When Did Synthesizers Replace Live Orchestras?

Discussions about anything analog, digital, MIDI, synth technology, techniques, theories and more.
User avatar
calaverasgrande
Expert Member
Expert Member
Posts: 1038
Joined: Thu Jul 07, 2011 5:29 pm
Gear: MG1, MP201, MF101, MF102, Taurus 3, SH09, KPR-77, Streichfett, Dark Energy, X0Xb0x, Dronelab, Synsonics Drums, Machinedrum, Modular.
Band: N.S.V.
Location: New York
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by calaverasgrande » Tue Nov 11, 2014 5:39 am

it is kind of silly to make the contention that a live orchestra is different than a sampled one. That is of course evident. I have classical musicians in my family. Also used to work for the SF symphony for a while. I have experienced symphonic music up close many times. The force of it is certainly quite visceral. But it is not something that can be reproduced via your typical cinematic 5.1 or 7.1 loudspeaker arrangement. No amount of surround sound or sub woofer steering can replicate massed strings with percussion. Even that is only a facsimile of the actual event. yes even John Williams.
The thing is, even full orchestra recorded for film is multi miced. They may use a Decca tree or spaced pair arrangement to get a cohesive sound, but the sections and often individual chairs are miked up and brought in on the mixer per fader. In true control freak film budget fashion of course.
Point being, what does it matter if those individual channels are multi-sampled representations of the instrument or actual musicians?
From an objective point of view I honestly do not think a well composed piece, with care taken for expression and nuance can be distinguished from one recorded en masse.
Subjectively I'm opposed to this kind of automation of what is artisanal craftmanship.
As the digital age has unfolded in cinema, it appears to me that we have lost a lot of the craft of film. You watch even ordiinary films from 30 years ago or more and there is an obvious greater care and attention paid to tasks such as lighting, cinematography, blocking, set dressing and so on. Especially the handcrafted end of filmmaking is being left behind. There is no longer a perceived need for many of the more esoteric special effects like slit scanning, miniatures or matte painting. "We can do it all in post".
So while I think it is easily possible to make a score on hardware that can be taken for one performed by hand, knowing the schedules and budgets that are allotted for the thematic music, I do not think it will feasible for any ordinary person to achieve this on a regular basis. Simply due to the immense amount of nuance that would be need to be detailed in order to render a convincing score.
Whatsoever thy hand findeth to do, do it with all thy might; for there is no work, nor device, nor knowledge, nor wisdom, in the grave

User avatar
joeboy
Junior Member
Junior Member
Posts: 102
Joined: Tue Jul 17, 2012 1:22 am
Gear: ax-80, jx-8p, v-synth, monotribe, ПИФ, tx81z, sy35, tg77
Location: 熊本

Re: When Did Synthesizers Replace Live Orchestras?

Post by joeboy » Tue Nov 11, 2014 9:25 am

ItalianStallion81 wrote:While I know this trend is a black eye for the remainder of the live orchestral string musicians, but in the far future, what if the same thing happened to live rock bands? The people that are the guitarists, bassists and drummers?

Imagine attending a Rolling Stones concert with only Sir Mick Jagger putting out simulated guitar and bass sounds using only one (or two) keyboard(s) and replacing Charlie Watts's role - drums - with a drum machine?? We can do our part and slow down this trend!
The other day I saw a Sparks video where it's just Russell and Ron with everything coming out of a single D-50. It was baller.

As the ranks of dadrock groups start thinning out it'd be cool if they streamlined themselves in a similar way rather than hiring replacements.

Ashe37
Moderator
Moderator
Posts: 3903
Joined: Mon Jan 19, 2009 3:43 pm
Real name: Unpronounceable
Gear: Ensoniq SD-1/32,SQR,VFX,ESQm
Virus Indigo, M3-61 , MS2000BR, Volca Bass
Emu XL-7, Matrix 6r
TG-33, K3m, Blofeld, Micron, Mopho, BS II, JV-1080
Band: Eridani V
Location: Central VA

Re: When Did Synthesizers Replace Live Orchestras?

Post by Ashe37 » Tue Nov 11, 2014 10:18 am

commodorejohn wrote:Yes, and every issue with every one of those problems will resolve itself by the sheer magical force of Progress and next year we'll all have hoverboards. I've heard that song before.
But... we will have hoverboards next year!


User avatar
CS_TBL
Synth Explorer
Synth Explorer
Posts: 1677
Joined: Wed Jul 12, 2006 3:47 pm
Gear: All "In-The-Box"
Mainly FM8
Location: NL
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by CS_TBL » Tue Nov 11, 2014 7:28 pm

commodorejohn wrote:Yes, and every issue with every one of those problems will resolve itself by the sheer magical force of Progress and next year we'll all have hoverboards. I've heard that song before.
These days it's easier to predict the future than it was ages ago. Remember the early samplers like the Synclavier? Instrument players were all like 'OMG our jobs!'. Now, if you see at how big libraries are these days, and if you keep into account the maximum size of such a Synclavier (was it 768 MB?), then the fear of the Synclavier was a bit unrealistic. But it was the time when people were expecting to have flying cars and androids (who've seen c-beams near the Tannhauser gate) in 2019. So far the only android that plays any role in my life is the one in my phone.

Nowadays it's much easier to predict the future because we know of a constant factor that makes predicting the right figures so much easier: computers. To be exact: CPU, system RAM and storage space. In 1997 I bought a PC with 32 MB RAM, now I've a PC with 32 GB RAM, a factor 1000 in 17 years. While 32 TB in 2031 seems insane to think of, so was 32 GB in 1997 - just 3 years earlier I had 4 MB RAM! Same with harddisks; In 1997 I had 3.2 GB iirc. I've 14 TB worth o' external HD's laying around on my desk, and my PC has 1.25 TB worth o' SSD drives. Again: see how this will have grown when we're in 2031. In the 80's, no one knew jack s**t about what computers could and couldn't do (and so predictions didn't make sense at all); that bit of naivety is now behind us, I think.

Now, what to do with all that space! Let's for sake of argument assume that you'll have oodles o' RAM and storage space in (let's make it a bit closer) 2020. Let's assume that every single instrument and vocalist of an orchestra+choir has been recorded (solo). You'd have like 150 instruments. Now let's assume that each instrument will have like, oh, 16 velocity layers and 16 articulation styles, meaning 256 times the same note, played in some kinda way. Even if each single set o' chromatic notes would be 4 GB in size (which would be insane), we're still talking about only 1 TB in size. Using such an orchestra/choir would be doable with a decent CPU and enough RAM, especially by then. Let's double that for sake of argument (the choir may need more vowels and vowel-transitions). No, let's make that 10x, because we want round robins too: 10 TB. In 2020 this figure wouldn't be much out of the ordinary. I've more than that on my desk already, just to store p**n useful stuff.

If, by then, we have the sounds of individual players, with enough round robins and articulations, then I'd challenge you to do a blind test of an authentic orchestral recording and a virtual orchestration! I think this prediction has more fundamental ground than the Synclavier ever had.
"You know I love you, CS, but this is bullshit." (Automatic Gainsay)
s: VSL/FM8/EWQL/LASS h: DX7/FS1r/VL70/SY77/SN2r/JD800/JD990/XV88/Emu6400/Poly61/Amek35:12:2/genelec1030 r: Violin/AltoSax/TinWhistle c: i7-4770/RAM32GB/SSD
FM8 vids

User avatar
calaverasgrande
Expert Member
Expert Member
Posts: 1038
Joined: Thu Jul 07, 2011 5:29 pm
Gear: MG1, MP201, MF101, MF102, Taurus 3, SH09, KPR-77, Streichfett, Dark Energy, X0Xb0x, Dronelab, Synsonics Drums, Machinedrum, Modular.
Band: N.S.V.
Location: New York
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by calaverasgrande » Tue Nov 11, 2014 10:54 pm

I dont know why there is so much emphasis on 'someday' this will be possible.
audio, even huge amounts of multitrack audio, is very low bandwidth compared to video and 3d graphics. It is not a problem to have enough data on hand for replicating each instrument or voice in a symphony. Heck I've got everything I've ever recorded on a 4TB raid array. And it isn't even half full!
With computing being pretty mature at this point in terms of moving huge amounts of data, playing back those samples shouldnt be much of an issue either, even on a plain old middle tier HP or Dell.
I still argue the primary roadblock at this point is in rendering the nuances of fingering, phrasing and delivery that a well oiled orchestra can.
I actually know a guy who is working on something along these lines. I don't want to blow his patent, but it involves some fancy math. He's played a few demos just using GM tone modules. And it was pretty impressive using such mediocre tone generation. I kind of wince imagining how well it could work when it is integrated into sampletank or something similar.

Put another way, Sin City was 9 years ago, Planet Terror 7 years ago. And those films were done pretty much entirely greenscreen with sets and quite a lot of the SFX done digitally.
Whatsoever thy hand findeth to do, do it with all thy might; for there is no work, nor device, nor knowledge, nor wisdom, in the grave

commodorejohn
Synth Explorer
Synth Explorer
Posts: 1586
Joined: Mon Aug 12, 2013 2:39 am
Real name: John
Gear: Roland JX-10/SH-09/MT-32/D-50, Yamaha DX7-II/V50/TX7/TG33/FB-01, Korg MS-20 Mini/ARP Odyssey/DW-8000/X5DR, Ensoniq SQ-80, Oberheim SEM
Location: Sacramento, CA
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by commodorejohn » Tue Nov 11, 2014 11:06 pm

CS_TBL wrote:Nowadays it's much easier to predict the future because we know of a constant factor that makes predicting the right figures so much easier: computers.
If you seriously think that Moore's Law is an honest-to-God constant and we're never going to hit a wall with it, you need to brush up on basic physics, particularly the part where you can't keep squeezing larger and larger amounts of stuff (i.e. circuitry) into a finite space forever and ever without creating a black hole.
If, by then, we have the sounds of individual players, with enough round robins and articulations, then I'd challenge you to do a blind test of an authentic orchestral recording and a virtual orchestration! I think this prediction has more fundamental ground than the Synclavier ever had.
Again, maybe you can convince on a recording, but good luck getting the same results in a live environment from a handful of speakers as opposed to 60+ individual sound sources with their own resonating bodies.
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73

Ashe37
Moderator
Moderator
Posts: 3903
Joined: Mon Jan 19, 2009 3:43 pm
Real name: Unpronounceable
Gear: Ensoniq SD-1/32,SQR,VFX,ESQm
Virus Indigo, M3-61 , MS2000BR, Volca Bass
Emu XL-7, Matrix 6r
TG-33, K3m, Blofeld, Micron, Mopho, BS II, JV-1080
Band: Eridani V
Location: Central VA

Re: When Did Synthesizers Replace Live Orchestras?

Post by Ashe37 » Tue Nov 11, 2014 11:17 pm

commodorejohn wrote:
CS_TBL wrote:Nowadays it's much easier to predict the future because we know of a constant factor that makes predicting the right figures so much easier: computers.
If you seriously think that Moore's Law is an honest-to-God constant and we're never going to hit a wall with it, you need to brush up on basic physics, particularly the part where you can't keep squeezing larger and larger amounts of stuff (i.e. circuitry) into a finite space forever and ever without creating a black hole.
People keep predicting the end of Moore's Law- which says that the number of transistors in an integrated circuit doubles approximately every two years. It doesn't say the chips, and the processes to make them, will keep getting smaller (even though they are, with some exceptions). It just says the number of transistors. They *can* start making the die sizes physically larger again, any time they want.

My proposed corollary to Moore's Law is that the number of people predicting the end of Moore's Law doubles approximately every two years.

User avatar
calaverasgrande
Expert Member
Expert Member
Posts: 1038
Joined: Thu Jul 07, 2011 5:29 pm
Gear: MG1, MP201, MF101, MF102, Taurus 3, SH09, KPR-77, Streichfett, Dark Energy, X0Xb0x, Dronelab, Synsonics Drums, Machinedrum, Modular.
Band: N.S.V.
Location: New York
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by calaverasgrande » Tue Nov 11, 2014 11:41 pm

they are of course working on possible successors to the current transistor/photolithograpy paradigm.
Quantum computing has been making some progress lately.
There is also talk of biological computing using DNA or RNA as computing power.
Another area of research has been into growing circuits using some forms of genetic engineering to achieve scales smaller than lithography can. And then there is research, related to quantum computing, into using particles other than electrons.
Perhaps by the time we really start hitting a wall on silicon chips, we will be ramping up positron computing or some form or 'wet' computing.
Whatsoever thy hand findeth to do, do it with all thy might; for there is no work, nor device, nor knowledge, nor wisdom, in the grave

User avatar
CS_TBL
Synth Explorer
Synth Explorer
Posts: 1677
Joined: Wed Jul 12, 2006 3:47 pm
Gear: All "In-The-Box"
Mainly FM8
Location: NL
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by CS_TBL » Wed Nov 12, 2014 12:42 am

calaverasgrande wrote:I still argue the primary roadblock at this point is in rendering the nuances of fingering, phrasing and delivery that a well oiled orchestra can.
If those musicians can describe in words why they do what they do (all the fingering, phrasing 'n such details) then it can be programmed. They aren't doing random movements you know, they have reasons to do the things they do. I think the next step in virtual orchestras isn't the big lot o' samples alone, I think some smart a*s will create a virtual player which triggers the right samples. This virtual player is actually a virtual human who really plays in the only way an instrument can be played. In other words: if your MIDI data couldn't be played in real life, such a virtual player couldn't play it either.

There's one catch though. A real player can anticipate to what's coming, and this player could choose to e.g. increase its loudness prior to a new note. Current software doesn't, it reacts to whatever is happening 'now', it doesn't look ahead. Years ago there was Synful, remember that? With the 1 sec delay-mode it could look-ahead. Naturally it worked awfully in combination with other stuff. So, the big breakthrough in DAW-land would have to be software that looks ahead and sends such info to the plug-ins. This would rule out real-time playing (as it doesn't know what you're going to do) but it would work perfectly during playback.

But we'll get there, sooner or later..
"You know I love you, CS, but this is bullshit." (Automatic Gainsay)
s: VSL/FM8/EWQL/LASS h: DX7/FS1r/VL70/SY77/SN2r/JD800/JD990/XV88/Emu6400/Poly61/Amek35:12:2/genelec1030 r: Violin/AltoSax/TinWhistle c: i7-4770/RAM32GB/SSD
FM8 vids

User avatar
CS_TBL
Synth Explorer
Synth Explorer
Posts: 1677
Joined: Wed Jul 12, 2006 3:47 pm
Gear: All "In-The-Box"
Mainly FM8
Location: NL
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by CS_TBL » Wed Nov 12, 2014 12:51 am

commodorejohn wrote:Again, maybe you can convince on a recording, but good luck getting the same results in a live environment from a handful of speakers as opposed to 60+ individual sound sources with their own resonating bodies.
Of course, I've once conducted 33 strings, 3 ww and piano in a medium concert hall for a small student film in 2001. I was at the conductor's spot (the conductor left and I had to improvise some undocumented bits with which I created the opening shots of the short :D). I heard and could locate each individual string player, so yea: in a stereo recording (or 5.1/7.1 for that matter) that's missing. I know.

Would I care about that for a film/game OST? Not one bit. If a virtual orchestration would at some point sound the way a regular LSO recording sounds on CD, I'll be the happiest CS_TBL in the world!

Do note that I'm merely talking about production work, for which there's usually time nor budget to work with real musicians. I'd still love to visit the concert hall now and then, I plan to visit a local orchestra's open rehearsal in January.
"You know I love you, CS, but this is bullshit." (Automatic Gainsay)
s: VSL/FM8/EWQL/LASS h: DX7/FS1r/VL70/SY77/SN2r/JD800/JD990/XV88/Emu6400/Poly61/Amek35:12:2/genelec1030 r: Violin/AltoSax/TinWhistle c: i7-4770/RAM32GB/SSD
FM8 vids

User avatar
CS_TBL
Synth Explorer
Synth Explorer
Posts: 1677
Joined: Wed Jul 12, 2006 3:47 pm
Gear: All "In-The-Box"
Mainly FM8
Location: NL
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by CS_TBL » Wed Nov 12, 2014 12:59 am

commodorejohn wrote:If you seriously think that Moore's Law is an honest-to-God constant and we're never going to hit a wall with it, you need to brush up on basic physics, particularly the part where you can't keep squeezing larger and larger amounts of stuff (i.e. circuitry) into a finite space forever and ever without creating a black hole.
It's true that individual cores won't grow that much compared to the increasing figures we saw in the 90's and early 00's. The MHz figures seem stuck, instead the amount of cores grows - I've 8 in my i7. My PC case is quite empty as nearly all of the stuff I need is on the mainboard itself. I'd say that leaves plenty o' room for huuuuuge blocks o' RAM, huuuge CPUs 'n such. So, growth will likely come from quantities o' regular units, not from smaller components. Iirc I've now 4x 8 GB RAM chips. If the maximum size for RAM chips grinds to a halt, I'm sure we'll see mainboards with 8, 12 or 16 RAM slots.

Point is: how is not important, I'm just sure that the PC I buy in 2018 is a lot faster than my current i7.
"You know I love you, CS, but this is bullshit." (Automatic Gainsay)
s: VSL/FM8/EWQL/LASS h: DX7/FS1r/VL70/SY77/SN2r/JD800/JD990/XV88/Emu6400/Poly61/Amek35:12:2/genelec1030 r: Violin/AltoSax/TinWhistle c: i7-4770/RAM32GB/SSD
FM8 vids

Ashe37
Moderator
Moderator
Posts: 3903
Joined: Mon Jan 19, 2009 3:43 pm
Real name: Unpronounceable
Gear: Ensoniq SD-1/32,SQR,VFX,ESQm
Virus Indigo, M3-61 , MS2000BR, Volca Bass
Emu XL-7, Matrix 6r
TG-33, K3m, Blofeld, Micron, Mopho, BS II, JV-1080
Band: Eridani V
Location: Central VA

Re: When Did Synthesizers Replace Live Orchestras?

Post by Ashe37 » Wed Nov 12, 2014 1:08 am

CS_TBL wrote:
commodorejohn wrote:If you seriously think that Moore's Law is an honest-to-God constant and we're never going to hit a wall with it, you need to brush up on basic physics, particularly the part where you can't keep squeezing larger and larger amounts of stuff (i.e. circuitry) into a finite space forever and ever without creating a black hole.
It's true that individual cores won't grow that much compared to the increasing figures we saw in the 90's and early 00's. The MHz figures seem stuck, instead the amount of cores grows - I've 8 in my i7. My PC case is quite empty as nearly all of the stuff I need is on the mainboard itself. I'd say that leaves plenty o' room for huuuuuge blocks o' RAM, huuuge CPUs 'n such. So, growth will likely come from quantities o' regular units, not from smaller components. Iirc I've now 4x 8 GB RAM chips. If the maximum size for RAM chips grinds to a halt, I'm sure we'll see mainboards with 8, 12 or 16 RAM slots.

Point is: how is not important, I'm just sure that the PC I buy in 2018 is a lot faster than my current i7.
Oh, you can buy a PC *now* that is faster than your i7...

User avatar
vicd
Supporting Member!
Supporting Member!
Posts: 312
Joined: Sat Sep 03, 2011 9:26 am
Real name: Victor
Gear: Blofeld, Formanta (EMS01, Maestro, Mini), Fusion 8HD, JD800, Morpheus, MP7, Plugiator, TG77, TOM1501, TR626, UltraProteus, V-Synth XT
Band: ROBOPOP
Location: Siberia
Contact:

Re: When Did Synthesizers Replace Live Orchestras?

Post by vicd » Wed Nov 12, 2014 5:46 am

CS_TBL wrote:
calaverasgrande wrote:I still argue the primary roadblock at this point is in rendering the nuances of fingering, phrasing and delivery that a well oiled orchestra can.
If those musicians can describe in words why they do what they do (all the fingering, phrasing 'n such details) then it can be programmed. They aren't doing random movements you know, they have reasons to do the things they do. I think the next step in virtual orchestras isn't the big lot o' samples alone, I think some smart a*s will create a virtual player which triggers the right samples. This virtual player is actually a virtual human who really plays in the only way an instrument can be played. In other words: if your MIDI data couldn't be played in real life, such a virtual player couldn't play it either.

There's one catch though. A real player can anticipate to what's coming, and this player could choose to e.g. increase its loudness prior to a new note. Current software doesn't, it reacts to whatever is happening 'now', it doesn't look ahead. Years ago there was Synful, remember that? With the 1 sec delay-mode it could look-ahead. Naturally it worked awfully in combination with other stuff. So, the big breakthrough in DAW-land would have to be software that looks ahead and sends such info to the plug-ins. This would rule out real-time playing (as it doesn't know what you're going to do) but it would work perfectly during playback.

But we'll get there, sooner or later..
We're already there - Roland's Articulative Phrase synthesis on V-synth GT, Roland's SuperNATURAL on Integras'. Actually, we were there for at least the last two decades or so - the physical modeling like VL synthesis touches a lot of the nuancing ground; Band-In-A-Box with its algorithmically generated accompaniments, soloists and melodists is available since 1990 (and yes, it also does read-ahead to predict its variations, but strictly from MIDI data).
There is also a lot of (rather clumsy, but still) undergrad thesis work on this happening in the CS area - just because youngsters love to think that performance and composition can be formalized, and they sometimes achieve really interesting things.
Someone just has to put all the already existing knowledge and achievements together, and set up a reusable engine and framework with libraries of styles, fingerings, adlib fill-ins generation etc. - this is pretty much the way that third-party developers are doing with Yamaha's Vocaloid engine.

It's also tempting to make it able to self-educate by "listening to" and then decompozing/analyzing/formalizing various audio renditions of the same (sheet or MIDI) music.

Well, as you said, a "virtual musician". The future is now.
Last edited by vicd on Wed Jul 29, 2015 7:43 pm, edited 1 time in total.

I don't like the drums, but the drums like me!

User avatar
Walter Ego
Senior Member
Senior Member
Posts: 897
Joined: Thu Feb 09, 2012 7:17 pm
Gear: Various noisemakers hidden around my home and classroom. And a great quantity of caffeine.
Band: 12"VS
Location: Near Boston

Re: When Did Synthesizers Replace Live Orchestras?

Post by Walter Ego » Wed Nov 12, 2014 3:03 pm

Umm...since the '20s?

Image

Image

Image
Walter Ego
seamonkey wrote:I nominate this for STUPIDEST THREAD ever in the history of the internez. ;)

ItalianStallion81
Newbie
Newbie
Posts: 68
Joined: Mon Feb 19, 2007 5:26 am

Re: When Did Synthesizers Replace Live Orchestras?

Post by ItalianStallion81 » Fri Nov 14, 2014 10:47 am

If you go back to the 1980s, when synthesizers had only just begun to replace some live musicians, I'm wondering if there are any pictures available of such orchestra pits in that time frame that had these synthesizers?

The hot sheet:
RMI Keyboard Computer -- this 1974 invention, which used punched-out cards to read sounds, actually began the trend toward digital synthesizers, not the DX7 as many assume!
Fairlight CMI
NED Synclavier II
Yamaha DX7 (famously used for Les Miserables)
Kurzweil K250
Roland D50
Korg M1

~Ben

Post Reply