Jump to content

Synthesizer programming Megathread.


Yoozer

Recommended Posts

  • Members

Welcome to the Synthesizer programming Megathread. Yes, we all know about the vast archive of knowledge at Sound On Sound, but this only covers so much.

 

The Idea

You heard a sound and want to know how it's made, or your don't have the fancy huge $10,000 synthesizer your idol has and want to look for something that can do the job

 

OR

 

You have a bunch of interesting ideas in terms of programming synthesizers and making certain kinds of noise?

 

If one of these people is you, welcome to this thread.

 

The Rules

 

Requests

If you have a request on how to program a certain sound, provide an example.

 

Either use one of the many services available to put up a snippet of an mp3 file that contains the sound, or refer to a good quality Youtube clip (no cellphone camera recordings).

 

http://www.putfile.com/

http://www.zshare.net/

http://www.rapidshare.com/

http://www.divshare.com/

 

Important (aka making an ASS out of U and ME)

Do not assume people will know the song. Like nobody might've heard of that obscure underground hiphop group somewhere in the Detroit slums, there are also people who never heard the B-side live version of Famous Track by some prog rock supergroup.

 

Do not assume people will do the job of searching for a song for you. Also, do not assume everyone to understand immediately what you're talking about when you say "well, that one chord sound" - specify the exact time when it is heard in a track for the first time, or a timerange. If you have a 30 second mp3 and the sound kicks in at 0:12 plus it sounds like a bass, that's generally enough information. If you have 2 nebulous sentences and a misspelled arist name, no song title, you fail. Take the effort to thank people if they give you a description that works, let us know if you have made any progress.

 

Also, not every sound rolls as a 100% duplicate straight out of a synthesizer, so learn about stuff like equalizers, compressors and effect devices.

 

Responses

Be complete. Be correct. Make sure your description either assumes an initialized patch or a soundalike and fill in at least the most important details. Make sure your description actually works (test it on your own synthesizer). If it's bull{censored}, you'll be told so.

 

Offers/Programming tricks

Your programming tricks depend often on some detail. You do not have to specify numerical values in 6-digit accuracy - switches are on or off, and you can write the values of an ADSR as 0, 10, 5 (0%, 100%, 50%) or sketch the general curve of the thing using MS Paint or an equivalent. Use screenshots for software synthesizers. Image hosting (no silly, we can't read your C:My DocumentsUserPictures folder) can be found at ImageShack or Photobucket.

 

Drawing diagrams is not rocket science, and for all I care you take screenshots in Word. Put up your results in audio form; otherwise you're describing a juicy steak without letting anyone take a bite from it.

 

Drawing stuff

For screenshots - save as jpg or png in high quality, MS Paint mangles these things like a dog. Alternative: http://www.getpaint.net/ or http://www.gimp.org/ (but that may be a bit overkill).

 

For drawing diagrams, you can use Inkscape, or if you have Word, use that to do the job.

 

Recording music

If you're a luddite and have no computer, or you have one but only use it to post on messageboards, here's what you can do. Up there you already saw hosting services. No way your 100 mb wave file is going to fit in there, and even if it does, it's useless.

 

You can use Audacity ( http://audacity.sourceforge.net/ ) to record your examples or remove the silent bits. You can also use this to save as mp3 - you'll need to install the so-called LAME encoder, then. A different method is to use this separately - you save the .wav in Audacity and then drag it into something like RazorLAME too. This has as an advantage that it can do multiple wave files in a row and you don't need to hit Save and wait every time.

 

Take a little time to figure out how it works, it's not complicated - and you can reduce your 100 mb .wav file to a mere 12 mb or so.

 

Educational Tools

The Reason 4 demo version works for 20 minutes, a time I've found generally sufficient to tweak a sound until it matches and then take a screenshot. Also, it works on a Mac.

 

The Synth1 synthesizer has a clear, no nonsense UI and is free to download.

 

Both are excellent educational tools; if you have anything else, please suggest it.

 

Go.

Link to comment
Share on other sites

  • Replies 59
  • Created
  • Last Reply
  • Members

I'll go first with this 'Victim of Love' Intro sound I mentioned in this thread.

 

Sound can be heard here, it is the first sound you hear in the intro:

 

http://www.youtube.com/watch?v=VDKioRMxj1c

 

It seems like a very basic sound, but for some reason, it doesn't sound like a basic "VCO-VCF-VCA" type sound to me for some reason, but perhaps someone can prove me wrong. I though it was perhaps a D-50, but it turns out the D-50 wasn't out yet when this song was released. Someone mentioned it possibly being a Prophet VS, which I could believe.

 

If it happens to be a vector or wave type synth, is there a decent free vst that has wave table synthesis?

 

Anyways, I know it seem like a basic sound, but I want to see how it is programmed.

Link to comment
Share on other sites

  • Members

I'll go first with this 'Victim of Love' Intro sound

 

To my ears it sounds like sampled fiddle "chops"...a staccato bowed note on a violin (possibly a viola or cello), maybe someone who has classical training can provide the correct term ;)

 

Any sampled violin sound with a sharp bow attack should work, set the release to 0 and play short notes.

Link to comment
Share on other sites

  • Members

It sounds analog to me. It has a really quick decay or release on the VCF/VCA so it fades out quickly. It's also played polyphonically (chords) so that makes it a bit harder to isolate the individual voices, but start with a basic violin-like analog patch and then give it a quick decay and no sustain and/or release. It sounds like the filter is affected by the envelope, with little or no resonance.

Link to comment
Share on other sites

  • Members

To my ears it sounds like sampled fiddle "chops"...a staccato bowed note on a violin (possibly a viola or cello), maybe someone who has classical training can provide the correct term
;)

Any sampled violin sound with a sharp bow attack should work, set the release to 0 and play short notes.

 

Funny how that works. I used to wonder how Robert Rich got those beautiful overdriven, sustained and bending sounds. Later found out it was a steel guitar run through some of his modules. :)

Link to comment
Share on other sites

  • Members

Resampling

 

After the hardware sampler died and got second place in workstations (or a limited role in the Akai MPC series), and software samplers became mostly multi-gigabyte romplers, people tend to overlook the potential and the earlier role of the sampler in the 1990's studio.

 

This was an expensive device that could take over the job of a synthesizer - you can offload certain duties, freeing up your machine to do something else (no excess luxury when it's monotimbral anyway).

 

You're no longer bound to the {censored}ty drum computer sample set but if you're too poor for a 909, you can load a set in there and have (most of) one.

 

Of course we all know about the 909 trick already, but there's more you can do with samples. One important role is played by your software sequencer - as long as it can freeze tracks, it's good. Best part? It works in a 100% hardware setup, it's just more work that way.

 

The mp3 demo is at the bottom, and it's divided in 3 parts.

 

1) We start with a plain sinewave - single oscillator.

 

2) Take a reverb with a long, long tail and add a compressor so that the volume stays constant. What you want is a reverb tail that lasts pretty long, so choose a big hall preset with a pretty wet mix (40%) and a long decay time.

 

Play the sinewave as a stack of chords. It depends on your mood, but anything bigger than 5 notes is usable. Experiment with the following: instead of a sinewave, use a narrow pulsewave (also single osc), or play a dissonant chord. Don't hold the chord down - it's the reverb that should do the work, and less original signal = more ethereal signal.

 

3) Resample. What you do is you record the result, cut off the attack and choose a fragment of the tail which you then start looping. If you can loop and fade in/out to smooth the transitions, excellent - the result will shimmer pretty much every time anyway. Software/hardware sampler, I don't care.

 

3b) already done in this example: filter with a bandpass filter and let the ADSR control the filter cutoff. Cutoff should go down after pressing the keys, up again when you release them. Add a little reverb to smooth out any kinks, maybe a chorus, and amplitude modulation to add another shimmer (not in sync with the one already in the sample).

 

If you do this correctly, you'll notice where a lot of the waveforms in the JV/XP boards may come from, and you might not need Omnisphere for a while ;).

 

Another option I learned in the Nord Modular manual, but that's hard to do on most systems. What you do is the following:

 

- you sample 3 tracks of white noise

- you use an equalizer to deliver a boost to a certain range - it helps if the EQ starts to resonate when it's cranked up to that point (the Nord used bandpass filters with nearly full resonance, but some EQs may do this just as well).

- you try to "tune" the tracks EQs in such a way that you get a chord, a fifth or whatever goes together nicely. Resample this, put reverb on it to smooth it out again, and you have a similar result.

 

This gives you breathy choir pads/sounds/ambience with ease.

 

Demo: http://www.theheartcore.com/megathread/resample_verb.mp3

 

Picture:

 

resample_verb.png

Link to comment
Share on other sites

  • Members

Someones already explained it as modulating osc1 with an lfo

This generally means that something's happening with the pitch (or the PWM), and it does not sound like that.

 

Try a fast (this is about 8hz) phaser on a basic string/brass sound, like this:

 

http://www.theheartcore.com/megathread/numan_phaser.mp3

http://www.theheartcore.com/megathread/numan_phaser.png

 

I've used the Synth1's default "1: Synth1 Brastring" preset with the cutoff a little higher - with Live's phaser. It's fast, and whatever phaser you pick - you have a "silver lining" effect with most of 'm, and you should find a way to avoid this (because you need a more nasal sound). I think but I'm not sure that this is made in a similar way - only of course with much more interesting and vintage gear ;).

 

On the AN1x you could either use the phaser or the bandpass filter with a little dash of resonance. Ideal would probably be parallel bandpass filters with LFOs out of phase, but you can't have everything :D.

 

How do you get that sound in "Jump"?

 

Try to rebuild this on your Nord Lead 2X:

http://www.theheartcore.com/megathread/vanhalen_jump.png

Link to comment
Share on other sites

  • Members

 

Anyway, first 8 seconds on Mr Numans finest

 

 

The actual sound was produced using a Polymoog Keyboard (the Vox Humana preset) and a phaser stomp box at a fast setting (a MXR Phase 90, if I remember correctly). Start with a two osc PWM pad... Each osc's PWM modulated by their own LFO, global pitch vibrato adjusted to taste, and for further authenticity have the PWM LFOs track the keyboard (wilder modulation in the upper octaves) and use a phaser...

 

But I bet you could use just a detuned saw pad into a phaser...

 

- CM

Link to comment
Share on other sites

  • Members

The actual sound was produced using a Polymoog Keyboard (the Vox Humana preset) and a phaser stomp box at a fast setting (a MXR Phase 90, if I remember correctly). Start with a two osc PWM pad... Each osc's PWM modulated by their own LFO, global pitch vibrato adjusted to taste, and for further authenticity have the PWM LFOs track the keyboard (wilder modulation in the upper octaves) and use a phaser...


But I bet you could use just a detuned saw pad into a phaser...


- CM

 

I think he means the weird oscillating sound in the first eight seconds.

Sounds a bit like a thin pulse wave going into a highly resonant filter which has a fast delta LFO on the cutoff frequency.

EDIT: And run it through a high-pass filter before going into a resonant low-pass one. There might be ring-modulation involved as well, but I'm not sure.

 

This question led me to try it on my Andromeda, getting off course as usual and ending up with a cool dogfight sound effect :facepalm:

Link to comment
Share on other sites

  • Members

 

I think he means the weird oscillating sound in the first eight seconds.

Sounds a bit like a thin pulse wave going into a highly resonant filter which has a fast delta LFO on the cutoff frequency.

EDIT: And run it through a high-pass filter before going into a resonant low-pass one. There might be ring-modulation involved as well, but I'm not sure.

 

 

Well, maybe. But when I had a 280a, I got that sound using Vox Humana and a phaser. It was a different phaser, so the sound wasn't exactly the same, but it was veeery close. Listen to the PPNM file here.

 

- CM

Link to comment
Share on other sites

  • Members

Well, maybe. But when I had a 280a, I got that sound using Vox Humana and a phaser. It was a different phaser, so the sound wasn't exactly the same, but it was veeery close. Listen to the PPNM file
here
.


- CM

 

Wow, sorry, my bad. I didn't expect that sound to be based on the Vox Humana sound, that's pretty amazing. He really must love that preset :love:

Link to comment
Share on other sites

  • Members

1. the erasure sound is a simple one. the amp is modulated with an env with an attack at roughly 20% time slope. the filter is opened at slowly with the same envelope but is only allowed to open to 70% maximum (adjust cutoff to around 50% and use positive envelope amount to boost slowly)

basic wave is some saw waves with very minimal detuning

this was probably made on an arp 2600 as it's vince's favorite howvere it's an early song so maybe a roland sh series.

 

2. gary's cars warbling sound...

the real trick to getting it right is bandpass filtering. use a wide bandpass with 30% resonance on the basic sound (the polymoog has a filter bank on it and the basic sound vox humana was altered using this 'resonator' bank) to get it close then use the phaser set fast or alternatively use a second bandpass filter with no resonance and a smaller width than the first and set it's cutoff to an lfo.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...