The No.1 Website for Pro Audio
 All  This Thread  Reviews  Gear Database  Gear for sale     Latest  Trending
Won't computers produce a convincing analog sound someday?
Old 6th March 2013
  #1
Lives for gear
 

Won't computers produce a convincing analog sound someday?

Just wondering about computing power and analog- I realize that analog is actual electricity with all the fineness this implies. But humans have a limit in how well they perceive sound and the real question is: at what point will computers produce a sound that is identical to analog as far as humans can tell? How close are we? (85% of the way there? 95%?, two years away at present rates?)
Old 6th March 2013
  #2
Ged
Lives for gear
 
Ged's Avatar
Good question - but.....by "Sounding Analog" what do you specifically mean???
Old 6th March 2013
  #3
Lives for gear
 
kirkelein's Avatar
 

I'd say 15 years. And 30 if they're going to have the same latency as well.
But it will be highly usable and much more practical long before then; already some people don't mind the sonic difference.
Old 6th March 2013
  #4
viz
Here for the gear
 

Which humans?

There are many people who can't hear a difference right now.

What "limit"?

If it's frequency range, 24/192 recording can theoretically capture/reproduce everything in the audible spectrum ("what humans can hear") with enough room beyond that to avoid audible artifacts. You'll still see people arguing that higher sample rates are needed to capture harmonics outside the audible spectrum. They may be right. But my first point still applies.

If you're talking about recreating the dynamics present in, say, an analog oscillator - the tiny fluctuations in pitch, volume, frequency content, etc., that human ears are very sensitive to - that comes down to how you generate and apply (pseudo)randomness in a computer. Which at this point is mainly an issue of designer knowledge/skill than raw computing power.
Old 6th March 2013
  #5
Lives for gear
 
aTelecine-Lex's Avatar
Quote:
Originally Posted by kirkelein View Post
I'd say 15 years. And 30 if they're going to have the same latency as well.
But it will be highly usable and much more practical long before then; already some people don't mind the sonic difference.
Just curious, (no ill-will intended)

What factors informed your guess(es) of 15, and 30 years?

Best,
Alexa
Old 6th March 2013
  #6
Lives for gear
 
lysander's Avatar
 

Please...not...this.....thread.......again !!!
Old 6th March 2013
  #7
Lives for gear
 
frans's Avatar
A number of things. First, the things you can buy in a shop don't equal "the best" you can buy, they just want you to believe that. So there are already machines out there with enough computing power. You just can't buy them.

Second, another stepping stone is to write something in code, you have to fully understand first what's happening inside the circuits and components. As 'science' still struggles to understand what's matter, energy, time or other things this could take some time. Well, some parts of science have an okay grasp on these things but this news didn't so far make the headlines in the rest of the scientific community, not even speaking from behind the fence in other parts of science, which is split in tiny parts so the facts uncovered don't spread too fast.

Third, it's all about code, not computing power. The shops want to sell you newer machines so they repeat, repeat, repeat that you need more, more, more CPU, it's just marketing. It's not true or the end-all. Marketing = lies.
Specialized CPU is much faster than jack-of-all-trades Intel junk. Look up which CPUs the NASA uses for space probes. Hint, hint: not a single Intel. You want (available for civilians) computing power you can buy now? Look for something with FPGA.

The lowdown: it's possible, now. But why would you do that? As long there is hardware, who needs obsolence-built-in software? Go try to run a softsynth in a few years from now. But you still can play a Moog from the 60s.
Old 6th March 2013
  #8
Lives for gear
 

Yes, eventually (source: undergrad in physics, masters in math and comp sci)...it's pointless speculating a time-line as this depends on the development of cheaper and better computational power, the willingness of programmers to actually do it, etc etc blah blah blah
Old 6th March 2013
  #9
Lives for gear
 
mainesthai's Avatar
 

Quote:
Originally Posted by Lune View Post
Just wondering about computing power and analog- I realize that analog is actual electricity with all the fineness this implies. But humans have a limit in how well they perceive sound and the real question is: at what point will computers produce a sound that is identical to analog as far as humans can tell? How close are we? (85% of the way there? 95%?, two years away at present rates?)
It depends on the sound. A lot!

Clean bass sounds with no resonance and no modulation can be very convincing right now.
But cross-modulating oscillators and then ring-modulating them will take some years. Its difficult to say how long because of the exponential rate at witch technology grows and the implementation of that technology on synthesizers. My guess: 10 years???
Old 6th March 2013
  #10
Gear Addict
 

Happened a few years ago as far as I'm concerned.
Old 6th March 2013
  #11
Lives for gear
 
login's Avatar
Even when we get to have 3d printers and you can print an analog synth (any replica of classic ones) people will complain about it lacking mojo, karma or soul.
Old 6th March 2013
  #12
Lives for gear
 

Well it's like, at a certain point, with the effort and expense of the hardware and software, why not just go analog?
Old 6th March 2013
  #13
Ged
Lives for gear
 
Ged's Avatar
Quote:
Originally Posted by login View Post
Even when we get to have 3d printers and you can print an analog synth (any replica of classic ones) people will complain about it lacking mojo, karma or soul.
LOL, True!!
Old 6th March 2013
  #14
Gear Nut
 

Given a choice between two synths I will ALWAYS pick the one with better karma, definitely.
Old 6th March 2013
  #15
Lives for gear
 

Quote:
Originally Posted by frans View Post
Third, it's all about code, not computing power.
Tell that to owners of DIVA! I have the latest, fully stocked iMac and DIVA still hiccups on me (and it doesn't even fully model every part of the analog signal path).
Old 6th March 2013
  #16
227861
Guest
Call me weird, I find a bond with a piece of hardware.
Old 6th March 2013
  #17
Lives for gear
 

Quote:
Originally Posted by draven5 View Post
Call me weird, I find a bond with a piece of hardware.
Yes there is a completely different user experience when you are able to physically manipulate multiple things at once...BUT, eventually we will have AFFORDABLE 3D printing capable of printing that physical interface, and we will also have touchscreens/surfaces with dynamic tactile input (pop-out buttons, knobs, etc).
Old 6th March 2013
  #18
Lives for gear
 
gremlin moon's Avatar
 

Why doesn't someone post up four or five examples of pure non-digitally processed analog and I, and others, will see if it can be recreated using a softsynth. Post your patch parameters if you want to get technical. However, no saying things like, "Soundcloud altered the audible analogocity of this," or "Youtube's algo-compression ruined the sound," or "my converters are weak. . ." Ideally, do two or three bread and butter and then a some zero-through cross mod inversion run through a analog filter bank. Let's see what type clothes the Emperor (or Monarch) is wearing. Then we will have a reverse challenge to see what type of digital that analog can do.
Old 6th March 2013
  #19
Lives for gear
 

It's not just about recreating an analog sound. The interface and limitations of some analog synths can help creativity and produce exciting results. A lot of the shootouts on here are about creating some very specific preset sound, which can miss the point. When you are working with analog synths many times you tend to change the sound not the presets and it's easy to stumble on things that sound good that you would be difficult to reproduce.

And when you are recording to audio rather midi you tend to resample yourself more rather than adjusting midi notes, moving things on a midi grid, and then getting frustrated and changing a preset again. The process is different not just the sound.
Old 6th March 2013
  #20
Lives for gear
 
gremlin moon's Avatar
 

The OP was asking about a "convincing" analog sound -- not workflow. That is also why I suggested a wide representative variety of sounds.

[edit: I am saving for some form of modular rig right now -- it would be nice to a/b what i am really saving for in terms of sound characteristics in relation to what i can presently do with my current digital setup (SW/HW).]
Old 6th March 2013
  #21
Lives for gear
 

The process leads to the sound.

Softsynths are convincing for a lot of things, but some bass sounds and the way filters resonate and some VCAs overdrive are still tough to nail. It's not something anyone should lose any sleep over though.
Old 6th March 2013
  #22
Gear Guru
 
Yoozer's Avatar
Why spend a million on R&D and writing code if $30 in parts gets it right the first time?
Old 6th March 2013
  #23
Lives for gear
 

Yeah I'm not buying any test unless the analog stays in the analog domain all the way from source through speakers.
Old 6th March 2013
  #24
Lives for gear
 
kirkelein's Avatar
 

Quote:
Originally Posted by aTelecine-Lex View Post
Just curious, (no ill-will intended)

What factors informed your guess(es) of 15, and 30 years?

Best,
Alexa
Nothing!

It was merely a guesstimate based on the soft sounds I have heard today, compared to what was available a few years back. And those figures are for perfect (as far as humans are concerned) emulation. For more practical goals I think commercially available tech will take 5 years at most; there are already some things that can be done almost flawlessly.

The latency thing was just something I threw out there. Given the immense capability of certain hardware maybe it is just a question of cost; as soon as the appropriate algorithm is found, you pair it with a processor fast enough for your needs. Nanosecond latency seems hard though. But I've never measured my analogues so I don't know if they're really all that fast.

Quote:
Originally Posted by draven5 View Post
Call me weird, I find a bond with a piece of hardware.
+1

Edit: I also think there is something intuitively bothersome with emulation. Even if it is perfect, I still want the real thing. Like some foods not being as appealing when you find out how they were made. Go reality!
Old 6th March 2013
  #25
Lives for gear
 

If anyone wants to watch a conversation go absolutely no where keep refreshing this pages.

On another note, how long before analogs convincingly reproduce digital sounds?
Old 6th March 2013
  #26
Lives for gear
 
kirkelein's Avatar
 

I actually thought the initial question was interesting: How long before even the savviest analogue freaks cannot hear even the slightest difference, on any settings. I'm interested in hearing what people think, but so far there hasn't been much input on that point. We've heard 10. I've said 15. Anyone else?
Old 6th March 2013
  #27
Lives for gear
 
Sotsirc's Avatar
 

Don't know what year it will happen, but I guess it will be in the spring.
Old 6th March 2013
  #28
viz
Here for the gear
 

Quote:
Originally Posted by mekohler View Post
Tell that to owners of DIVA! I have the latest, fully stocked iMac and DIVA still hiccups on me (and it doesn't even fully model every part of the analog signal path).
I think what Frans is getting at is that with a fixed level of raw computing power, different approaches to coding will produce different results.

Assume DIVA is written in C++. It'd run faster in assembly. It'd run even faster on the raw processor (without OSX). Both of those steps are harder for developers, though.

That's not even touching on the issue of serial vs. parallel architectures that he alluded to with FPGAs.
Old 6th March 2013
  #29
Lives for gear
 
gremlin moon's Avatar
 

Quote:
Originally Posted by kirkelein View Post
I actually thought the initial question was interesting: How long before even the savviest analogue freaks cannot hear even the slightest difference, on any settings. I'm interested in hearing what people think, but so far there hasn't been much input on that point. We've heard 10. I've said 15. Anyone else?
I think the best way to tell is by a/b comparison with what is currently available in software tools and referencing that between developments in other related fields like video games, imaging software, movie sounds, etc... Then we have the analogue freaks do a blind test. There are some scientists on this board who can probably help it run like a legitimate experiment. From combining those factors, in conjunction with trying it with old and new softsynths I think we could come up with a reliable estimate.
Old 6th March 2013
  #30
Lives for gear
 
kirkelein's Avatar
 

Quote:
Originally Posted by gremlin moon View Post
I think the best way to tell is by a/b comparison with what is currently available in software tools and referencing that between developments in other related fields like video games, imaging software, movie sounds, etc... Then we have the analogue freaks do a blind test. There are some scientists on this board who can probably help it run like a legitimate experiment. From combining those factors, in conjunction with trying it with old and new softsynths I think we could come up with a reliable estimate.
You'd be hard pressed to find much support for an experiment that tries to prove something many (on this forum) are convinced is not true, and in some cases also take pride in not being true.

I'm convinced myself that we're not there yet, but that some things can be done convincingly today. Perhaps as someone has pointed out, it isn't really tech that's in the way but code. Still, research needs to be done.
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump