Messing with Vocoding

[JSynth]

So, I decided to mess around with vocoding in Logic. This is what I came up with.

https://soundcloud.com/jasper-synth/vocals/s-OhDb5

I could definitely use some feedback on this. Like, how easily can you make out the words if at all?

[Mr. Bigglesworth]

it's kind of difficult to understand at times.

[Navron]

Vocoders have 2 common uses for vocals.

1. A short segment (ex. "We're getting 20% cooler," in Alex S's 20% Cooler Remix) can be used to add a robotic voice prior to a drop or closure of a verse.

2. A continuous vocoder can be used on the main vocals of a song throughout its entirety, but often you will have to layer it with non-vocoded vocals, otherwise they're hard to make out.

Gotta find the right blend.

[the4thImpulse]

Use a harmonic rich carrier to make the vocals 'cleaner'.
Flatten some of the peaks in the frequency response with the EQ bands.
Mess around with the vocoder envelopes, I think this one needs more sustain/release.
Give the carrier some reverb.

[WoodenToaster]

Give the carrier some reverb.

Oh hell naw

Carriers and Modulators should both be completely dry, only add delay and reverb after the vocoding.

[the4thImpulse]

Give the carrier some reverb.

Oh hell naw

Carriers and Modulators should both be completely dry, only add delay and reverb after the vocoding.

I guess our vocoding tastes are different, there are no rules in music so don't make it sound like there is.

[Alycs]

In my messing with Vocoding, I've found that when you mess with the carrier and modulator independently it can start sounding muddy. As the point of Vocoding is to blend the instrument and the voice, it seems like it would make more sense to treat them as one, and add effects after blending them.

What I find works well if you really don't like your voice is to have two carriers, but only use one at a time; that way everything still has that distinctive vocoded sound, but they still sound different on account of the separate instruments modulating them.

[WoodenToaster]

Give the carrier some reverb.

Oh hell naw

Carriers and Modulators should both be completely dry, only add delay and reverb after the vocoding.

I guess our vocoding tastes are different, there are no rules in music so don't make it sound like there is.

But there are guidelines on what usually sounds good and what doesn't, and we're going for vocal clarity in this case. Reverb before the vocoding would be very destructive for that.

[the4thImpulse]

Oh hell naw

Carriers and Modulators should both be completely dry, only add delay and reverb after the vocoding.

I guess our vocoding tastes are different, there are no rules in music so don't make it sound like there is.

But there are guidelines on what usually sounds good and what doesn't, and we're going for vocal clarity in this case. Reverb before the vocoding would be very destructive for that.

I am guessing you have never tried, some reverb setting would ruin it like you say, but if you program it right it will give you added harmonic warmth giving the vocals more clarity (and programming right isn't hard at all). Experimenting is a large part of electronic music, I put reverb on the carrier and I liked it so I will recommend everyone tring it out, regardless of what the guidelines say.

In terms of clarity with vocoding that comes from the harmonic content of the carrier, the more harmonics (combined with the proper vocoder envelope, depth, EQ ect.. settings) the more clarity you can get from the vocals. I agree there should be reverb after the vocoding process like you would with regular vocals, but I like reverb with the carrier as well.

[Raddons]

I don't know what you're using for a carrier, but supersaws tend to give you good results most of the time. Be sure to keep what you're feeding and the carrier mono as well and do you panning afterwards.

[Stars In Autumn]

And randomly, WoodenToaster shows up.
But I've been messing with a song using a vocoder as well and I think it takes a lot of experimentation. One thing I notice is the more bands you add, it can add more clarity. And obviously the carrier is important. The carrier needs to have a signal during the vocals. And mixing in the dry vocals can help too (unless you're really creating something weird).

[JSynth]

Give the carrier some reverb.

Oh hell naw

Carriers and Modulators should both be completely dry, only add delay and reverb after the vocoding.

*dies of a heartattack*

But I've been messing with a song using a vocoder as well and I think it takes a lot of experimentation. One thing I notice is the more bands you add, it can add more clarity. And obviously the carrier is important. The carrier needs to have a signal during the vocals. And mixing in the dry vocals can help too (unless you're really creating something weird).

Now that I am dead, I will look into playing with the bandwidth. I also plan on re-recording the main vocals with better articulation.
Mixing with the dry might be a bit complicated, seeing as I lack tools like melodyne to do a decent job pitch correcting with. But I'll experiment with that as well. This is a pretty new department for me.

[hackd]

Carriers and Modulators should both be completely dry, only add delay and reverb after the vocoding.

...there are guidelines on what usually sounds good and what doesn't, and we're going for vocal clarity in this case. Reverb before the vocoding would be very destructive for that.

... I would take his advice.

But as it has been established, try a stronger carrier. It sounds really "airy" at the moment.

[the4thImpulse]

I don't know what you're using for a carrier, but supersaws tend to give you good results most of the time.

A super saw is great place to start when making a carrier.

anyone know why??

Like I said above its all in the harmonic content, super saws have a lot of harmonic content so they often work well as a starting point. Sine waves have no harmonic content and usually sound horrible as a vocoder carrier (unless its that particular sound you want).



Here's an example of what I did with a vocoder in 20 minutes:

https://soundcloud.com/the4thimpulse/vocoding-example

The first part is completely dry, the second has a reverb on the carrier before the vocoder, the third has reverb after the vocoder. I made small changes to the reverb so it would sound decent on the two examples, every other parameter stays the same. The synth is a very simple saw wave played in a chord.

[WoodenToaster]

Fished out this old thing I wrote up

http://wooden-toaster.tumblr.com/post/3 ... torials-on

[the4thImpulse]

Fished out this old thing I wrote up

All your carriers should be monophonic, if you want more harmonies add more vocoders

Carriers should be completely dry, no reverb and no delay, those should be added onto the finished product

Compress the shit out of the vocoder channels

Vocoding is mono, so pan out harmonies and add stereo imaging to some layers to give it some clarity and life.

Sorry to hear you live such a boring life with all your rules here, music would be very stagnant without experimentation and rule breaking.

[WoodenToaster]

Fished out this old thing I wrote up

All your carriers should be monophonic, if you want more harmonies add more vocoders

Carriers should be completely dry, no reverb and no delay, those should be added onto the finished product

Compress the shit out of the vocoder channels

Vocoding is mono, so pan out harmonies and add stereo imaging to some layers to give it some clarity and life.

Sorry to hear you live such a boring life with all your rules here, music would be very stagnant without experimentation and rule breaking.

Sorry to hear your vocoding sounds like ass.

[froggy]

My sides are moving on their own

[JSynth]

Fished out this old thing I wrote up

All your carriers should be monophonic, if you want more harmonies add more vocoders

Carriers should be completely dry, no reverb and no delay, those should be added onto the finished product

Compress the shit out of the vocoder channels

Vocoding is mono, so pan out harmonies and add stereo imaging to some layers to give it some clarity and life.

Sorry to hear you live such a boring life with all your rules here, music would be very stagnant without experimentation and rule breaking.

Well, yes that is true, but for someone who is just starting to vocode a bit, I am going to want to know what the general "guidelines" are. I can always experiment and rule break later.

[the4thImpulse]

Well, yes that is true, but for someone who is just starting to vocode a bit, I am going to want to know what the general "guidelines" are. I can always experiment and rule break later.

His 'guidelines' will give you great results, he is completely right to do it the way hes suggested. What I don't like is when he says its the only way, there are no rules. Try it his way and listen, then try it the exact opposite of everything hes said and listen again. Turn knobs for the fun of turning knobs and discovering its effect on the sound.

What vocoder do you use in logic?

[Navron]

Sorry 4th, but unless you can outdo Glaze in the vocal department, I think he has the best say in the area.

From a logical perspective, it doesn't make sense to add reverb prior to the vocoding, in the same sense that you wouldn't add compression before de-essing.

Reverb is in its true essence, is just a very short delay, with varying amounts of feedback. If you were to insert reverb prior to a vocoder, you'd have the vocoder synthesizing not only the sound of the vocals, but the large amounts of tiny delay as well, which in the end, would only serve to make the vocoded sound more muddy, or at the least, harder to differentiate the words.

I can see some possibilities in your technique in creating some interesting sounds (especially with longer reverb times), but if your goal is a synthesized voice with more presense and warmth, adding reverb to the final product is definitely the better way to go.

[the4thImpulse]

Sorry for voicing my opinion everyone, I can see no one wants me here. All I did was encourage the OP to experiment and gave him my personal workflow for vocoding, I never said anyone's method was wrong or inferior to my own.

Navron, I am compelled to add to your definition of reverb. Reverb is a simulation of a room, it's much more than just one delay and even a short one at that, it factors in material and shape just as much as size. It's far more complex than a couple of delays with EQs which results in an effect that can be used in many ways beyond simulating a real world room.

[JSynth]

Well, yes that is true, but for someone who is just starting to vocode a bit, I am going to want to know what the general "guidelines" are. I can always experiment and rule break later.

His 'guidelines' will give you great results, he is completely right to do it the way hes suggested. What I don't like is when he says its the only way, there are no rules. Try it his way and listen, then try it the exact opposite of everything hes said and listen again. Turn knobs for the fun of turning knobs and discovering its effect on the sound.

What vocoder do you use in logic?

Well, I do plan on experimenting with different nobs too see what happens. That's how I came up with what I posted earlier.

Now to get back on topic, I am using the EVOC 20 PolySynth.

[Navron]

Sorry for voicing my opinion everyone, I can see no one wants me here. All I did was encourage the OP to experiment and gave him my personal workflow for vocoding, I never said anyone's method was wrong or inferior to my own.

Navron, I am compelled to add to your definition of reverb. Reverb is a simulation of a room, it's much more than just one delay and even a short one at that, it factors in material and shape just as much as size. It's far more complex than a couple of delays with EQs which results in an effect that can be used in many ways beyond simulating a real world room.

Should have worded the definition a little better. Yes, it is much more than a single delay. An algorithmic processing of multiple delays to simulate the shapes and sizes of rooms.

There are no rules in music, but there are guidelines, and the guidelines when it comes to any type of sound usually states that any time based effects should be used close to the end of the signal route.

We're not saying you're wrong, or saying that the OP has to do it this particular way. It's simply saying, "This is the standard."

If somebody was asking for help on making dubstep, and somebody advised that they make dubstep at 80-90bpm with the snare on the 2nd and 4th beat, you can bet there will be at least 3-4 people who come in to say, "You want 140bpm and the snare on the 3rd beat," because that's the standard.

H8_Seed was recently talking to me how people should never use compression on vocals, yet (in EDM especially) there are very logically sound reasons why compression on vocals is important, and often used.

You need to stop getting defensive and bent out of shape when others prefer that people are being taught the standard techniques. A musician doesn't need to be told to follow the rules. That's their decision. To assume you have to tell them that they don't have to listen to the advice is (quite frankly) insulting to them, because that's making it seem like they don't have the capability to make that kind of decision on their own. I'm sure JSynth has the capability to experiment with different signal flows and values himself. So for future reference, if you're going to suggest something that goes against the norm, don't get defensive when others correct you on the standard.

Nobody is trying to be mean here, but you're really, really, really taking this the wrong way.

[FIMprov]

I found out while experimenting that a saw wave and a square wave put together have really big harmonic content. Experiment with the relative levels of the two oscillators until your spectrum analyzer is pretty much flat--about 60% / 40% (or, in a dB ratio, 3:2 to 5:3) in favor of the square wave seems to be about right--and a spectrum analyzer is the easiest way to check if the harmonics of either oscillator are overpowering the other's. Now, I'm pretty new to vocoding myself, so I don't really have much to show for using this particular kind of carrier, but I thought that this information might be helpful. The square wave has some different harmonics than the saw wave, and more harmonics equals more clarity, after all.