Working on some demos of the Atomic Pack sounds.
Toybox release Atomic Pack
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
Here's a deep dive of the Atomic Oscillator included in the pack by CinningBao:
https://www.youtube.com/watch?v=J7HslqKPqLE&t=489s
https://www.youtube.com/watch?v=J7HslqKPqLE&t=489s
-
- KVRist
- 102 posts since 2 Jan, 2019 from London, England
Watching the deep dive video convinced me of two things.
One, cpu for (what are likely to be) common settings is high. I'm assuming he has a pretty good machine (would be great to know the specs so I can compare to mine), and it's showing between 25%-30% – with what looks like around 2/3 max partials – and 40% (max partials). FM on/off didn't seem to make a significant difference. Polyphonic sequences probably would, but my use case for this would most likely be leads, so monophonic with maybe occasional note overlap.
Two, I really want that oscillator. I like the concept of additive synthesis, but was, until now, never fully sold (Pigments came close).
One, cpu for (what are likely to be) common settings is high. I'm assuming he has a pretty good machine (would be great to know the specs so I can compare to mine), and it's showing between 25%-30% – with what looks like around 2/3 max partials – and 40% (max partials). FM on/off didn't seem to make a significant difference. Polyphonic sequences probably would, but my use case for this would most likely be leads, so monophonic with maybe occasional note overlap.
Two, I really want that oscillator. I like the concept of additive synthesis, but was, until now, never fully sold (Pigments came close).
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
To give you an idea, It's using around 7% min partials, 15% medium partials and around 22% max partials on my 1019 Macbook Pro.
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
-
- KVRist
- 102 posts since 2 Jan, 2019 from London, England
Could not resist. Using a prebuilt rack which includes the Atomic Oscillator, it's possible to overload my current machine, but I have to push the boundaries of what it can do. I found that the number of "max" partials can also be edited, default (for the rack I loaded) is 24, but it can go up to 72 (which is what is used in the video). Set to 24, with the knob on "max", I get slightly higher CPU already than in the video. Pushing the "max" up to 72 gets me high enough to start glitching. It sounds good with 24 max partials, so I'm okay with that for now, and can push to between 36 and 48 and still have a reasonably stable session.
Current timeline to new computer is March/April 2023. Until then, if I have to freeze tracks...
Questions, from a novice user of additive synthesis:
- How many partials do people commonly use?
- Does number of partials only affect the harmonic richness of the sound? If I understand the correctly, I can create a saw wave with a fundamental of 100Hz, and harmonics up to 1500Hz using 16 partials (root at 100Hz plus one additional at every integer multiple, i.e., 200Hz, 300Hz, ..., 1500Hz). A square wave would be similar, but only using odd-integer multiples, so to get to 1500Hz I'd need fewer partials.
- Why stop at 72 max? I think Pigments goes to 256.
And something that I don't get about how the oscillator works. In the section where you can set partials/base safe/unison, there are level controls for oscillator 1 and 2. These don't, as far as I can tell, do anything. Set to full or zero, I'm not hearing an obvious difference. What am I missing?
Current timeline to new computer is March/April 2023. Until then, if I have to freeze tracks...
Questions, from a novice user of additive synthesis:
- How many partials do people commonly use?
- Does number of partials only affect the harmonic richness of the sound? If I understand the correctly, I can create a saw wave with a fundamental of 100Hz, and harmonics up to 1500Hz using 16 partials (root at 100Hz plus one additional at every integer multiple, i.e., 200Hz, 300Hz, ..., 1500Hz). A square wave would be similar, but only using odd-integer multiples, so to get to 1500Hz I'd need fewer partials.
- Why stop at 72 max? I think Pigments goes to 256.
And something that I don't get about how the oscillator works. In the section where you can set partials/base safe/unison, there are level controls for oscillator 1 and 2. These don't, as far as I can tell, do anything. Set to full or zero, I'm not hearing an obvious difference. What am I missing?
-
- KVRAF
- 2475 posts since 15 Apr, 2004 from Capital City, UK
Hi fold24,fold4 wrote: ↑Fri Dec 09, 2022 2:50 pm Could not resist. Using a prebuilt rack which includes the Atomic Oscillator, it's possible to overload my current machine, but I have to push the boundaries of what it can do. I found that the number of "max" partials can also be edited, default (for the rack I loaded) is 24, but it can go up to 72 (which is what is used in the video). Set to 24, with the knob on "max", I get slightly higher CPU already than in the video. Pushing the "max" up to 72 gets me high enough to start glitching. It sounds good with 24 max partials, so I'm okay with that for now, and can push to between 36 and 48 and still have a reasonably stable session.
Current timeline to new computer is March/April 2023. Until then, if I have to freeze tracks...
Questions, from a novice user of additive synthesis:
- How many partials do people commonly use?
- Does number of partials only affect the harmonic richness of the sound? If I understand the correctly, I can create a saw wave with a fundamental of 100Hz, and harmonics up to 1500Hz using 16 partials (root at 100Hz plus one additional at every integer multiple, i.e., 200Hz, 300Hz, ..., 1500Hz). A square wave would be similar, but only using odd-integer multiples, so to get to 1500Hz I'd need fewer partials.
- Why stop at 72 max? I think Pigments goes to 256.
And something that I don't get about how the oscillator works. In the section where you can set partials/base safe/unison, there are level controls for oscillator 1 and 2. These don't, as far as I can tell, do anything. Set to full or zero, I'm not hearing an obvious difference. What am I missing?
Regarding performance, you could try increasing your buffer size to give the computer a touch more time to think about the sound. If you're doing patches which kind of run themselves, like drum machines or sequencers, you can probably increase the buffer size a lot since you wouldn't looking for instant key-down feedback.
Theoretically, you can keep adding partials for most complex waveforms and after <a quality value only determined by the user has been met> there will be a point when you can't hear the difference between one more or one less. In some situations you might decide that you need as many as possible because you want all the high-end partials there to brighten the sound. In another situation you could be making a sound you are quite happy with but then realise taking 10 of the partials from the top end won't qualitatively remove anything from the sound, giving you back a few CPU cycles Additive synthesis isn't always about perfectly reproducing another sound or waveform (not to say the right additive tools can't be used for that); it's about using the tools the additive oscillator gives you to make the sound _your_ sound. Drawing your own partials, describing the partial harmonic relationships, changing the partials from sine to another shape..
- How many partials do people commonly use?
---- entirely depends what you want to achieve. sometimes I'm completely content with 6, other times I want hundreds.
- Does number of partials only affect the harmonic richness of the sound? If I understand the correctly, I can create a saw wave with a fundamental of 100Hz, and harmonics up to 1500Hz using 16 partials (root at 100Hz plus one additional at every integer multiple, i.e., 200Hz, 300Hz, ..., 1500Hz). A square wave would be similar, but only using odd-integer multiples, so to get to 1500Hz I'd need fewer partials.
------ I kind of answered this in the 2nd paragraph above. how many partials do you need to make you happy in the moment of making the sound? There aren't any rules about this; for example. if you want to put the sound through a saturation/distortion, 16 partials might be more than enough for you to get the sound out of the following stages. It depends what those following stages are doing. I'm not going to say "the rules of music say you need a minimum of x partials for your sound to be considered an actual sound" because those rules don't exist. It's probably the most boring answer of all, but ultimately only you know how many partials you need
- Why stop at 72 max? I think Pigments goes to 256.
---- All the devs have different coding frameworks and within those frameworks, limitations. Many limitations exist because of system resources. It _may_ be possible to increase the partial count in future but for the moment 72 (which is in fact 144 because there are 2 oscillators) works for us. And since the Atomic oscillator can also FM each of those partials against another, and those partials can also be any waveform from the morphing waveform block, we thought the sonic possibilities were broad enough!
And your final question about the oscillator volumes:
--- There are two pairs of outputs on the Atomic oscillator; one pair labeled L and R, which outputs a mix of the oscillators, with the partials spread around the stereo field. The volume knobs affect the oscillator volumes at this output. The other output labeled 1 and 2 are direct feeds from the oscillators and are not affected by the oscillator volume setting; it's full volume the whole time.
I hope some of that helps
-
- KVRist
- 102 posts since 2 Jan, 2019 from London, England
I'm at max buffer (2048 samples). I usually program my sequences rather than play them (my keyboard playing ability is poor in both timing and in keeping up with playing what I hear in my head, at best it's useful for coming up with ideas I then have to quantise, and any evolutions of those ideas will be programmed by hand) so that's not a problem, but I can't take it higher. I am running at 48KHz so I could drop to 44, but I doubt that will make much difference. I have an old machine (2015 Macbook Pro) that isn't performing as well as it used to. An upgrade is very much due.
In case that wasn't clear: my aging computer is the problem here, not the Atomic Pack.
The rack I'm using doesn't have the L/R out attached to anything, so now I know why those controls do nothing.
Thanks for the response, that was all very helpful.
In case that wasn't clear: my aging computer is the problem here, not the Atomic Pack.
I completely agree with this. My question was more curiosity than anything else. Playing with the oscillator, I found some very nice sounds I got with around 36 partials, some with more, and then on one sound decided I preferred it with about 14. Adding more created interesting things going on higher up, but I felt they distracted from the core sound, and were something that on a subtractive synth I'd find myself filtering out. So I'm fully convinced I don't need a lot of partials. The interplay of FM and other ways you can manipulate partials are the main reasons I wanted this pack.CinningBao wrote: ↑Fri Dec 09, 2022 4:25 pm Additive synthesis isn't always about perfectly reproducing another sound or waveform (not to say the right additive tools can't be used for that); it's about using the tools the additive oscillator gives you to make the sound _your_ sound. Drawing your own partials, describing the partial harmonic relationships, changing the partials from sine to another shape..
I realised after I posted that in fact there are 144 across the two oscillators. The rest makes sense as well.- Why stop at 72 max? I think Pigments goes to 256.
---- All the devs have different coding frameworks and within those frameworks, limitations. Many limitations exist because of system resources. It _may_ be possible to increase the partial count in future but for the moment 72 (which is in fact 144 because there are 2 oscillators) works for us. And since the Atomic oscillator can also FM each of those partials against another, and those partials can also be any waveform from the morphing waveform block, we thought the sonic possibilities were broad enough!
The rack I'm using doesn't have the L/R out attached to anything, so now I know why those controls do nothing.
Thanks for the response, that was all very helpful.
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
Yes exactly, due to the fact that each partial of the two additive oscillators can be 'FM-ed', 'ring-modulated' or 'sample & held' by the corresponding partial from the opposite oscillator, very rich and detailed timbres can be produced with just a few partials.fold4 wrote: ↑Fri Dec 09, 2022 5:20 pm I completely agree with this. My question was more curiosity than anything else. Playing with the oscillator, I found some very nice sounds I got with around 36 partials, some with more, and then on one sound decided I preferred it with about 14. Adding more created interesting things going on higher up, but I felt they distracted from the core sound, and were something that on a subtractive synth I'd find myself filtering out. So I'm fully convinced I don't need a lot of partials. The interplay of FM and other ways you can manipulate partials are the main reasons I wanted this pack.
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
We just made a new cheat sheet for the Atomic Oscillator from the pack.
The Atomic Oscillator is a super versatile block that fuses advanced additive synthesis with FM, great for thick, harmonically rich timbres. Each partial of the two additive oscillators contained in the block can be 'FM-ed', 'ring-modulated' or 'sample & held' by the corresponding partial from the opposite oscillator. Therefore very complex and detailed timbres can be produced with just a few partials. Check out the cheat sheet here:
https://www.toyboxaudio.com/pages/atomic-oscillator
The Atomic Oscillator is a super versatile block that fuses advanced additive synthesis with FM, great for thick, harmonically rich timbres. Each partial of the two additive oscillators contained in the block can be 'FM-ed', 'ring-modulated' or 'sample & held' by the corresponding partial from the opposite oscillator. Therefore very complex and detailed timbres can be produced with just a few partials. Check out the cheat sheet here:
https://www.toyboxaudio.com/pages/atomic-oscillator
-
- KVRAF
- 3027 posts since 6 Nov, 2006
Any hope for pdf versions of the various cheat codes?
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
Here's a pdf of the Atomic Oscillator cheat sheet -->
https://www.dropbox.com/s/xq4kv205w4fiu ... t.pdf?dl=0
-
- KVRAF
- 2991 posts since 22 Dec, 2004
Besides the Atomic Oscillator, are there any other new or revised modules in this pack?
- KVRist
- Topic Starter
- 352 posts since 15 Apr, 2019
Yes, the blocks are either brand new blocks, or updated and improved versions of older blocks with new DSP etc.Genetic_Junk wrote: ↑Mon Dec 12, 2022 11:27 pm Besides the Atomic Oscillator, are there any other new or revised modules in this pack?
-
- KVRian
- 756 posts since 7 Dec, 2018 from Saint-Petersburg
Damn. Can’t wait NI actually do Silicon native support for Reaktor. Your take on Reaktor modules is something truly unique and so versatile. Almost real modular experience without spending thousands $ on modules.