How to get rid of the metal sound of a supersaw - Sound Design

[Mohammadyarahmad]

Hi,
Basically, the supersaw is created as following steps:
1. Select a saw wave, and add unison which is slightly detuned.
2. Select another saw wave. One octave higher. And again, add unison which is slightly detuned.

But we will find a little metal sound which is not desired.

How do you remove this metallic-sounding from your supersaw?

[Heigen5]

Hi! Without saying how to do it without sounding metallic, have you downloaded this community Refill in this thread yet? viewtopic.php?f=8&t=7517930
There's many patches inside that Refill so you might find less metallic ones too.

[Loque]

Modify the basic saw if possible, add a LPF before the saws are mixed and maybe also after they are mixed. A chorus like FX can also make it smoother. Light saturation can help. Multi band processing with compression and other FX can help. Add a reverb to wash or the sound, maybe just a shorter one with high damping and a bit more of wet...

I guess there are plenty ways to smooth a sound. There also specialized plugins available for such tasks.

Best way to change the sound is take a different synth. Sometimes a synth just sounds harsher at all for whatever reason

[RobC]

Metallic sound most commonly comes from a comb filter effect. That effect is caused by a slight delay or for example when oscillators are out of phase.

In this case, it seems like the "core" saws are left raw, and the unison effect won't solve the issue afterwards. Unless the oscillators are independent.

Either way, try detuning the oscillators from each other.
That, or add more unison layers and detune them more.

That said, unison could be just multiple modulated delays, which also can have phasing issues, adding a nasty comb filter effect, which results in a metallic sound.

If you want a perfect super saw, then layer and detune as many as you want. But make sure that the tuning is different for each oscillator. For example +25 -25 +12 -12 +6 -6. The lower the octave the more you detune; the higher, the less the detuning value (so that it doesn't sound off).

You can further enhance it with vibrato.

Forget unison. It causes more trouble that good.

[selig]

For me I’ve never added the octave up oscillator. I’m more ‘hard core’ because of owning the original Super Saw machine for many years (JP-8000). The other possible contributing factor is the detune amounts. When you detune by the same interval for every detuned oscillator you add, you get a less organic sound IMO (the comb filter effect). This is why the original detuning tables were so vital for my tastes, as originally defined by Adam Szabo in his published paper on the original super saw many years ago (google is your friend).

There is no low pass involved in the original (unless you dial it in yourself) but there IS a high pass that is said to contribute to the original sound.

And finally, the magic for me is somewhere between 5-7 detuned oscillators - any more/less and you loose the “magic”, but again I’m a fan boi of the original so take my opinions with a large grain of salt!

[bitley]

I understand what you mean OP and it is the sound of non movement. We need to add movement to the character and thus pulsewave with PWM is a thousand times better than the boring sawtooth waveform. Nick Batt has driven me insane with his sawtooth. The only interesting thing with a supersaw design is if we can add proper and continously moving detuning. Only then can we achieve or superswell the lively sound you get from PWM. Dive deeply into a 1 osc poly like Alpha Juno to open your mind to how important PWM is. Unison has to be blended in very carefully because too much of it creates an undesireable and mushy sound. I rather listen to Yazoo's simple but effective clear raw synth sounds for life.

[selig]

The lower the octave the more you detune; the higher, the less the detuning value (so that it doesn't sound off).

First, it should be the other way around. Second, no super saw adjusts tuning by frequency, the key is to detune by pitch and not by frequency, which is the default in most synths. In fact, not every synth gives you the option to detune by frequency, so it is typically not an issue!

Thor to the rescue once again:
You can explore the difference between pitch vs frequency detune in Thor if anyone is interested in the subject. Use the mod matrix to assign a Rotary to oscillator FM (frequency mod, not pitch mod) on one oscillator but not the other(s). You get an interesting effect, which is that the beat frequency is constant at every pitch/octave! BUT, you’ll also notice how much more detuned the oscillators sounds at lower frequencies than high since there are fewer cycles per second between lower pitches than higher.

By detuning by a pitch based ratio (cents/semitones/octaves) rather than frequency (Hertz, CPS), you get different beat frequencies for every different note which sounds richer/fuller to our ears because of this.

Adam Szabo explains this in his excellent paper:
https://www.adamszabo.com/internet/adam ... er_saw.pdf
(From page 8):
“The spectrum shows that the center oscillator is at 523.3572 Hz playing a C5 note, and the six
other oscillators have been detuned by a set value. This value cannot be defined by a frequency,
because the interval in an octave has half or double of its frequency. A number must be found
which can be used to detune the oscillators by the same amount regardless of the current
frequency.
To find this number, one must look at the relationship of the six oscillators with respect to the
center oscillator. The center oscillator must always be “1”, because it will have to be multiplied by
the frequency of the played note. If each oscillator is divided by the center oscillator, the
relationship between them is found, however, since the six oscillators must have the same
frequency as the center oscillator when the detune value is 0, it must be written in the form: 1+-
offset. This offset is then multiplied by the detune value which has to have the range from 0 to 1. If
the detune value is 0, all the offsets will also be 0 making the overall detune amount 1, thus having
the same frequency as the center oscillator (table 1).”

[Mohammadyarahmad]

Thanks all for the great responses.
Reason's users are elite musicians <3

[RobC]

The lower the octave the more you detune; the higher, the less the detuning value (so that it doesn't sound off).

First, it should be the other way around. Second, no super saw adjusts tuning by frequency, the key is to detune by pitch and not by frequency, which is the default in most synths. In fact, not every synth gives you the option to detune by frequency, so it is typically not an issue!

Thor to the rescue once again:
You can explore the difference between pitch vs frequency detune in Thor if anyone is interested in the subject. Use the mod matrix to assign a Rotary to oscillator FM (frequency mod, not pitch mod) on one oscillator but not the other(s). You get an interesting effect, which is that the beat frequency is constant at every pitch/octave! BUT, you’ll also notice how much more detuned the oscillators sounds at lower frequencies than high since there are fewer cycles per second between lower pitches than higher.

By detuning by a pitch based ratio (cents/semitones/octaves) rather than frequency (Hertz, CPS), you get different beat frequencies for every different note which sounds richer/fuller to our ears because of this.

Adam Szabo explains this in his excellent paper:
https://www.adamszabo.com/internet/adam ... er_saw.pdf
(From page 8):
“The spectrum shows that the center oscillator is at 523.3572 Hz playing a C5 note, and the six
other oscillators have been detuned by a set value. This value cannot be defined by a frequency,
because the interval in an octave has half or double of its frequency. A number must be found
which can be used to detune the oscillators by the same amount regardless of the current
frequency.
To find this number, one must look at the relationship of the six oscillators with respect to the
center oscillator. The center oscillator must always be “1”, because it will have to be multiplied by
the frequency of the played note. If each oscillator is divided by the center oscillator, the
relationship between them is found, however, since the six oscillators must have the same
frequency as the center oscillator when the detune value is 0, it must be written in the form: 1+-
offset. This offset is then multiplied by the detune value which has to have the range from 0 to 1. If
the detune value is 0, all the offsets will also be 0 making the overall detune amount 1, thus having
the same frequency as the center oscillator (table 1).”

Hold on a second. I meant detuning by cents, not Hz. And I also meant detuning pitch.

The phase thing was about helping the OP find the cause of the metallic sound, aka comb filtering.

That said, if we're already on topic, it's fun to play around with both oscillator phase and pitch, mixing both fixed and modulated instances. Not to mention in stereo and mono, even mid/side.

I once made an organ type sound, where I also layered multi-octave samples. It actually sounded pretty cool. It even worked for playing chords - though I remember that bass and higher octaves had different samples.

That said, Thor already had a module for supersaw, which could be tuned for any taste. Whether multi octave, or single octave with lots of color.

I think Europa does it, too to some degree.