How to make 1990s style music soundtrack?

[tmelram]

From your example tracks, it kind of sounds like you should just listen to early Warp artists and imitate that style. Check out the Artificial Intelligence 1 and 2 compilations. The specific gear won’t be as important as getting the aesthetic right.

Thanks, I had a listen. I'm glad I heard a Roland R8 used a lot in them.

I'm not sure the Korg MS synths were particularly common or popular [in the 90s]

Well I checked Wikipedia and there were a ton of artist that were prominent in the 90s that used one that I've listened to (Prodigy, Aphex Twin, Erasure, The Crystal Method, etc).

Effects were Midiverb II, very common as it was cheap for a digital reverb at the time, Yamaha R100 reverb, Boss SE-70, guitar pedals into a Mackie CR1604 desk.

The thing to remember is if you had a compressor, you probably just had the one, so none of this compressor on every channel that you can do in your DAW nowadays. Likewise reverb tended to be the one effect, so you would just use the sends to alter the amount of each instrument going to the one reverb, rather than having a plethora of reverbs to choose from.

Thanks. Do you know how phaser, flanger, and distortion was done back in those times ?
(the 90s)

[tmelram]

[left blank due to error]

[Mr Arkadin]

I'm not sure the Korg MS synths were particularly common or popular [in the 90s]

Well I checked Wikipedia and there were a ton of artist that were prominent in the 90s that used one that I've listened to (Prodigy, Aphex Twin, Erasure, The Crystal Method, etc).

Well I'm sure, like many of us, they were just using whatever was at hand. If it was cheap you bought it! However, I think when you listen to those artists, the signature sounds aren't MS-20s. Most people lusted after 303s, 808s, 909s and Juno 106s, stuff like that. Lots of Roland sound. Liam from The Prodigy used a lot of JD-990 and loved the U-220 string sound and had five or six Boss SE-70s for effects. Of course this doesn't exclude the odder stuff and also when it comes to cheap digital people loved the M1 and Wavestation of course, as well as some of the odder Kawai stuff (I have an R-50e drum machine, LFO used the Kawai K1 to great effect).

Do you know how phaser, flanger, and distortion was done back in those times ?
(the 90s)

Well I would have used those effects from the Boss SE-70, also Phase 90, Boss HM-2 guitar pedals. I had two Boss T-Wah pedals so I could auto-wah on stereo keyboards! I was obsessed with T-wahs (also Crybaby Wah Wah for funk guitar). It was all about whatever you could get to hand.

[thecontrolcentre]

Back then I was using a Midiverb II and a Yamaha SX500 ... an old Vesta DIG delay unit too (still got it). It was guitar pedals for everything else. Synthwise I had a Korg Poly 61, Bass Station & Teisco 60F ... after starting with a Yamaha PSS680

[aciddose]

Kids today don't know how to multi-track. Even if you only had one compressor you could use it as many times as you had tracks available. The issue was noise floor growing on tape-to-tape processing, so usually any effects chain you laid down you'd apply before recording to a track.

Afterward however it was possible to mix these tracks together and apply any subtractive or linear effects (chorus, wah-wah, subtractive EQ = bandpass, notches, low/high shelving cuts) without additional noise. That would be the "mastering" stage before mix-down and going on the master.

How do you do this today? Cubase called it "freeze", in Reaper you can "render track to new take" and so on.

What either young or inexperienced people don't realize is that the modern way of using 100s of plug-ins on dozens of tracks all simultaneously is absolutely ridiculous. A lot of people worked with 6-track consoles so at most you'd have 6x6 = 36 tracks, in groups of 6 per "part" and the effects (inserts + sends) would be shared on each 6-channel "bus".

That process has a huge influence on how you develop parts and radically changes your focus. You can't go back and forth between tweaking the bass preset and adjusting the level of the snare... they're already recorded by that point.

[aciddose]

Also, for the most part ignore guys on here who have a severe case of GAS and buy every plug-in in the universe never using them. They're all about "sounds" or "gear" and rarely or never tell you about technique. Ultimately it doesn't matter in the slightest what effects or sources you use when you have no idea what the hell you're doing.

Do you want to produce like it's 1993? Set a budget, figure out what job you were doing and do the math. How long did it take to save this up? What was actually available and what were the prices?

If you're coming up with numbers like $100000 working as a painter part time, get real. A lot of the time people either had minimal equipment or used what the studio had. The "bedroom studio" is something that appeared near 2010, not 1990. The list you've provided is fine, but you need to add some core components to that like monitors, console/mixer, multi-tracks, destination (DAT tape?) and add up the cost. Or figure out what studio time would cost, what a studio had and start watching the clock.

https://youtu.be/VfiwhMFoawI

It's a bit rare to find raw "home studio" material from the 90s, but Sublime's first release "Robbin' the hood" was claimed to have been recorded in an apartment. The noise, techniques and instruments used here give you a really great example of 1990s "part time painter" sound:
https://youtu.be/Ah2MzAZDlAs

[tmelram]

Also, for the most part ignore guys on here who have a severe case of GAS and buy every plug-in in the universe never using them. They're all about "sounds" or "gear" and rarely or never tell you about technique. Ultimately it doesn't matter in the slightest what effects or sources you use when you have no idea what the hell you're doing.

I really only plan on using:

1.Korg-MS-20 Mini (an analog synth)

2 .Roland Sound Canvas (a 90s era rompler)

3. Drum machine samples from the Roland TR-909 and Roland R-8 (a lot actually come included with the Sound Canvas since they're made by the same company)

4. Editing and sampling sounds with FL Studio or some old Casio keyboard I have.


I definitely want to keep things simple and minimalistic and focus more on composition, instead of complicated music programming (like modern DAWs enable you to do). I wanted to go this direction specifically because I was overwhelmed with the amount of free/demo VSTs I had and deleted them all.

Do you have any specific techniques you would recommend when going for this style??

[aciddose]

You need to decide what you're trying to model. Are you the rich guy in the studio with all the gear and sequencers? Or are you the fool in a basement somewhere with used old trash who tries to slap stuff together piecemeal as best as possible (where electronic music really comes from.)

So you need to decide on your whole analog set up including which sequencers or other devices you're trying to model.

My advice is basically limited to:

• Minimize the number of tracks and complexity you use (do as much manually as possible.)
• Always think about noise. NOISE NOISE NOISE. That's the biggest thing you're always worried about. Don't use peak EQ without mixing in a -60+ dB noise floor first to see what it would really sound like on an analog console.
• Avoid anything that relies heavily on modern software features. Can you add a swing beat to the whole track using a quantize sequence function when you're relying on the TB-303's sequencer? No.

Stuff that seems simple in FL Studio might have been absolutely impossible to do with real gear. Those little things are what will create the unique techniques you use by forcing you to work and think differently that result in the sound you're looking for.

Every time I've attempted it, the results I've gotten have been way better. One thing I notice is having the console EQ in your face means you never miss the chance to fine-tune the EQ on each part as it's being recorded. Remember boost = during record, cut = subtractive, so doesn't add noise during mastering.

In software today you never need to worry about noise and it allows you to do things differently... which gives a different result in a lot of cases.

None of that is really direct advice applicable to your situation since I both do not know what you're aiming for exactly, and I haven't got the experience in every little niche to be able to say anyway.

Here's a very loosely on-topic example:
https://soundcloud.com/xhip/megamonopoly

I used 4x unison layering by doing 4x takes of each part. Only a single mono-synth was used on a single mono console channel. This really doesn't sound very good (it's a bit too bright and not "mixed" or "mastered") and it doesn't include any percussion. The important thing is that other than the 4x unison takes, the whole process was done in a way that is 100% compatible with old gear. Each take is unique, no presets, no "project files" with all the console knobs remembered... I just had to "wing it" which is what can produce really great results when you're trying hard. Unfortunately this isn't an example of that... since in this case I just wanted to do a quick&dirty demo.

[aciddose]

I don't know any details about it, but this is a good example track using that "chaos self-oscillation" sound you get when you use PWM into a low-pass with the cutoff slightly detuned:
https://youtu.be/0DH365BZYYM
https://youtu.be/sHkdH294UAc

This is quite minimal style stuff and vaguely in the genres I'd guess you're aiming for.

[tmelram]

are you the fool in a basement somewhere with used old trash who tries to slap stuff together piecemeal as best as possible (where electronic music really comes from.)

I am the poor fool.
All of the ambient tracks posted were pretty contemporary music made to sound 90s. So I know it can be done.
Especially this:
https://sndup.net/8tkn/beach.mp3

[aciddose]

Additional advice that might help is, music is just like legos. Once you have the right "parts", you just snap them together and they look really cool without anything else. You don't need to paint them like a model or anything, they're all pre-made just right so they all fit together.

This was made by just slashing together a few parts, total 4 stereo tracks with only a couple insert effects just to test an idea. All material taken from elsewhere, "pre-bounced" tracks:
https://soundcloud.com/aciddose-1/gaaahgagagapark
Doesn't sound great, but roughly gets the job done with minimal effort because the source parts "lego pieces" were already well made.


Usually any take that sounds great on its own will also sound great in a mix. So you need to make sure each track you record sounds like pure bliss before you mix it. It's always possible to make fine-tuning adjustments to fit a take using EQ (mostly all cuts) and basic effects like saturation, compression, gates and so on.

For multi-track takes you need to think in terms of using multiple multi-tracks. It was always possible to play a little ditty on a casio keyboard using a whole four takes and mix that down to a cassette tape (or a VCR was often used for the higher audio fidelity and ease-of-use.) That would add a little noise, but you could then copy that back either solo or mixed with another part on a single track on the multi-track at a later point.

So much like mixing records and scratching, you can have multiple sources into the console at once and be tweaking levels and eq while recording into a single take. By "bouncing" multiple takes like this it means you can't adjust them later, but you can free up additional tracks to use for additional parts.

One of the emergent techniques you get by manually doing that "bouncing" is when you're trying to line up the mix again to "bounce" back to the multi-track. Without a click track and all the timing gear it's hard to get it right... but then you notice things like how delaying a drum track or bass track to put it slightly out of time/sync can sound really cool. You'd never discover stuff like this without that process and techniques that allow you to learn. To get the same effect you might put a delay on your click track before input to the "bounce" device, or in a DAW you can just use a delay on a whole track set back by a beat to get -1 to +1 beat alignment control. Or a plug-in can be made with a 1-beat "lookahead" latency which gives you that automatically.

[GRUMP]

Hard to tell ... https://www.youtube.com/watch?v=yIyVjuoSoPc

But OK - we used to call it "Acid", "Acid Lines" - and usually suspected a 303 behind it.

And the KISS-Principle. Not because we were all clever, but because we just couldn´t be stupid!

[tmelram]

You need to decide what you're trying to model. Are you the rich guy in the studio with all the gear and sequencers? Or are you the fool in a basement somewhere with used old trash who tries to slap stuff together piecemeal as best as possible (where electronic music really comes from.)

So you need to decide on your whole analog set up including which sequencers or other devices you're trying to model.

My advice is basically limited to:

• Minimize the number of tracks and complexity you use (do as much manually as possible.)
• Always think about noise. NOISE NOISE NOISE. That's the biggest thing you're always worried about. Don't use peak EQ without mixing in a -60+ dB noise floor first to see what it would really sound like on an analog console.
• Avoid anything that relies heavily on modern software features. Can you add a swing beat to the whole track using a quantize sequence function when you're relying on the TB-303's sequencer? No.

Stuff that seems simple in FL Studio might have been absolutely impossible to do with real gear. Those little things are what will create the unique techniques you use by forcing you to work and think differently that result in the sound you're looking for.

Every time I've attempted it, the results I've gotten have been way better. One thing I notice is having the console EQ in your face means you never miss the chance to fine-tune the EQ on each part as it's being recorded. Remember boost = during record, cut = subtractive, so doesn't add noise during mastering.

In software today you never need to worry about noise and it allows you to do things differently... which gives a different result in a lot of cases.

None of that is really direct advice applicable to your situation since I both do not know what you're aiming for exactly, and I haven't got the experience in every little niche to be able to say anyway.

Here's a very loosely on-topic example:
https://soundcloud.com/xhip/megamonopoly

I used 4x unison layering by doing 4x takes of each part. Only a single mono-synth was used on a single mono console channel. This really doesn't sound very good (it's a bit too bright and not "mixed" or "mastered") and it doesn't include any percussion. The important thing is that other than the 4x unison takes, the whole process was done in a way that is 100% compatible with old gear. Each take is unique, no presets, no "project files" with all the console knobs remembered... I just had to "wing it" which is what can produce really great results when you're trying hard. Unfortunately this isn't an example of that... since in this case I just wanted to do a quick&dirty demo.

Thank you for the advice! Btw, I saw an F-zero song on your Soundcloud and I always wanted to know if they used MS-20 synth for the music?

https://www.youtube.com/watch?v=D_9PLdWoC3c

[tmelram]

(left blank due to error)

[aciddose]

There are two versions of the MS-20 (actually five revisions?) based upon two different filters that sound radically different with high res. The "KORG-35" filter is a Sallen-Key with resistors replaced by NPN transistors, making the circuit highly asymmetric. Later revisions use a standard OTA Sallen-Key filter which is far more common and easier to model in software.

The only other distinctive element of the MS- series is the triangle waveshaper which generally ages and produces a sharp narrow pulse at lower frequencies. This can be partially cured by replacing the electrolytic capacitors with larger fresh ones, but the circuit is flawed due to using RC high-pass filters rather than being DC-accurate so can never produce a perfect triangle shape.

In other words while specific versions of the MS-20 are distinctive sounding, generally they'll blend right in and be impossible to identify.

A third possibility for identification is the single NPN used in the VCA which tends to "snip" the tail off long releases. This would only be possible to identify easily if you had recordings of solo sounds with a long release.

The MS-20 is not capable out-of-the-box of sync, so it's impossible that the sync sound in your linked clip is a MS-20. The ressy bass sound is most likely not a KORG-35 but rather any synth (possibly a MS-20, possibly not) passed through an overdrive pedal or VST.

The main identifying characteristic of the KORG-35 is the asymmetric clipping causing oscillation to be focused on one polarity (positive vs. negative) of the waveform, along with cutoff frequency modulation by the input signal: http://xhip.net/temp/mp3/ms10vcfselfosc.mp3