A Faster S.S. Head?

[seamonkey]

Fun applet

Fourier Series Applet

 

[Bob Lee (QSC)]

Square Waves are the "Man"!

In "Magic Man" and "Lucky Man" the synth starts with a square wave. Many analog synth hit solos are square.

Those were either sawtooth or pulse waves, but not square. Square waves have a more hollow, reedy sound--clarinet-like, really--and we don't often hear them used in analog synths.

 

[JHAz]

You should try it yourself. It should be trivial to recreate the apparatus today, compared to 40 years ago.

As I said farther above, I spent some time listening to massively mangled (from a phase perspective) square waves on a web site I don't remember any more and could not hear a difference. I don't feel an intense need to recreate that experience. Again, changing the amplitude of the harmonics of course changes the sound. It's part of what makes a piano sound different from a tuba.

 

[Bob Lee (QSC)]

As I said farther above, I spent some time listening to massively mangled (from a phase perspective) square waves on a web site I don't remember any more and could not hear a difference. I don't feel an intense need to recreate that experience. Again, changing the amplitude of the harmonics of course changes the sound. It's part of what makes a piano sound different from a tuba.

I agree. Timbre is mostly defined by the harmonics that are present (including the fundamental) and their relative amplitudes. Relative phase is not important unless it's so great that it produces noticeable gaps.

 

[okcrum]

I agree. Timbre is mostly defined by the harmonics that are present (including the fundamental) and their relative amplitudes. Relative phase is not important unless it's so great that it produces noticeable gaps.

If the reproducing electronics, transducer(s) and your cochlea are all perfectly linear in response, this is true. Practically speaking, the occurrence of all three is exceedingly rare. For super extra credit, what is it, then, that we Link Removed?

 

[agedhorse]

A square wave is ideally an infinite number of related signals. When you mess with them relative to each other, you sure as shootin' can hear it in an A/B test as long as the system can reproduce enough of the component frequencies. How do I know this? One of my work study projects in school was building a Fourier series demonstration unit for the physics lectures, which allowed you to mix a 400 Hz fundamental + harmonics and vary the phase and amplitude of each harmonic separately, for the first 16 harmonics of the fundamental. The lectures got to see the changes on a display and hear them too. And hear them you could. Not just the differences between saws, triangles, and squares either.

Correct IMO, because you are hearing them all RELATIVE TO EACH OTHER and not out of context.

 

[agedhorse]

I agree. Timbre is mostly defined by the harmonics that are present (including the fundamental) and their relative amplitudes. Relative phase is not important unless it's so great that it produces noticeable gaps.

If you can shift the phase of say a 200Hz harmonic with respect to the 100Hz fundamental, add them together, you will (IME) pretty clearly hear the difference as the harmonic is shifted.

I suspect that we may be describing different things however...

 

[Bob Lee (QSC)]

If the reproducing electronics, transducer(s) and your cochlea are all perfectly linear in response, this is true.

They don't necessarily have to be perfect. Very high quality is fine. You actually can shift relative phase of harmonics in a signal without causing audible changes to timbre.

If you mix the shifted signal with the original, though, it can be very audible.

 

[bongomania]

Square waves have a more hollow, reedy sound--clarinet-like, really--and we don't often hear them used in analog synths.

Whoa, whoa, whoa. Sorry but I have been using analog synths since 1985, and every bleeding one of them had a square wave, and many of the most famous analog synth sounds include it. Flutey/airy tones, bells/clangs, screamy leads, all routinely made with square waves (as well as other waves).

 

[dukeorock]

I think 'fast' is a perfectly good term...the first time I fired up my GB SM12.0 I said to myself, "Man, this thing is fast!" I had to laugh because I knew that term had no literal meaning with regard to sound, but we use 'dark' and 'muddy' when the lighting doesn't change and there's no mud to be found Heck, even 'punchy' doesn't work for real...we've just used the term long enough to agree on what it means.

Now I'm going back to the studio to calibrate the perfect 'speed to bap' ratios on my rig

 

[Bob Lee (QSC)]

If you can shift the phase of say a 200Hz harmonic with respect to the 100Hz fundamental, add them together, you will (IME) pretty clearly hear the difference as the harmonic is shifted.

In that example you could hear a shift in the second harmonic's frequency as its phase is being shifted. The more rapid the change in phase shift, the greater the frequency change.

What would not sound different is, say, the first harmonic + the second with no phase shift versus the first plus the second with 20° relative phase shift, unless it was accompanied by amplitude-related anomalies.

 

[Bob Lee (QSC)]

Whoa, whoa, whoa. Sorry but I have been using analog synths since 1985, and every bleeding one of them had a square wave, and many of the most famous analog synth sounds include it. Flutey/airy tones, bells/clangs, screamy leads, all routinely made with square waves (as well as other waves).

Yes, that is true, except that flute tones tend to start with triangle or sine waves.

The really classic, brassy-sounding analog synth sounds--"Just Enough for the City," "Lucky Man," "Frankenstein," "Suffragette City," et al--are sawtooth or pulse, with ample amounts of both even and odd harmonics. As a kid I noticed these sounds are so rich in harmonics that they just leaped out of even a cheap AM transistor radio with a 2" speaker.

 

[msmucr]

Hello,

interesting discussion here. I hope, it won't mind to add something to OP.

IMO and IME (Maybe pure IMO and IME to borrow some nice English phrase from Mr. KJung, perceived "quickness" of produced sound by amp (or other device in audio chain) relates "mainly" to its varying transient (short peak) response and specific distortion products and not so much to "static" delays and phase. It inherently relates to both frequency and time and psychoacoustics plays significant role in perception.

Practically, it is quite common during studio work, to use some devices for creative transient manipulation.. It could be in both
directions, soften or enhance (examples.. "undigitize" modern synths, make more pleasant vocals versus adding more punch to bussed drums). It's not necessary to use those devices to extremes, sometimes you can use compressor without hearing actual gain reduction, because just passing through circuits works pleasantly for your intention (result feels slower), sometimes you can use preamp or eq without actual volume gain, because it nicely emphasize midrange aggression that fits to mix (feels faster). In those cases, if you do standard measurement of frequency response with white noise, you'll find, that those devices are very close to flat. Still you can have feeling, that some frequencies are slightly pronounced or tammed in you signal. My answers to this contradiction are in transient response and harmonic distortions in circuits, which naturally emphasize some fundamental frequencies. Similar things could be also found if you'll shift your point of view to amplitude response in time.

To general discussion, I didn't spent much time with amp construction or repairs except for some stuff during high school years, but I enjoyed time with several programs and hardware for DSP "emulation" of various studio audio devices.. (actually it was started when i saw first guitar FX by Line6 it was technically very adventurous to explore techniques and processes used for sampling those "real" audio devices (static convolution, dynamic convolution, Volterra kernels) or algorithms used for electronics circuits modelling in software (a la real time SPICE). It was quite revealing for me as real analog circuits are very complex to replicate using mentioned techniques and it's always some kind of compromise (especially due to available computing power - you can have real time dual triode simulation which eats your Desktop PC processor power). But that wasn't main thing for this topic. Interesting point for me was always comparison with original devices, and folowing finding about how many variables are involved. For instance.. how my typical measurements and assumptions doesn't correctly represent different states of circuits.. how differently emulation behaves with sweeping pure tones and then with real complex time varying signals.. what will happen if i change gain staging and overload some parts in emulated device. It definitely shifted my view when describing and comparing sound of amps to more holistic point.. from, let's say, that sounds nice due to its frequency response. Just for fun, when i started to read posts on this forum.. my biggest quest was to imagine something related to sound behind English words woolly, nasal and thumpy, so maybe i'll also find something about "slow" at this thread...

Have a nice day

Michal

 

[Bob Lee (QSC)]

Here's an interesting article on psychoacoustics:
Invalid Link Removed

 

[seamonkey]

Yes, that is true, except that flute tones tend to start with triangle or sine waves.

The really classic, brassy-sounding analog synth sounds--"Just Enough for the City," "Lucky Man," "Frankenstein," "Suffragette City," et al--are sawtooth or pulse, with ample amounts of both even and odd harmonics. As a kid I noticed these sounds are so rich in harmonics that they just leaped out of even a cheap AM transistor radio with a 2" speaker.

i.e. Lucky Man = Square Wave.
It's a preset on about every synth since the song released and every VSTi synth out there.

Give one a try.

Invalid Link Removed

Square waves are fat and sweeping filters love them.

Not sure why there's so much hate for squares.

 

[Bob Lee (QSC)]

No, "Lucky Man" would sound like a square wave if it were a square wave.

 

[seamonkey]

Read the PDF that describes the sounds on the link.
It sounds like a square wave 'cause it is.

 

[Bob Lee (QSC)]

Thanks. The PDF explains why it doesn't sound like a square wave. It's three square wave oscillators in near unison, so it has more complex content than a single odd-order series. I thought it was two pulse oscillators in near unison. I stand corrected but still would stress that a square wave by itself would sound distinctively different.

Keith Emerson has just gotten the Moog when he recorded the "Lucky Man" solo and hadn't yet figured how to use the envelope generator and VCA. That's why there's no amplitude modulation, just one continuous tone with portamento to create glide between the notes. In a "Contemporary Keyboard" interview back in the mid 70s he said something to the effect that he would've liked to come up with something much better.

 

[bongomania]

Bob, just FYI, lots of flute patches back in the day started with a square wave. Not always triangle, never sine. I realize this is a teeny tiny point, not really worth belaboring, but seriously square waves were one of the main staples of analog subtractive synthesis, from Buchla and Moog on forward. That hollow "tubular" sound was immensely popular and widely used. Back in the late 80's I used to write banks of patches and swap them (on cassette and cartridge) with many other synth heads, and there was no shortage of square waves throughout. In fact, the only patch banks I can remember NOT having much square wave action were groups of string, horn, or piano patches.

 

[Bob Lee (QSC)]

OK, but when I created analog flute patches I never used square waves because there was too much harmonic content that I would've had to filter out anyway. I usually used a triangle wave, because it was the closest the EML synths had to a sine, along with some noise for breath and ADSR shaping for tonguing.

A flute tone is almost purely sinusoidal. A triangle gives a more recorder-like sound, and a square wave is more like a clarinet. I never did get to analyze saxes.