Make Noise - New Spectraphon Module

[Stefken]

Well, it s a pretty tedious process to map what s going on in a video (with talking on top) to a spectrogram, but i gave it a little try with the panic girl video .

For SAM mode.

With the slide knob, you can see that the harmonic distribution slides along the horizontal axis of the spectrum, when you turn the knob.
With a low setting, you see a lot of upper harmonics, and as you set slide to a higher setting, these distributions get mapped to lower frequencies, giving it a more bassheavy tone. Off course, the source material plays an important role too. It would be interesting to test with e.g. material with a lot of mid frequencies (say, a voice ? ) and then see that midfrequency distribution slide and get mapped to the lower harmonics.


With the focus knob. I took a printscreen with the slide knob at a high setting and then

1) focus knob on low.

low focus.png

2) focus knob on high

high focus.png

On low focus, you have a fuller, denser spectrum with somewhat apparent 'formant mountains'.
On high focus, you get a much thinner spectrum with highly discrete err... focused spikes of sine waves. The overall amplitude of the spikes is also lower.

* This exercise was not conducted under scientifically proven circumstances
* No animals were hurt during the course of this exercise

See it more as an appetizer to spark your interest.

[Stefken]

Some nice sounds in here

[Stefken]

so, where are those Make Noise groupies when you need them?

[elxsound]

Some nice sounds in here

That was very cool. At times it was hard to hear spectraphon, but subtle performance use cases are nice too.

[Stefken]

Some nice sounds in here

That was very cool. At times it was hard to hear spectraphon, but subtle performance use cases are nice too.

I wonder what spectraphon will do with ambient material and pads.
From the little we ve seen it will most likely create a layer of dirt/harshness that you can blend in with the original as it s character seems rather harsh.

Very different from panharmonium. A pity Spectraphon won t be doing any ambient duties.

[foosnark]

I wonder what spectraphon will do with ambient material and pads.
From the little we ve seen it will most likely create a layer of dirt/harshness that you can blend in with the original as it s character seems rather harsh.

Very different from panharmonium. A pity Spectraphon won t be doing any ambient duties.

The harshness and dirt I've noticed all comes from strong upper harmonics and growly FM. Dialing in Shift and Partials should keep it in check pretty easily. One of the patch ideas I want to try is using it to create a sort of washy harmonic reverb using a lot of Focus (maybe ducking Focus a little with the envelope follower, so it doesn't blur too much except during the tail).

But then, I always felt that Panharmonium always wanted to sound either like chiptunes or bad MP3s

[Stefken]

In 5 hours

[foosnark]

Hopefully that comes with the "now shipping" announcement

[elxsound]

Hopefully that comes with the "now shipping" announcement

Yup, it’s announced on the YouTube channel and IG.

I’m also waiting on Ctrl-Mod… hopefully they get their stock soon

[elxsound]

RTFM time

https://www.makenoisemusic.com/content/ ... manual.pdf

[foosnark]

Focus in SAM mode is doing more than I expected. I figured it was just a kind of slew for the amplitudes, but it seems to be a bandwidth control for the partials. Or maybe I was wrong about the filter speculation and it really is using FFT, or wavelets or some other thing. It doesn't really matter I guess

Something Walker mentioned in the chat that I didn't see in the manual: the arrays are stored on a memory card as small data files that you can save on a computer. Which implies to me that one could also generate new arrays using some other method and transfer them I imagine we will have our hands full just using it normally for a good while though!

[elxsound]

Absolutely, hands will be full. I’m really interested in the array creation walkthrough he mentioned, but I have a feeling that might not show up for a couple weeks (or longer). It would be helpful to know the thought process behind keeping the SD card in the back of the module.

And not that I’m knowledgeable about these things, but in the video demos they kept saying FFT, and just as you added that Focus relates to frequency bandwidth, reducing partials to the fundamental. I just forget which direction narrows and which direction expands.

[Stefken]

Something Walker mentioned in the chat that I didn't see in the manual: the arrays are stored on a memory card as small data files that you can save on a computer. Which implies to me that one could also generate new arrays using some other method and transfer them I imagine we will have our hands full just using it normally for a good while though!

Yep, i don t have the skills to program an editor but i will be looking into those files and maybe i can edit them by hand. But i would be awesome if you could do things like in Iris 2 like cutting out stuff from the spectrogram. And some way to generate them would be awesome too. At only 1024 frames there is a limit to consider though.

[Stefken]

they kept saying FFT, and just as you added that Focus relates to frequency bandwidth, reducing partials to the fundamental. I just forget which direction narrows and which direction expands.

I reckon this being a fully digital module with a good processor, they actually do FFT.
Setting the focus knob to high, strenghtens the focus, narrowing the spectrum to isolated spikes.
Just have a look at the spectra i posted on the previous page:
viewtopic.php?t=596530&start=304

[foosnark]

My understanding of FFT is partial (heh) but I know it has an inherent latency, regardless of processing power.

The resynthesis method it uses only cares about actual harmonics rather than arbitrary inharmonic partials or phase. So I speculated, maybe the mention of the Buchla filterbank as inspiration was somewhat more literal than people have been assuming, and the implementation is using a parallel set of bandpass filters with envelope followers. With enough processing power, you can run a lot of them at low latency.

OTOH, Tom Erbe is a hell of a lot more knowledgeable at this stuff than me, so it could be some kind of optimized hybrid FFT thing or wavelets or something else scary from the math department.