2

Can someone please explain to me the difference between the sound of one violin and the sound of an ensemble of violins? Take for example an ensemble of only violins playing a Concert C. Why is it that they not only sound louder, but slightly more like a choir rather than just one violin playing a Concert C. The reason for my question is that I am trying to synthesize the sound of a violin ensemble but it does not sound like many violins, rather than like an electric pad. It would be helpful if you had any knowledge you would be willing to share regarding different timbres in the same instrument.

My theory from what I know about music (I am a clarinet player myself): Each violin has a slightly different timbre because no violin is exactly identical, therefore causing an ensemble of violins to sound like more than one violin.

Here is the sound I'm going for (skip to the violins), and here is the sound that synthesizer software makes. You can quickly tell there is an important difference in the sound, I just need help pinpointing what exactly it is.

Thank you very much!

  • As for getting that "choir"-like sound from a synth, this is what chorus pedals were invented to do. It (essentially) creates multiple copies of the input sound with slight delays added to get the different copies of the sound out of phase with each other. – S. Burt Sep 6 '16 at 4:16
5

There are at least two effects at work. They're related and connected, yet I'd keep them apart:

Transient smearing

Bowed strings have naturally a pretty bright, edgy sound – in synthesizer terminology, it's actually similar (albeit more complex) to a sawtooth wave. These hard edges in the signal sound kind of raspy. (Particularly pronounced when beginning notes with a hard attack at the frog.)

However, they only happen at discrete points in time. If multiple violins play together, then the transients will never be really in sync, even if all violins are playing exactly in tune. The result of this is a waveform that has the same spectral composition, yet sounds much smoother than a single violin, because the transients don't add up and are “lost in the crowd”.

Note that this effect doesn't require there to actually be multiple violins at all: it can also be achieved if you only record a single violin with a close mic multiple times, and layer the recordings (doing that gives a result quite similar to 70's “string synths”: thick and smooth, yet somewhat unnatural and not really like a true string section).

Spatial widening

If you hear a single violin, your ears can pretty well focus on it (the sharp transients help with this, but they aren't required). Your ear zoom in, if you will. This can give the kind of intimate closeness that's needed for a folksy fiddle tune, but it also brutally highlights any playing details that aren't so desirable.

If multiple violins play physically together, they will necessarily sound all a bit different: different instruments have different timbre, the players will be a few cents off in pitch, and they can't sit all on the same chair. Hence there is no single point in space to focus on: you get a wide ensemble effect. Reverb also factors into this.

I have in the past tried to also emulate this effect with single instruments. It's quite tricky; you can get an approximation when actually setting up mics in a large room as if recording an orchestra, but then skip from chair to chair for each individual voice. Ideally also switch between a couple of different instruments, but this is actually not that crucial.

http://dime-chime.bandcamp.com/track/remember-the-sun

You definitely need a bit of stereo spread to pull this off convincingly. Which is what's most lacking in the recording you linked to (that actually sounds a lot like good old mellotron strings).

4

Your hypothesis that the violins don't sound exactly alike because they are not perfectly identical is indeed part of the reason why a violin section has a chorus effect, rather than sound like a louder violin.

Here is the list of things that would cause a section of violins to have the chorus effect compared to an individual violinist:

  • No two violins are identical and have slightly different acoustic signatures (mainly in the overtones)
  • Players are ever so slightly out of tune with one another. Hopefully this is minimized, but there will always be some slight pitch discrepancies.
  • Similarly to the previous reason, players won't always attack at the exact same time. This is sometimes desired, if you do want a fuzzy attack.
  • Most violin work uses vibrato, unless specified otherwise. More so than any of the above reasons, the differing vibrato speeds and intensities account for the difference between the sound of a single violin and a whole section of violins.
  • 2
    And the phase mismatch, as ttw pointed out. BTW, not only are the violinists out of phase as they produce the sound (at the string), but they all are different distances from the listener! – Carl Witthoft Sep 6 '16 at 11:14
  • Thank you for the breakdown, it's good to know this in musical terms as you put it. – IgnisImperatrix Sep 9 '16 at 18:05
3

Each violin in a section will be out-of-phase with the others. The peaks and troughs of the various sound waves will not occur at the same time. Thus N violins sound about Sqrt(N) times louder than one violin. Were the violins in phase, the sound would be N times louder.

The practical effect is that one needs a section to sound like a section. The string sections of an orchestra do not sound like a loud string quartette. For making electronic sounds, just being louder does not a section make.

There may also be differences in tuning that enhance this effect.

  • Wow, the mathematical way of putting it helped me understand a little better, thank you! – IgnisImperatrix Sep 9 '16 at 18:06
  • You are describing the increase in volume but not the other aspects of the question. It's not just volume that the OP is asking about. Is he/she wrong about the "other" effects? Is an increased volume the only effect? – ggcg Nov 22 '18 at 13:23
  • My point (though not fully explained) is that one gets less volume because of the phase difference between instruments in addition to the slight difference in tuning It's a different tone quality. I'm not sure my Sqrt(N) is correct on re-consideration (I guess I should look it up). It may be that pure in-phase tones would give an N^2 loudness rather than N. The ratio is correct but the formula may be wrong. – ttw Nov 22 '18 at 14:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.