Watch Joe Hanley, the creator of Syntorial, as he demonstrates how to take an audio recording of the word “Serum”, import it into Serum, and create a wavetable that says the word “Serum”. In this excerpt from the Serum Lesson Pack, you’ll learn about the different import types, and how to take what you get in an import and clean it up to create a nice wavetable.
Get 4 more Serum videos for free via the Syntorial Demo.
PSSHT. Looking for the forum link? We got you.
In this video I’ll be showing you how to take an audio file in which I say the word “serum,” import it into Serum, and create a wavetable that says the word “serum.” So we’ll learn about the different import types and how to take what you get in an import and clean it up to create a nice wavetable.
Now this video is actually an excerpt from the Serum Lesson Pack for Syntorial. Syntorial is a synthesizer training app that teaches you how to program synth patches by ear. It does this by combing video demonstrations with interactive challenges in which you program patches on a built-in soft synth. The Serum Lesson Pack adds 55 videos that show you how to take everything you learn in Syntorial and apply it to Serum, as well as covering all of the many additional features that Serum has to offer, and you can get the first four videos from the Serum Lesson Pack for free by going to syntorial.com, clicking try for free, and downloading the Syntorial Demo for Mac, PC, or iPad.
This has the first 22 Syntorial lessons but also the first four video from the Serum Lesson Pack. Once you download Syntorial, just go into this drop-down, download lesson packs, and you’ll see the Serum Lesson Pack at the top, just click the download button. We also have lesson packs for Massive, Sylenth, and a couple others. And like I said, the demo will have the first four or five videos from each of these packs. When you buy Syntorial, all of the packs and all of their videos are included. All right, onto the excerpt.
Serum Custom Wavetables Part Four. In this video, I’m gonna show you how to import audio files and create wavetables from them. And it’s really interesting, it’s not a sampler. We’re not just bringing in an audio file and playing it back. Instead what Serum does is it analyzes the audio file, breaks it up into little subtables, and you essentially capture the essence of the sound, not an exact replica of it, and it’s a great source for creating interesting waveforms and wavetables.
So, let’s import an audio file I created in which I say the word “serum.” Here’s what the audio file sounds like. So I’m intentionally over-pronunciating the “S”, the “R”, the “um”, the “M.” All right, so let’s drag it in onto our big wave editor, and we are given several import options. Each of these options analyzes the file in a different way, and thus we get a different wave table result.
These top three are all about pitch. They like to look at the files pitch, analyze its pitch, and break it up into different ways according to that. Now, this audio file, “serum,” me saying a word, it’s not really about pitch, it’s just a word. So these aren’t really appropriate. So we’re gonna use one of the FFTs.
Now these numbers at the end of the FFT determine how big the chunks are that are broken up into subtables. So what FFT does is that rather than look at pitch, it is going to just break up our sound into individual chunks and then analyze the FFT information, so our harmonic series, our phase series, and that’s how it generates the subtables. And the higher the number, the larger chunks of audio are crammed into each subtable.
So for example, let’s try the first, 256. Now right off the bat you’ll see we get a whole set of 256 subtables. And keep in mind, no matter what kind of import you use, its always going to fill up all 256 subtables. The number at the end of our FFT imports, the FFT 256, 512, that’s got nothing to do with the number of subtables. It’ always going to be 256. That number is instead referring to, again, how big of a chunk is going to be crammed into each subtable.
Now I’ve got some silence at the beginning of the audio file, so at the beginning there’s nothing. But I’m gonna scan through using my mod wheel. Now you can hear the word “serum” in there but it’s pretty garbled up, so we’re gonna go over how to smooth it out and make it sound nice in a second, but first I want to show you the difference between the different FFT imports. So our “S” kicks in around 60 subtables, and then… Our “M” kind of rings out until about 180, so we’re looking at about 100-120 subtables total dedicated to the actual word “serum.”
Let’s try this one, 512. Now, going to the beginning. “S” kicks in around 30. And it goes to around 80 or 90, now we’re looking about 50 to 60 subtables, so the first one gave us almost 120, this one gives us half as many, 60. Its cramming more into each subtable. So that’s one of the things you want to think about, you know. How many subtables do I want to work with. How much detail do I want to be able to get in and mess around with.
In this case, I actually want less subtables to work with, because we’re actually gonna reduce them even more in a bit, and I’ll show you why. So I like the lower number for that reason, but you also do end up with different sound options, in this case we end up with different varieties of voices coming from different mouths, if you will, depending on which import we use. Let’s go back to 256. Let’s listen to a couple different pitches, like this C.
Okay, now let’s try this one. I actually kind of like this. I just like the way that sounds, right around there. So you just kind of end up with a different sort of overall tone in our import, and I just prefer this 512. So it’s kind of the perfect sweet spot for me. Okay, so we’ve got to fix up and edit this wavetable now.
And the first thing we want to do is get rid of all the silence in the beginning, so let’s start from the beginning and… Okay, right around there, the “S” kicks in. It’s about 31, maybe… Yeah, about 30. So I’m gonna select the one before it, 29. I’m gonna go into Add/Remove, and Remove: beginning to selected. That’s gonna get rid of all the wavetables up to the one I selected. Okay.
Okay, so now we’re starting at the “S,” I lose the “M” around there. Okay, so now where am I? I’m all the way at 70. So now I’m gonna Remove: selected to end, right. So now it’s gonna remove 70 all the way up. Now what have we got? Okay, good. So I’ve got rid, I’ve trimmed the ends off my wavetable. Now it sounds kind of rocky, right. It’s really jumping between all these different versions of each vowel. Like listen to the “uh.”
Well, maybe all we need to do is smooth out the transition between each subtable, right. Let’s morph it. Let’s go with spectral. Okay, it’s definitely better, but… It didn’t help that much. Like the “uh” for example still kind of , it’s kind of all over the place. So let’s remove the morph, and I want to reduce the number of subtables here. I don’t want, you know, five or six for each letter, 10 for each letter, what I want is just a few at the most. So, let’s go into Add/Remove, Reduce to, I want to keep a quarter of the current tables, so get rid of 3/4. All right, so now… Only 17 total.
Let’s morph that, and now what do get? There we go. Now it sounds like “serum”. And by reducing our subtables we also create more room for interpolated tables, which allows for smoother, more gradual morphing between each subtable.
Next, I want to take a piano sample and import that.
To see the rest of this video, and 54 other Serum videos, you’ll need the Serum Lesson Pack for Syntorial. But again, you can get the first four videos for free if you download the Syntorial Demo here, and then in Syntorial go to download lesson pack, and download the serum lessons.