distortion and mainly FM. These are all actually pretty simple, in the super conditions with huge sampling rates, like 1000k . But we'll get there .

IncarnateX: Whau, cool stuff!!! Especially the Bieber thing . Who knew the satan is already among us!
Btw. about the repeatitive vs. nonrepeative stuff. I actually believe it is partly caused be sex. Men are used to protect and fight, so for example for me is easy to sleep when a repeatitive sound is playing, but random stuff, no way, I'll spend the whole night checking them out and getting nervous . While women can ignore the random stuff, that's what they have men for . It really seems to work like that.
I very much agree with you in pretty much everything! Yep, our brains are sooooo imperfect , or more like "washable"

Tagirijus: Exactly!

[jmg8]

What are your thoughts on us getting dynamic convolution?

[IncarnateX]

Btw. about the repeatitive vs. nonrepeative stuff. I actually believe it is partly caused be sex. Men are used to protect and fight, so for example for me is easy to sleep when a repeatitive sound is playing, but random stuff, no way, I'll spend the whole night checking them out and getting nervous . While women can ignore the random stuff, that's what they have men for . It really seems to work like that.!

Well you got yourself a fine hypothesis there, which seems logical according to an evolutionary perspective and further is suggested by your own experience. Now the usual problem with such kind of hypotheses is putting them to a test without getting your Willy into the feminist-gender-research machine and being accused of discrimination. Dangerous territory

[mesaone]

processing at higher sampling rates makes sense and more often than not even desired, especially for non-linear processes. recording and listening doesn't.

I disagree. Recording at high sample rates may not make sense for everyone in all situations, but not everyone is doing the same thing or tracking/mixing in the same manner. There's a huge variety of workflows and equipment out there. What may not make sense TO YOU may make a lot of sense to someone else.

When I record through CLASP and a console, I always use 96 kHz. Same deal when I patch in outboard gear with hardware inserts in Pro Tools. Not just for sound, but for function - for example, the fastest attack of an 1176 is faster than a sampling rate of 44.1 or 48 kHz.

[MeldaProduction]

IncarnateX: Yeah , damn what it a just said? My career is over?

mesaone: It's more like a technical question "if it makes sense to use higher sampling rates". From the auditory point of view it shouldn't matter, we are just unable to hear that. When it comes to some of your plugins, that's a different story. It's often problematic to make the plugins work in both high and low sampling rates, but Melda plugins should be usually fine .

[IncarnateX]

IncarnateX: Yeah , damn what it a just said? My career is over?

Nah, it's still a male's world. Just keep quit about it in your manuals and you are safe. You know what they say: Make money, not war...or something. Keep up the good work

[goldglob]

IncarnateX: Yeah , damn what it a just said? My career is over?

You could always go into politics

[mesaone]

mesaone: It's more like a technical question "if it makes sense to use higher sampling rates". From the auditory point of view it shouldn't matter, we are just unable to hear that. When it comes to some of your plugins, that's a different story. It's often problematic to make the plugins work in both high and low sampling rates, but Melda plugins should be usually fine .

Vojtech, I was referring to hardware. You should check out CLASP, it's a very cool system made by a company called Endless Analog.

But yes, plug-ins too - working at a high sampling rate (or oversampling with Melda plugs) works just fine, especially since Melda EQs have HF response cramping. I generally prefer de-cramped, and working at 88.2 or 96 kHz (or using 2x oversampling with other sample rates) takes care of that.