SOUND ON SOUND

[Kwackman]

In last months "Nugen Audio: How A Plug-in Is Made" article, there's a box explaining "Additive Synthesis & Resynthesis".

There's a bit that I don't understand. 3rd paragraph.

"Why go to all this trouble, rather than simply sample the clarinet? Because a resynthesized sound is infinitely mutable, in a way that a sample is not."

What does "infinitely mutable" mean?
Thanks.

[desmond]

It means you can always change it - all the constituent parts are always editable, unlike a sample, where the timbre changes and so on are baked into the sample.

"Mutable" is a concept used a lot in programming to, at least on the Mac, where you can have for example "strings" that you can't change after creation and "mutable strings" where you can change them as much as you want.

[Kwackman]

Cheers Desmond. :thumbup:

[CS70]

In last months "Nugen Audio: How A Plug-in Is Made" article, there's a box explaining "Additive Synthesis & Resynthesis".

There's a bit that I don't understand. 3rd paragraph.

"Why go to all this trouble, rather than simply sample the clarinet? Because a resynthesized sound is infinitely mutable, in a way that a sample is not."

What does "infinitely mutable" mean?
Thanks.

Yeah, something that you can change as you like.

If one wanted to be pedant, also an instrument sample is "infinitely mutable", in the sense that it can be changed whenever you want. After all, an "instrument sample" is just a short digital recording made by PCM samples (ie. say 24 bit words) and therefore you can apply FFT transforms like with any other sound. You do it every day when you mix.

But what the guy means, surely, is that it's far easier to change a synthetized sound towards a specific goal that doing the same with an instrument sample, because you can reason in terms of the specific components that make up the sound in a way that often is less easy and efficient with a sequence of PCM samples. And the whole sound generation and playback mechanism is radically different, of course.

[desmond]

The difference here is that you can manipulate a sample, but you can't change it's component parts. You can eat the cake, slice it up, chuck on some ketchup, whatever, but it's still the cake at heart.

Resynthesizing *breaks the sample into it's component parts*, which can then be modified - the cake is split into it's components, the egg, flour etc, which can then be changed, swapped out, reassembled in different quantities, add in some ketchup flavouring etc.

[CS70]

The component parts of an instrument sample are the PCM samples that make it up - or even better, their bits. You change them, voila'. The two systems are turing-equivalent - which is why you can do both inside a computer. But of course for certain tasks it's easier to work with one or the other.

[Dr Huge Longjohns]

It also means that, if it's a sax sample for example, you can turn it off for as long as you like.

[CS70]

What about ukulele samples? :D

[desmond]

The component parts of an instrument sample are the PCM samples that make it up

For the example in the article - you sample a clarinet - that *is* a PCM sample. it doesn't have component parts as such (although you can derive some from FFT but you're now on the way to resynthesis...)

For example, think about pitching a sample up - with a sample, you transpose everything, with a resynthesized voice, you can transpose the pitch fundamentals up but keep the timbre consistent.

Anyway, I thought the example point was well made, and I'm less interested in getting into a pedant's definition-war, so I'll bow out there... :thumbup:

[Kwackman]

Anyway, I thought the example point was well made,

Your (initial) answer was all a layman like myself who knows nothing of programming needed. Thanks for the clarity and brevity!!

[CS70]

Sure - I did not see it as a particularly controversial remark. But just for clarity: an instrument sample it's not *one* PCM sample - it's a whole lot of them, as an instrument sample usually lasts a little more than 1/44100 of a second (or whatever sample rate). :-) One PCM sample = one word (these days, often a 16 or 24 bit word). Maybe you meant PCM stream. If someone's learning to code plugins, the distinction is relevant.

[desmond]

Ok, true, I see what you meant, thanks.

In the context of the example though, a "sample" didn't mean "one sample point", it referred to a single digital recording of a sound, so that was how the wording was intended.

[Eddy Deegan]

What about ukulele samples? :D

They are infinitely mute-able.

[blinddrew]

::: stops unpacking new ukulele :::

::: puts it sadly back in its bag :::

[desmond]

They are infinitely mute-able.

These things just remind me what a good hang this place is...

:clap: