Help With Harmor

[tony tony chopper]

If so, would the resynthesis be even better if the Timbre waveforms are set to something other than the saw and square waves they are on by default?

no

[bharris22]

Is every sample you drag and drop for resynthesis automatically combined with the Saw/Square Timbre waveforms?

[tony tony chopper]

yes

[bharris22]

Sorry to be so dense about this - I'm just trying to get a better understanding of how Harmor works. If every sample that is dragged and dropped for resynthesis is automatically combined with the saw/square Timbre waveforms, how does this result in a more accurate resynthesis than if the original resynthesized sample was not combined with any other waveforms? Thanks again!

[tony tony chopper]

look at the mapping envelope that defines the saw, it's all flat, so it's doing nothing to the sample.

[bharris22]

Hi Again,

I have since bought Harmor and learned a little more about additive synthesis in general. To follow up on this last question, if a saw wave has only harmonic partials, how are the non-harmonic partials of a resynthesized sample generated? In other words, I understand from this discussion that the amplitudes in the timbre waveform are multiplied by the amplitudes in the sample to be resynthesized, so frequencies that are not common to both will be ignored. If a saw wave has zero amplitude for non-harmonic partials, how would any non-harmonic partials in the sample to be resythesized make it to the final sound (as they would presumably be multiplied by zero and silent)?

Thanks again for your patience and explanation!

[tony tony chopper]

they are harmonics by order, but made inharmonics by the freq plane, so it's the gain plane that's multiplied with the shape, regardless of the freq plane.

[bharris22]

In other words, it is only the gain levels of the frequencies in the timbre waveforms that is multiplied against the gain level of the corresponding frequencies in the resynthesized imported sample? Are all of the frequencies of the imported sample (both harmonic and inharmonic), then, retained in the resynthesized sample?

The timbre waveforms only affect the harmonic levels of the imported sample, but don't filter out any frequencies in the imported sample - is that correct?

[tony tony chopper]

Forget harmonics vs inharmonics, just think of a bitmap plane. That's what's multiplied, the gain plane. The subfreq plane is affecting the pitch of the partials of the gain plane. It's easier if you think "partials". The resynthesis uses 200 to 500 partials, they're the same ones used by the waveform. Those 500 partials don't have a fixed frequency, they're more 'bins' of frequencies.

[bharris22]

1) So the timbre waveforms only affect the harmonic levels of the partials in the imported sample, but don't filter out any frequencies in the imported sample - is that correct?

2) If so, what happens if the amplitude of a certain partial in the timbre waveform is zero - wouldn't that make the corresponding partial in the resynthesized sample (if there is one) zero and effectively filter it out?

3) Earlier in this thread I think (if I understood correctly), you said that using a saw wave as the timbre waveform does not affect the imported sample for resynthesis, because the harmonic amplitude envelope of the saw is flat. Doesn't the harmonic amplitude of a saw wave decrease as the frequency increases, though? Isn't that why the envelope is flat for the saw (because it's based on Brownian noise)?

4) If so, wouldn't using a saw wave as the timbre waveform decrease the amplitude of the higher frequencies in the resynthesized sample?

Thanks again for all your help. I'm just trying to understand exactly how the timbre waveform affects an imported resynthesized sample.

[tony tony chopper]

1) So the timbre waveforms only affect the harmonic levels of the partials in the imported sample, but don't filter out any frequencies in the imported sample - is that correct?

I don't know if it makes much sense to tell about "filtering frequencies", but no
(the freq plane is also more a freq drift plane)

2) If so, what happens if the amplitude of a certain partial in the timbre waveform is zero - wouldn't that make the corresponding partial in the resynthesized sample (if there is one) zero and effectively filter it out?

yes

3) Earlier in this thread I think (if I understood correctly), you said that using a saw wave as the timbre waveform does not affect the imported sample for resynthesis, because the harmonic amplitude envelope of the saw is flat. Doesn't the harmonic amplitude of a saw wave decrease as the frequency increases, though?

normally yes, but Harmor's "world" is naturally linear for the human brain, that is, flat=brownian noise=sawtooth, even though we could argue that natural for us could be pink noise, that's subjective. Most of the things we hear have an average EQ of pink or brownian noise, way more than white noise. So it's much easier to use musically. It also allows storing stuff in the gain planes more obviously.

4) If so, wouldn't using a saw wave as the timbre waveform decrease the amplitude of the higher frequencies in the resynthesized sample?

no because it's everything that's linear as brownian noise, that includes the gain planes

[bharris22]

This is very helpful - thanks!

2) If so, what happens if the amplitude of a certain partial in the timbre waveform is zero - wouldn't that make the corresponding partial in the resynthesized sample (if there is one) zero and effectively filter it out?

yes

One last (I think ) question about this point. As a saw wave has only harmonic partials in it, don't the nonharmonic partials in the saw wave have an amplitude of zero and therefore filter out nonharmonic partials in a sample imported for resynthesis? If so, how are nonharmonic partials in a resynthesized sample generated in Harmor?

[tony tony chopper]

One last (I think Smile) question about this point. As a saw wave has only harmonic partials in it, don't the nonharmonic partials in the saw wave have an amplitude of zero and therefore filter out nonharmonic partials in a sample imported for resynthesis?

It's always the same question that you ask in different ways, while this is what you have to understand:
take each pixel of a column of the gain plane, from bottom to top, those are the same as each point in the shape's envelope. So what happens is that each column is multiplied by an envelope that defaults to a flat line of value 1, thus isn't changed.

[bharris22]

Thank you.

[Totolitoto]

thank you, I feel so dumb now.