What makes analog so analog?

[Lawnmower Of The Damned]

Alright, I have an idea, and I want you to tell me if you think it would work (Or not).

My belief is that the analog sound comes from the oscillators. If you were to run an analog oscillator through a digital filter, I believe it would still sound analog.

So, let's say we have a specialized audio card. This audio card has a midi port, an analog in/out and a nice analog/digital converter on board. It also comes with a piece of software similar to Synth Edit, or Reaktor.

Here's a step by step process of how it would work.

1) Midi data from your external keyboard goes into the soundcard. This data is taken into the editor for further manipulation.

2) The software gives you a digital envelope. You can use ADSR, or whatever other cooky envelope you might want. You could also set up an arpegiator, step sequencer, or pitch modulation, whatever you like.

3) The sound card converts that midi data to voltage and passes it out, into a standalone analog oscillator box.

4) The audio passes back into your sound card, loving every moment of the high quality analog/digital conversion that keeps all of the analog sound intact. The data will remain audio data from this point on, without any conversion back to midi.

5) The software editor allows you to set up filters, LFO's, all that great stuff, in the digital domain.

So now, we have an analog/digital cyborg version of Reaktor. Fully flexable, probably affordable as a good oscillator costs what, $40? A basic sound card with a midiport... The only expensive part will be the nice A/D - D/A converters.

The question is, will it sound analog? What do you think?

[meister eder]

Dr. Wackler, could you expand upon the instantaneous feedback? Do you mean the kind of feedback that I know and love as a guitarist, or do you mean zero latency, or something else altogether?

In a way I mean both:
The feedback that you know and love as a guitarist, is instantanious feedback. This means, the feedback itself is happening in the very moment that the original signal that causes the feedback is happening. In the digital domain however, you will always be confronted with a copy of the original signal which is delayed at the least by 1 sample (so it's about latency as well in a way, but the important part is the 'copy').

One may think that even the analog signal needs a (however small) fraction of time between the original signal and the fed back signal - but that's not what feedback is. Analog feedback actually does happen the very moment that the original signal that causes the feedback is happening.

IMO this fact is not only of interest when you are actually creating feedback, but it is also an essential aspect of what makes analog so analog. However esoteric that may seem to some (or many).

[nutekk]

it is impossible for an a/d converter to keep
the "analog sound" intact

[tee boy]

Alright, I have an idea, and I want you to tell me if you think it would work (Or not).

My belief is that the analog sound comes from the oscillators. If you were to run an analog oscillator through a digital filter, I believe it would still sound analog.

So, let's say we have a specialized audio card. This audio card has a midi port, an analog in/out and a nice analog/digital converter on board. It also comes with a piece of software similar to Synth Edit, or Reaktor.

Here's a step by step process of how it would work.

1) Midi data from your external keyboard goes into the soundcard. This data is taken into the editor for further manipulation.

2) The software gives you a digital envelope. You can use ADSR, or whatever other cooky envelope you might want. You could also set up an arpegiator, step sequencer, or pitch modulation, whatever you like.

3) The sound card converts that midi data to voltage and passes it out, into a standalone analog oscillator box.

4) The audio passes back into your sound card, loving every moment of the high quality analog/digital conversion that keeps all of the analog sound intact. The data will remain audio data from this point on, without any conversion back to midi.

5) The software editor allows you to set up filters, LFO's, all that great stuff, in the digital domain.

So now, we have an analog/digital cyborg version of Reaktor. Fully flexable, probably affordable as a good oscillator costs what, $40? A basic sound card with a midiport... The only expensive part will be the nice A/D - D/A converters.

The question is, will it sound analog? What do you think?

Im not an expert, but its my opinion that the filters are equally as important as the oscillators. I think if you run a VSTi synth through a guitar overdrive, then an outboard filter you will have a very analog sound. If you run an analog oscillator through digital effects, it will require the digitisation of the continuos signal and therefore present digital artifacts. I guess this is why many top engineers record and edit in PT, then mix on wacking great SSL's.

[neverwhere2012]

it's all to do with the quality of components. is it discrete or not? active or passive? FET? tube? tape? it all depends. a $100 behringer eurorack or $200 mackie board is analog but sounds arguably as bad as or worse than digital most of the time. and i don't buy into "signal degradation" because that's precisely what nyquist does to your digital audio.

the reason you get such great top end response out of nice analog gear, ESPECIALLY the vintage stuff like neve/telefunken/neumann is the fact that all the components used to build it were military spec and rated from 10Hz to 100kHz. it was all that was available. warmth and presence aren't created from signal degradation, just saturation or difference kinds. too much drive boosts different harmonics giving different tonal characters.

and in the end, it's all due to the way the electrical current is being run through the various components and written magnetically to the tape. perhaps the problem with digital lacking warmth is the fact that it's using math to reconstruct something that simply occurs in "nature" (acoustic energy, i'm not insinuating that there are squrrels out there making crunk... though that would be interesting too). all sound is energy and easily converted into different forms of natural energy. the "problem" with digital is that it's trying to "clone" these natural phenomena and it can only do it so well.

look at it this way:

science can clone or "create" a human being. science can create artificial intelligence. but can science create a soul? how exact a copy of the original can a computer make out of 1's and 0's?

however, the technology really isn't too far off. 24/192 is almost undiscernable in pure detail for analog. there's plenty of plugins coming out that actually have a character to them, that have a "color" like many analog pieces do (universal audio plugs being foremost of these in my mind).

but i know what you mean, my friend who's trying to make amp sims in synthedit has asked me many times what makes analog sound analog. the real answer is i haven't taken enough electronics classes to figure it out yet

[Lawnmower Of The Damned]

Dr. Wackler, that does make a lot of sense. Feedback can certainly create some high-end noise that could account for the percieved loss of presence in VSTi's. Unfortunately, I don't think my idea would address that at all.

Nutekk, I don't agree with you on that one. If that were true then every synth on a CD you've ever heard would sound like a VSTi, even if it was analog. CDs are digital, so all that analog sound needs to be converted to digital at one point or another. Not only is it possible, but it's not that hard either. The problem doesn't seem to lie in the virtual recording of the VSTi output, but in the creation of the sound itself.

Tee Boy, I wish I had an analog filter laying about to try it out. If there's anyone out there who could run a VSTi through an analog filter and post the sample, it would tremendously help.

Though it doesn't require Pro Tools rig to be able to capture that analog sound. You can hear the analog sound all over the place. Like I said, the problem doesn't seem to lie in the virtual recoding of the VSTi, but occours during the actual creation of the sound. And the very genesis of the sound takes place in the oscillators, which is why I would look there first when trying to improve the sound.

P.S. Good post, Neverwhere. Once again, something my idea doesn't tackle. I hope you guys are wrong, but your ideas make sense to me.

[Homebelly]

In a soft synth every thing is finite, it is written in code and there is no room for variance,, if variance has been written in as part of the code then it is limited to how much is allowed and under what circumstances it is to be present.
In an analogue synth there is infinite room for variance,, indeed as any one who has used a true analogue synth will testify,, this can lead to real problems in keeping the oscillators in tune,, but this also extends to the way the amplifiers handle the voltage being sent to them by the oscillators and how the filters alter that signal and so on,,
The introduction of distortion is also important in what makes an analogue synth in the real world sound " Phatt ",, this has little to do with the destruction of the sound in the way you might if you used an over drive or distortion pedal,, but it is similar,,
It's my understanding that in an analogue environment if a signal goes over it's head room it will compress slightly until it is unable to do so any more and then it will fold back on its self,,
In the digital environment it will simply cease to exist and will square out producing a horrible noise that hopefully, none of us hear to often,,
This is also a part of the reason that some people swear there hardware Virus sounds way better than there software virus,, it's probably because of the noise and distortion that is introduced in the D/A conversion and also in the cables and wires getting the signal into there DAW,,
Some software synths are programed with the ability to mimic this phenomenon to varying degrees of success, MOOG MODULAR V is one that springs to mind that dose this quite well,, but this often comes at a cost to your CPU,,,

[tee boy]

neverwhere2012,
Dude, you are probably spot on with this post. Iv heard so many arguements over which opamp is best in which micpre... its crazy! The saturation and general 'character' of classic analog designs is likely to play a major part in the whole 'sound' thing. I mean, run some beats through a 1073 with the gain cracked, and you know your in analog heaven. Try the same through your Behringer, and... errr need i say more!

thornemaelstrom
I dont have anything on me right now, but i can tell you quite categorically that this is the case. All the time i used this technique. Say Im doing some ambient stuff, and Iv create some textures using 100% digital processing. This can often sounds a little bland and 2D, so at this point i get the pedals out and warm it up with a little Big Muff overdrive or whatever. If you can get hold of a few nice overdrives, and maybe a couple of old analog filters, then you should be able to warm up your digital synths a treat. Infact, I wonder what Arturia Minimoog would sound like run through a Moogerfooger (I imagine rather nice ).

Funnily enough, this thread reminds me of a conversation i had with a mastering engineer a while back. He was saying to me that he much prefers to receive mixes which are a little dull sounding, cuz he feels that brightening stuff up in the analog realm always sounds great. Mixes which sound too bright in digital can tend to sound harsh in comparison. I guess this ties in quite nicely with what the original poster was saying about analog presence and air.

[Lawnmower Of The Damned]

Neverwhere, were you talking about Universal Audio as in these guys: http://www.uaudio.com/ ?

I wasn't and to find any instruments, just effects. I'm not really talking about effects, I'm really just talking about things that would require oscillators for their sounds.

Homebelly, Moog Modular V is one of the more analog-ish VSTi's that I'm aware of, but it still isn't quite there.

My solution would still be passing data through wires, and would still be using an analog oscillator, which should not fall prey to digital clipping, so long as the A/D-D/A converters are nice.

Possibly a built in pre-amp or compressor would be a good addition to the card?

[McLilith]

But still instantaneous feedback will never be possible in the digital domain.

It's also not possible in the analog world, if you want to be picky about things.


later,
McLilith

[Robert Randolph]

Most ME's prefer to receive dull mixes...

[meister eder]

But still instantaneous feedback will never be possible in the digital domain.

It's also not possible in the analog world, if you want to be picky about things.

later,
McLilith

One may think that even the analog signal needs a (however small) fraction of time between the original signal and the fed back signal - but that's not what feedback is. Analog feedback actually does happen the very moment that the original signal that causes the feedback is happening.

[Lunch Money]

Without rehashing all the old arguments, which we've all heard a million times, I'll take a different tack:

When I said "signal degradation" I didn't really mean it as an insult. I just meant that by the time a signal has passed through transistors, resistors, potentiometers, capacitors... there will be signal degradation. It's not a smart-assed comment, but a fact. I guess there's a negative stigma attached to "degradation" (for good linguistic reason!), but I only meant that when we're talking about synthesizers (which IS what we're talking about, not pristine mixing consoles) there will be certain changes in the signal.

As a guitarist, I LOVE that audible sound of "air" coming out of my amp. There's nothing quite like it. But it's because of (relatively) noisy componentry that we can hear it at all.

On a note of interest related to an earlier point, has anybody ever seen one of THESE? There was an idea doomed.... actual audio enthusiasts know well enough to put their tubes elsewhere. <chuckle>

Greg

[McLilith]

sampling rate (khz you speak of) has nothing to do with quantization, that is bitrate. Harmonic content is lost at lower sampling rates, however, quantization is an amplitude phenomenon.. its when varying levels of amplitude are "quantized", or restricted to a certain area of existance, it creates noise that is unremovable. It is most noticable at 16 and lower bitrates.

Firstly, you are confusing bitrate (a measurement for serial transmission of digital data) with bit depth. When you refer to "16 and lower", you are actually referring to bit depth, not bitrate. Secondly, quantization isn't limited to amplitude alone. You can also speak of temporal quantization, which would be directly affected by the sampling rate.

it is said there is alot more going on in realms higher than 20 khz (the top frequency of human hearing)

Now, you've gotten into the realm of a personal pet peeve of mine. 20 kHz is not some magic threshold that represents the highest frequency that humans could ever hope to hear--although that is widely and blindly accepted as the truth. That figure is a rough average for the upper frequency range of healthy human hearing, but it is by no means a "brick-wall" limit. The actual upper limit for an individual will vary on either side of 20 kHz mark.

Yes, some people can hear sounds above 20 kHz. The hearing response also does not fall off instantly. Do you really think that a person would hear 19.999 kHz and be absolutely totally deaf to 20.001 kHz? Of course not. The upper limit of human hearing is a somewhat fuzzy variable.

If you want to get a real shock, look in the Guiness Book of World records for some eye opening information about the highest pitches that have been heard.

I've talked to former sailors that heard the "ultrasonic" sonar used on military vessels, and even got terrible headaches from it. I've personally been able to determine when a 45 kHz ultrasonic burglar alarm was transmitting its signal and when it shut off. I couldn't detect any definite sensation of pitch at that frequency, and it was a subtle sensation, but it was a definite sensation just the same. I would describe it as feeling more like a sensation of "pressure" than "normal sound". One day, while suffering from a severe headache, I was able to detect a 200 kHz sine wave. Once again, I couldn't sense the pitch, only the presence of the signal, and it felt like a dull sensation of "pressure." I only know that it was 200 kHz, because it was being generated by a friend with a signal generator that maxed out at 200 kHz, and he was playing it at the highest frequency setting.

The main point that I would like to make is that there are lots of subtle nuances that humans (at least some of us) receive from frequencies outside the traditional "textbook" range of human hearing. This is an area that isn't isn't well understood yet, but there are definitely things above 20 kHz that affect people in various ways.


take care,
McLilith

[tony tony chopper]

what makes analog sound so... well, analog?

snobs