 Hey, it's Anfa, and you're watching Anfa Vlog. Today, I'm gonna show you how to make... VACUAL RISE! So in this episode, we're gonna learn how to use Evocoder Antisynthesizer to use your voice to make these cool robotic sounds. Follow me. What's Evocoder anyway? Evocoder is a special processor that takes two inputs. A carrier and the modulator. Then it analyzes the frequency response of the modulator, and it applies filtering to imprint the EQ shape of the modulator onto the carrier, and then it returns that. So if you take a saw wave carrier, which has all the harmonics, odd and even, and then put something like whoa, whoa, whoa, as the modulator, you're gonna hear something that is the saw wave, but doing the oh, whoa, whoa sound. Referring to a previous episode about formants, it's a way to sample formants from one sound and put them on another sound. Usually we sample the formants from our voice and put them onto a synthesizer sound. A quick note. Who made vocoders and why? The vocoder name stands for voice encoder. It was created to enable transmitting encrypted speech recordings way before the digital age, where we could just record a sound file, encrypt the sound file, send it, and we have encrypted sound. So the idea was that you separate the formant information that carries the meaning of the words, we say. You separate that from the tone of your voice, and you encode this in a separate band. So you basically split the sound in multiple frequency ranges, and then you put envelope followers that note the loudness of each band. Then these numbers are written down, they are encrypted, they are sent over, and then anybody turns on a synthesizer and uses filters that recreate the formants based on the numbers received and effectively recreate the speech, of course in a weirdly robotic way, but totally understandable. Enough history. There are two main open source vocoder plugins that I used. First is the TAL vocoder, which is very nice for quick and simple way because it has a built-in synthesizer. So you can quickly dial in a synthesized tone as your carrier and then simply plug in a MIDI controller or a MIDI track to supply the notes for the synthesizer, plug in your voice, and you can go. But there are some other options like MDA talkbox, which is a high quality but very simple in use vocoder. MDA talkbox is a simple but high quality vocoder. These settings teaching the amount of frames. Now it's foo. Now it was low quality. Take a look at the spectrogram. It's talkbox. It's a very interesting vocoder. It means to emulate the sound of physical talkboxes, but that's a different topic. So today we're gonna take the longer but more flexible route of using TAL vocoder, which is my personal favorite vocoder plugin. We're gonna use the newest version that can be downloaded and built from source from GitHub, but also it can be downloaded from TAL get package from Kix to your repositories. If you don't know what these are, there is a previous episode that talks about the Kix to your repositories and you can install these. Okay, let's get making sounds. Here is our Ardor 5 session. I have already created three tracks. The vocoder track, which is a stereo audio bus. That means it will process stereo sound from the input to the output all the time. And we're going to put our vocoder plugin in here. I'm gonna do it right now. If you don't have this plugin in your favorites, you right click here, go to new plugin, plugin manager, and then you type CULF vocoder. You can double click and then hit insert plugin. And this is what CULF vocoder looks like. I'm not going to talk much about what happens right here because it's much more interesting when we have sounds to hear what we're doing. Two notable changes to the stable version of this plugin is that it now allows for sharper Q values for the filter. So you can very have very awesome sharp sounds. And there's also another feature that I ask to be implemented that is manipulating the distribution of the filters. You can rise the lower limit, lower the upper limit, and you can do tilt. That means pushing the more bands toward the highest frequencies or the lower. The cool thing is that you can automate this and get some really cool sounds out of it. Let's close this plugin for now and set up the rest of our tracks. Below the vocoder bus, I have a carrier MIDI track. I call this carrier, this is going to be the source of the synthesizer sound that I'm going to shape with the formants of my speech. I have inserted a Helm synthesizer and it has its default basic patch. I'm just going to disable the second oscillator, so we have just one sort of wave and I'm going to leave it as that. Very simple, nothing special. I'm going to connect my MIDI keyboard and enable the record mode. It works. Now the modulator is going to be my voice. So I have the microphone input one selected already. If I enable now, you probably will hear me double. Now we need to do the routing. So both of these tracks are going to go to the bus. However, the bus is only stereo and we need to put their four inputs because stereo output from the carrier and stereo output from the modulator. Of course, this one is practically mono, but you can record like a choir or several vocals, put them into a bus and then route that bus as the modulator for the vocoder. That doesn't make very much sense because it's only going to smear out your result and make it less obvious to what's happening, but you can do this or you can record your vocal in stereo. So I'm going to disable the output on the bottom of the mixer strip. I click on the master and go disconnect. And here we have the dash and means that nothing is connected to the output of this track. That's modulator, the carrier. I'm going to do the same, disconnect. Now I'm going to the vocoder bus. I'm going for the input, routing grid. I'm going to switch to the order tracks tab. And here is carrier out. Let's put one dot here and one dot here. That connects the carrier out left to the vocoder in left and carrier out right to the vocoder in right. I'm going to close this. The connections are already done. Now if I play the carrier, you can see that the notes are in, but because we have this call vocoder plugin here, nothing will make a sound until we plug in the modulator. So I'm going to right click on this plugin and go for pin connections because we need to enable the sidechain input. Now it's enabled. Now I choose my track. It's the modulator. And now I just root the audio. So now the carrier, that synthesizer sound, is coming in for inputs one and two and the modulator, my voice, is going for inputs three and four. I can close this. Now if I open the call vocoder, we should be able to hear something. It is, however, very quiet because the mic input is very quiet. So I'm going to mute my mic input. I'm going to make this louder and you will hear the vocoding. Okay, so that's a quick demonstration. The problem is I'm getting feedback from the monitor, so I'm going to use headphones for this to get a cleaner sound. Okay, right now I have disabled the monitors and I will disable the voice recording and listening in. Ah, it's still loud. Ah, it's still loud. Is it still loud? Can you hear me now? All right, for that quick demonstration, I think I'm going to change the framing so you can see my keyboard as I play the notes because that's quite important, I think. Yep, having two cameras would be great for this kind of stuff. All right, so now you can see my face and you can see my hands as I play. That's going to make it a little bit simpler, I guess. All right, so let's talk about what the vocoder is doing. Right now, I'm going to pan the vocoder to the right so you can easily hear the difference between the vocoder and my voice. I figured out that actually using vocoder with a whisper is the most effective thing because otherwise the tone of the note that you say or sing like... It clashes with the tone of the carrier synth and it emphasizes some harmonics that you don't really want. But if you just go with a whisper like... Then you're creating a noise-based sound that isn't emphasizing any particular harmonics. It's just allowing the form and the voice to be clearly audible and sampled by the vocoder. So yeah, whispering is much more effective. So this is our basic vocoder sound and now let's try to make it a little bit more interesting. Wow! The first thing is I would use a gate on the modulator so we don't have notes when there shouldn't be no notes because if I press a key... Right, you just have to... gate. If I disable the gate and press a key... You can hear a very silent buzz of the synthesizer but we don't want this to happen unless I really sang something or said something. It's like... Alright, so if I put a gate... I need to of course put down the threshold. And that basically gets rid of that problem. We can also tune the release and attack and the ratio but the default values seem to work pretty well in this case so I'm not going to touch them. Also a good thing is to apply compression before the vocoder. So I'm going to add a compressor to my modulator which is my voice. Wow! You can see we're clipping sometimes into vocoder and that creates an unpleasant sound so I'm going to want to get rid of that with a compressor. This is real weird. We're not clipping on the mic preamp. We're not clipping. We're not clipping on the mic preamp. Alright, so I'm going to just lower the threshold a bit and make sure that we just don't cross a certain level. I think we need to tune the settings of our vocoder. Double-click and open the GUI. Here it is. Ah, yeah, yeah, yeah. Wow! Wow! Alright, so let's have fun and listen to what is happening in this vocoder. I'm actually applying quite a lot of gain to the modulator as you can see which might be causing the clipping. Yep, now we don't have the clipping. Now probably because I'm using high Q we can hear that the filters are ringing. That means they excite so much that they continue to produce a tone when the exciting tone is already dead. Yeah, the good example is when I hit the low C. Now let's change the analyzer to Carrier. And you can see that with that note I have a lot of energy right here. This is 100 hertz. This is probably like 50 or 40. And if I change the modulator, this analyzer mode to modulator we can see the shape of my voice. So that is effectively the sampled sound. Ah, yeah, I can whistle at 1K. But if I just blow into the mic you have this kick drum-like sound and you can see that even on these LEDs that the sound starts on the whole spectrum and then it dies off and the lowest frequency resonates longer. Let's change the tilt and see if that affects it a little bit. Yeah, it definitely rings longer probably because more filters are stacking up here because they are very close. So let's reset this to the default and also make this not high Q and less like that. Ah, I'm going to whisper. It's a very fun feeling when you start using a vocoder like that because you play and you whisper at the same time and you can create very interesting rhythmic patterns and I really like the sound of the skull vocoder when it is at really high Q because it gives me the sound that I really like when people are using stuff like vocodex and fruity loops it gives me this kind of sound when I turn up the resonance. I think it is quite quiet. I think I need more to make it louder. Well that immediately draws Daft Punk to your mind because that's what they are using, a vocoder. Alright, so let's take a look. We can choose the amount of bands we are splitting the signal, the carrier, into. The basic is the eight bands. Ah, yeah, yeah, yeah. It sounds very basic. It sounds very basic. It sounds very basic. It sounds very basic and robotic. The more bands we have the higher the resolution of our vocoding is. Going up to 12 bands. 8 bands. 12 bands. 16 bands. 20 more bands. 32 bands. 32 bands. It's the highest wave solution possible with a conf of coder. It's going to give you the most accurate sound in the form and analysis and audio production so you can make the most awesome sound with your vocoder. Yeah, yeah, yeah. Funny thing is when you start to beatbox through the vocoder well of course I need to turn down the modulator level I'm going to make it all stereo so I'm going to reset the panning and disable my voice. Okay, let's talk about more features of the vocoder. We're at the 32 bands. Now every band has some settings. We can change the level to change the overall balance of the EQ curve we have. I'm going to try to reset that. Oh, it's being carried on. Really nice. Is this a reset button? No, it's not. We need to make it louder. We need to make it louder. It's more or less the same as at the beginning. It's more or less the same as at the beginning. More or less the same. Now I'm pressing two notes that are very close and very low Pretty nice. The cool thing, if you want to accurately represent speech we need to represent the sibilance. And for that we have noise. We can add noise to the bands. Adding noise to the signal. Adding noise to the signal. Effectively adding noise to the signal. We're just changing. Whether it recovers the formance. We're actually making kind of a bypass for the several bands. I'm not sure how to explain it but the more noise, the more regular the sound is. Okay, right now let's play with the synthesizer itself to get this sound more interesting. The first thing I want to do is make sure we start with a clean slate which is the single saw. Then I want to make this more voices and turn up the unison. The problem is, this one is monophonic so I'm going to turn up the voices to 8. I have a problem that the carrier has too much low frequencies sorry the modulator, my voice and that causes bad booms. So I'm going to insert a filter, switch it to 36 decibels per octave high-pass turn it down so it is kind of like 200 Hz boom boom boom boom boom Yeah, much better now. Now we're clipping a bit. I'm going to insert a compressor after the vocoder. boom sake个匆匆 Another thing I want to do is play around with the modulation or with the timbre of the carrier tone of the synth sound itself itself. Let's try it. I want to use some modulation to make the sound sharper, more interesting. Let's use the sine wave and make it very sharp. Make it sharp, make it, make it sharp, make it high, make it pitch, make it high pitch, make it make the high pitch, make it make it high pitch, make it make it high pitch, make it make it high pitch. Give it some motion, give it, give it motion, give it some motion, give it, give it motion, give it some motion. Okay, cut down on the detuning, because it's going to be ruining the sound. Cut down on the detuning, cut down on it. Cut down, cut down, cut down, cut down, cut down, cut down, cut down. Oh, dang, shit, the detuning, yeah. Funny, another thing I want to do is use the trick I showed you in a previous video, with enabling multiple, with enabling multiple instance of helm, so we get a wide stereo. And that sounds much more interesting. Much more interesting. Po, po. Let's turn it down and pitch it down and listen to what happens. Wow. Po, po, po, po, po. Looney tunes. I'm going to enable, add some reverb. And this time let's use Guitarix Zeta Rev reverb, which is one of the nicer sounding reverbs in the open source world out there. So it's GX for Guitarix Zeta Rev one, which is reverb one. And there is no reverb two, so don't be confused. By that. This is the sound of the reverb. Zeta reverb. Oh, yeah. And be confused. By that. This is the sound of the reverb. Zeta reverb. Oh, yeah. The Zeta reverb is pretty interesting. It splits the sound in two bands. You can split it at frequency X. And then you have separate reverb tail times for the lower frequencies and the higher frequencies. And then you have the damping, which is the low pass filter for the upper frequencies. So you can make it very tall. Or you can make it very bright. You can make the low decay quick. Or you can make the low decay long. And the mid decay quick. Get it bright. Very nice. Very nice and interesting sounds. Also it has a two band semi parametric EQ. Because a parametric filter is the one that can select a frequency, the gain, and the width of the bell curve. So, you know, that's a parametric filter. These are semi parametric because you can't change the quality of the filter. So the width of the bell shape. And I'm not going to go into detail because that's not about this plugin. And we have full coding. Really nice. Okay, I think this is it for that tutorial. It was supposed to be quick and it is... I have 15 minutes recorded. Let's see how much will be cut off. You can download this session. The link will be in the description. I can play around with it, have fun, make sure that you have... And also, if you're wondering about how to use the helm synth, which is here, I have already two videos about it, about the older version and the newer version so we can learn it. It's a very nice synthesizer. I really like it. I'm using it more and more. The funny thing is, I started using it more after I recorded the first video. So, I learned myself and that's cool. I hope you're inspired. Get creative. Have fun. So, if you have any questions or suggestions for what I should cover in the next episodes, please leave them in the comments. If you feel thankful you can buy me a coffee by paypal.me, please consider supporting me on Patreon, which could eventually allow me to do this as a full-time job. That would be truly fantastic. I also want to make a shout-out to Harry VanHaren from OpenAV, who is literally my patron from day one. Thank you, anyway. Thanks for watching and I'll see you in the next video. Bye.