Dawesome Novum (A Granular Synth for Creative Sampling)

[melomood]

I thought it was more in padshop territory than pigment

[LKBorg]

Alright, the listening and watching stuff got me demoing it. Thx folks, but I couldn’t find a manual on the website (it’s blank for me at Tracktion), there are only 3 small tutorial vids.

So what is the thing with those 6 layers?
I can’t really grasp them. So it’s not that these are the voices of the engine, aren’t they?

[WasteLand]

Alright, the listening and watching stuff got me demoing it. Thx folks, but I couldn’t find a manual on the website (it’s blank for me at Tracktion), there are only 3 small tutorial vids.

So what is the thing with those 6 layers?
I can’t really grasp them. So it’s not that these are the voices of the engine, aren’t they?

those 6 layers, for instance, if you drop a sample on Novum, it will be seperated into 6 spectral layers. the method is explained somewhere in this thread. it is special kind of seperation.

so you have 6 layers, with spectral information. you can change per layer, with the Timbre Flower, the spectral contents, and the amplitude 'envelope'. (spectral breakdown; is frequency + amplitude + phase. frequency you can see (and change with the Timbre Flower)... amplitude you can change... phase, homogenize).

they are played, per layer with a granular processor.

so it has a quite unique approach for treating and playing samples.

it (the layers) aren't the voices. the voices are determined by the granular processor, so a layer can have several voices.

when you drop a sample, after INIT, it plays, when the base note is set right, the sample as is.

afterwards you can change the spectral information, the source. and the granular settings, the play domain...

the spectral content/source is played via a granular engine.

it is a combination.

and there are more ways to change the sound...

[LKBorg]

So far, I really like this. It's simple on the surface, as in: load a patch and you'll be absolutely fine, add some fx, dabble with the settings a bit and it already sounds nice.

A few little things upon venturing in I have a few little critique points:

- the patch selection could be made easier, yes there is the burger menu and you can click there and then further...

- speaking of them, when you select a patch, like one of the ORBITS for example, the beautiful patch selector on the left of the screen stays where it was before, so browsing a new patch folder would be much easier than going back 5 steps or have the patchfield be clicked forward/backwards

- a bit of screen real estate could be won by combining the search and the patch-field, much like the search and url bars in browsers are now combined (and users trained in it, so it's intuitively for them, see attached pic)

- another patch/browsing idea (can't remember if it was already mentioned) it would be awesome to have favorite locations be available to browse directly from the plugin (like: link your sample-folder(s) to it so browsing samples would be a little bit easier). Dunno how complicated that would be?

- and one last thing that doesn't "sound" quite right: I had just a handful of presets that had light crackles in them, much like you'd hear when you recorded a fireplace at a home. Very sublte, very in the background. Like the patch shown in the pic here "ALLELUJA HARM CHOIR" and it reminded me what folks here posted initially in the AD Quantum post about early crackles that were revealed as granular-engine problems later and patched?!

Kudos for the "WE STAND WITH UKRAINE" patch.
It's awesome when folks take a stance!
I think I'll stay with this for a while.

So what I mentioned earlier: these two fields could be combined into one very easily and I think it might enhance the usability, also if this field had direct access to the patch-browser structure (like for example another icon you could click, like a burger menu, but since there already is a burger menu, make it something else, like a book or so)

[LKBorg]

Alright, the listening and watching stuff got me demoing it. Thx folks, but I couldn’t find a manual on the website (it’s blank for me at Tracktion), there are only 3 small tutorial vids.

So what is the thing with those 6 layers?
I can’t really grasp them. So it’s not that these are the voices of the engine, aren’t they?

those 6 layers, for instance, if you drop a sample on Novum, it will be seperated into 6 spectral layers. the method is explained somewhere in this thread. it is special kind of seperation.

so you have 6 layers, with spectral information. you can change per layer, with the Timbre Flower, the spectral contents, and the amplitude 'envelope'. (spectral breakdown; is frequency + amplitude + phase. frequency you can see (and change with the Timbre Flower)... amplitude you can change... phase, homogenize).

they are played, per layer with a granular processor.

so it has a quite unique approach for treating and playing samples.

it (the layers) aren't the voices. the voices are determined by the granular processor, so a layer can have several voices.

when you drop a sample, after INIT, it plays, when the base note is set right, the sample as is.

afterwards you can change the spectral information, the source. and the granular settings, the play domain...

the spectral content/source is played via a granular engine.

it is a combination.

and there are more ways to change the sound...

Also thx

I still can't figure out WHAT the 6 layers are really and how they're chosen — like I threw in a drum sample and it had the drum layers seperated quite cleanly, but a clean bass sample wasn't sorted into like higher and lower spectrum — but I'm having fun so far and that's the main thing.

Oh and on that point: +1 for having a preview function for samples available, so you know what you're mangling there and mbe how to steer a patch into a different direction!

[SLiC]

I wonder if it will ever be possible for the user to have some influence (frequency/dynamics sliders?) on how the layers are separate- that would be next level!

[WasteLand]

I still can't figure out WHAT the 6 layers are really and how they're chosen — like I threw in a drum sample and it had the drum layers seperated quite cleanly, but a clean bass sample wasn't sorted into like higher and lower spectrum — but I'm having fun so far and that's the main thing.

it is not stem seperation, it is another kind of seperation, based on AI, NMF;

https://en.wikipedia.org/wiki/Non-negat ... torization

so, the developer, PeterVor, has learned a 'system', to get spectral elements, 6, from one sample. so of course the source, which can anything, determines the output, not all source will be seperated in the same way... some with enigmatic, but suprising results. or enigmatic...

it does not seperate into high or low, per se, it; well; it is the secret ingredient...

[Sampleconstruct]

I've tried to understand how Novum separates things but have come to no conclusion, e.g. drop a sax phrase into it and in one layer you will just find a single note from that phrase and the same note later in that phrase will be in a different layer together with several totally different notes and some key noises. Not that I really bother but understanding things is fun after all...

[Boy Wonder]

Just so I'm understanding Novum correctly: You can drag in up to six wave files and each wave file gets "spectralized" into six files, so six wave files would yield 36 "spectralized" files?

[WasteLand]

Just so I'm understanding Novum correctly: You can drag in up to six wave files and each wave file gets "spectralized" into six files, so six wave files would yield 36 "spectralized" files?

indeed.

and it is also, indeed, spectral. but how it is learned? the seperation, i repeat, i think PetVor did say something about somewhere, it took a lot of time, that i know, to make it. the process, the learning..

and as @Sampleconstruct says; it is of course done for musical interesting sources, but how it works... well....

but, i don't 'care', found sounds... and you can manipulate them in such a way, that well...

it is an instrument, yes yes yes i know, but it is an instrument...

[Sampleconstruct]

Just so I'm understanding Novum correctly: You can drag in up to six wave files and each wave file gets "spectralized" into six files, so six wave files would yield 36 "spectralized" files?

No, if you drag wav into one of the six slots it will not get separated, so you can stack six different ways.

[WasteLand]

Just so I'm understanding Novum correctly: You can drag in up to six wave files and each wave file gets "spectralized" into six files, so six wave files would yield 36 "spectralized" files?

No, if you drag wav into one of the six slots it will not get separated, so you can stack six different ways.

i read the post wrong..... but each wave file on one slot, will be transformed into a spectral file, indeed, no seperation.. but still spectral so; amplitude + Timbre Flower will work of course.

[Boy Wonder]

My confusion comes from the description of Novum: "You can drag and drop any sample source in the plugin window and Novum will decompose it into 6 layers of sound."

By that statement, to me, it means if 1 sample decomposes into 6 layers then 6 samples should decompose into 36 layers. If it's actually still only 6 decomposed layers regardless of the amount of samples dragged in, I'd understand.

[Dirtgrain]

That does not follow.

[Bako]

autmoation of effect and modulars not possible currently . you can automate everything in Abyss . i suppose it will be implement as Novum is still version 1