Simple questions on synth tech basics...

[aciddose]

It has always been considered somewhat immature and unfortunately as far as I'm aware remains quite out of date today.

Such a project is not at all simple. The fact it is freely available means you can count another possible tool in your set if it works.

If not, there are numerous other more expensive tools to do the same thing.

[Mayae]

Regarding the panning, just for fun I just put three instances of a stereo widening plugin in series after a synth and set each to 200%. Oddly, the pad sound did not really become any wider when I turned on the 2nd and 3rd instance, but sounded increasingly harsh. I knew from Sylenth that too much stereo width can make sounds "fall apart", but the harshness surprised me

In the stereo widening plugins I've used, 200% means only side signal survives, ie. the mid signal is completely erased - thus you can't "widen" it anymore after that. Any vectorscope will tell you the truth, however

[JCJR]

Maybe there is variety in design of Mid/Side tools, dunno much about them. For simple mid/side tools which do something like: Subtract Right from Left for the Side signal. Add Left and Right for the Mid signal. Adjust Mid/Side levels and then convert back to Left and Right outputs.

One possible interpretation of 200% widening, might be that the Mid Left+Right sum would be entirely missing in either speaker-- And the same "mono" [Left - Right] signal might appear on both speakers, except the Right version is 180 degrees out-of-phase from the Left version. IOW, the Left channel contains [Left - Right] and Right channel contains [Right - Left].

A purely mono input, "widened" thusly, would be silent. A "mostly mono" input would be greatly reduced in volume, perhaps drastically exaggerating any stereo reverb on the track. For instance if the input signal happens to be mono with some stereo reverb added. It would cancel everything except the stereo components of the reverb.

Any music common to both channels would be eliminated. Which is the working principle of many simple "vocal eliminator" effects. Subtract Right from Left (or vice-versa) and any purely center-panned vocals or other instruments would disappear except for any stereo reverb. Such a simple FX also kills off any center-panned rhythm/lead instruments, and any center-panned bass track, kick, or snare.

Some such "vocal eliminators" would high-pass one of the channels, to avoid removing the bass and kick.

Some were a little fancier, with adjustable bandpass center elimination-- Attempting to cancel midrange vocals, while preserving as best possible both the high-frequencies and low-frequencies of center-panned instruments. Though that would still play hob with any center-panned midrange instruments, guitar or sax or synth leads or whatever.

In an effect where 200% widening results in phase-flipped [L-R] on both sides-- Even if the input signal has a generous amount of stereo information, such a drastically modified track wouldn't be mono-friendly. If L and R are the same signal, except phase-flipped, then a mono mix of that "widened" stereo output would cancel and go silent. And it would be subject to fairly severe cancellations over speakers, depending on how well the Left and Right speakers mix together at the ear. The acoustics of the room would strongly influence what one would hear.

Some "synthetic stereo" effects use different techniques, using delays, phase shifts, or filtering to make fake stereo even with a mono input.

The result of "radical widening" could vary, depending on how a plugin or device operates.

[fluffy_little_something]

Sorry, had not seen your post before. I think I should delete some threads, sometimes I get so many reply notifications that I miss something

What you describe sounds messy
A mono signal can also be very wide, can't it? Doesn't mono simply mean that there is the same signal on both sides, no matter how far they are apart?
Yes, in general reverb seems to reduce the volume of the signal.



Anyway, why is it that with synths there is a certain high-frequency drop from the 6th octave or so up? I noticed it on various synths that when doing bells they sound metallic and crisp up until the 5th octave or so, but after that the typical sharp character gets lost. For some time I thought that maybe it is my aging ears that simply don't hear it anymore, but with my rompler that is not the case, bells sound crisp and metallic across the keyboard. Does it have to do with the way the oscillator makes the sound? Other than that a synth and a rompler work the same way, right?

[JCJR]

Anyway, why is it that with synths there is a certain high-frequency drop from the 6th octave or so up? I noticed it on various synths that when doing bells they sound metallic and crisp up until the 5th octave or so, but after that the typical sharp character gets lost. For some time I thought that maybe it is my aging ears that simply don't hear it anymore, but with my rompler that is not the case, bells sound crisp and metallic across the keyboard. Does it have to do with the way the oscillator makes the sound? Other than that a synth and a rompler work the same way, right?

Hi Fluffy

Dunno much about it as usual. Once you get fairly high in pitch, there are not many high harmonics remaining in the audible range. The high harmonics, above the fundamental, are primary timbre cues to the ear. The higher harmonics making the difference between a bell tone, or a trumpet tone, versus the tone of a pure organ sine wave.

For instance, sample rate 44.1 kHz, a 40 Hz wave could potentially have up to 500 harmonics before being brickwalled by the samplerate nyquist. Or alternately, pitches getting so high that "well-abused" ears can't hear any higher.

On the other hand, a pitch 3 octaves above A440 would have a fundamental of 3520 Hz. Virgin ears or nyquist for 44.1 kHz only extends to 20000 / 3520 = about 5 or 6 harmonics audible. If the ears have been properly-abused or well-aged to rolloff at 14 kHz, only hear about 4 harmonics. Many well-used ears don't even hear 14 kHz very well or at all.

At high pitches, the perceived difference between a pure sine wave and a complex timbre, is not a lot.

[aciddose]

Nevermind - not helpful.

[fluffy_little_something]

Does this table mean that 12.5kHz is the highest frequency that can be played on Midi equipment such as synths?
http://people.virginia.edu/~pdr4h/pitch-freq.html

[whyterabbyt]

Does this table mean that 12.5kHz is the highest frequency that can be played on Midi equipment such as synths?
http://people.virginia.edu/~pdr4h/pitch-freq.html

No. There's no such fixed relationship between midi notes and the resultant pitch (you know, since you can offset and transpose the pitch of oscillators).

[fluffy_little_something]

I mean, directly. Not via pitch bending

So, without pitch bending or transposing or tuning oscillators up, G9 is the highest note? Just tried it, seems to be correct. G9# doesn't sound anymore.

Interesting, sounds made with a wavetable waveform sound really weird on the highest octave, the pitch gets increasingly arbitrary the higher the note played. With the normal saw, however, the pitch increases as expected until it can longer be heard.

[aciddose]

It's 260 * (2^(1/12))^67, isn't it?

#60 = Middle-C = 261.6 Hz
(Middle-C = base-note * (2^(1/12))^(-9) = 440 * 12throot2 ^ -9)
(Or if you count it as the C above the base note = 523.25 Hz, although I do not think this is the case in MIDI.)

128 note range, 67 above Middle-C (since it starts at 0.)

That works out to 12542.6 Hz.

Yes, note 127 is the highest note. Yet this is at the point where you couldn't hear a single harmonic as the second would be at 25085.2 Hz, significantly above the range of hearing for ANY non-mutant or not-super-powered human.

7f2faab314cfbcbd17114c05231ecf1e.jpg

A typical adult male (40 years) can hear no higher than ~16000 Hz.

[JCJR]

(Edit Note: This message was composed as explanation for Fluffy, not as argument against what aciddose said. )

The note-naming may have standards, but notes are not always standard-named "in the wild". Many sequencers and such let you specify how you want the notes named.

Sometimes C0 might correlate to MIDI Note Number 0. Sometimes C0 might mean something bigger than MIDI 12. When C0 is not MIDI Note Number 0, then very low pitches have to be named in sequencer programs with negative numbers. Many musicians do not like working with negative note numbers. Actually there are numerous musicians who don't even like to see zero in numbering schemes. Just sayin, some people don't like seeing such as C -1 or even C0 in an event list editor or piano roll editor.

Though pitch can be standardized in such as general midi synths, "non standard" synths and samplers have always been arbitrary about pitch correlation with MIDI notes. For instance, if you have a 61 key synth, using general midi standard midi-to-pitch assignments, there are not enough keys on the left-hand side to play a decently low bass line. So often the short keyboards can be shifted down a couple of octaves. Or alternately, the bass patches are designed to naturally play 2 or more octaves lower than the short keyboard will go. A patch playing the C1 pitch when you play C3 on the synth keyboard or whatever.

Maybe about the only "absolute standard"-- Middle C is almost always agreed to be MIDI Note Number 60, though if you play MIDI Note Number 60, there's no guarantee the synth will play Middle C.

Similarly you can make patches which naturally play higher than the expected standard frequencies. For instance you could assign a high piccolo note to the low A on an 88 key sampler. Then that piccolo sample would get pitch-shifted beyond audibility before you reach the middle of the keyboard.

[JCJR]

(continued)

Analog subtractive synths, if built "hi fi" enough, probably won't have much that sounds like aliasing at high pitches, and possibly capable of cleanly playing notes too high to hear.

Maybe well-made virtual analog subtractive synths, especially running at higher samplerates, possibly also capable of going above the range of human hearing without aliasing.

A sampler or wavetable synth, if built hi-fi enough, ought to also be able to transpose a sample beyond audibility without aliasing, but I've heard plenty of sampler/wavetable instruments which are not built good enough to go all the way up without aliasing.

Just saying, some instruments, especially wavetable instruments, might sound "brighter" on high pitches, because the aliasing is making mid-pitch and high-pitch audible frequencies that wouln't be there in a pure-transposed wave. And perhaps the "purest" very high frequency pitches would be from a good quality mono modular analog synth, which wouldn't be as likely to make aliasing, harmonic, and intermodulation distortion way up high, unless you wire in a ring modulator or overdrive the modules.

So maybe it would be common to notice a rompler sounding "brighter" at high pitches, than a good analog synth, because of defects in the rompler. Going the other way, maybe if the patch programmer is trying to hide defects in the rompler, maybe he would program very high pitches "less bright" to hide the machine's defects.

Edit: Also, ears can become nonlinear at high pitches. And many speakers can make distortions if presented with loud high pitches. So when you hear dirt in a mid-range note, or especially very high notes, it isn't a foregone conclusion that the fault is in the synth. The dirt can quite possibly be in the speaker or the ear.

[fluffy_little_something]

I was just wondering if some synths use resonance internally even if the resonance knob is set to zero, like an offset. With some synths I seem to hear that nasal quality already at zero resonance.

[fluffy_little_something]

How can one synth sound clearly more intense and simply louder than another, while the volume meter of the DAW actually shows the opposite? I am not talking about some fancy patches, but even simple ones like init patches. Is there some voodoo at work? Or something more basic like a hidden compressor or similar effect that is not accessible to the user?

Here an example, identical patch and everything, but with volume controls set to deliver exactly the same meter reading in the DAW. Sylenth1 sounds clearly louder and more solid than Hybrid 3 to me.

https://app.box.com/s/vp0mexzc4nkltowqzwxn8g65dj47ivqg

[antto]

the second one is obviously louder, it has a bigger gain, the volume is not equal
the first one is highpass filtered

if you "normalize" both recordings individually - the second one will be louder still
the sawtooth reset edge in the first one is mostly negative (due to the highpass) thus it will peak sooner than the second recording, where the sawtooth's reset edge is equally "balanced" around zero

i wonder, didn't you see that when you recorded the wav?!