VOCODER

[Last Alternative]

I'm looking for tips to get a clean sound for vocoder vocals. Some of you have heard my song God Forsaken Land. I am revisiting it and fine tuning the demo into a finished song. Just search soundcloud for Last Alternative. I'd post a link but I don't want a huge picture attached. I'm sort of shooting for a Pendulum or Party Monster vibe. Here are some examples below. I know Rob Swire uses a talkbox and of course you have to abnormally enunciate when singing but any other tips, articles, or videos would help. I've been researching and haven't found much.

My main problem is getting solidness and clarity. I've messed with the vocoder knobs, its eq, and used Selig's leveler and de-esser- on the main vocals as well as the vocoder track. I've tried everything but it's a hard beast to tame. Thanx in advance.
http://youtu.be/B-l6N710u78
http://youtu.be/zENtEumtyBI

[kloeckno]

The carrier tone is very important. It needs to have as many harmonics present as possible since the vocoder subtracts away many and mostly leaves the harmonics present in the modulator (voice).

Filtering the carrier is not a good idea. Add in some white noise also to make "s" "t" and similar sounds, AKA fricatives. I've also found that distorting the vocoder output helps sometimes, and also makes it sound mean.

[joeyluck]

I've always gotten the best results from my Warpfactory and Talkbox.  But Reason has some more tricks up it's sleeve as of recent worth taking a look at (that I haven't had time to play with yet).

Parsec has a vocoder (which many users seem to prefer to the BV512).
And then there's Selig's DeEsser (which you could use on the formant) to bring out the syllables more (using the sibilance output to add separately to the mix).

Also, many folks tend to go for saw leads for the carrier, but I also find that things like a very clean solo trumpet or trombone work well as well.

This song (now 10 years old...geez time flies) was done using a trumpet on a DJX and the Warpfactory vocoder.

[Benedict]

If you are after intelligibility then mix Vocoder and Clean vocal so the vocode is the "backing". You don't need much to make a voice sound artificial as we are really keyed to human voices.

[JNeffLind]

Holy shit! A Reason related technique thread I can actually contribute to in a meaningful way.

Catching my breath here...

First, a note on recording. If you want your vocals as intelligible as possible, don't sing and then vocode. Speak instead. Speak in a clear tone but with expression, like you're reciting a monologue in tempo with the beat, and overacting. It takes a bit of experimentation to figure out how to fine tune the line between enunciating super-clearly to be understandable and over enunciating to the point where the vocoded signal doesn't sound like singing. But speaking (as opposed to singing) and then vocoding is a definite must for a finished product you can understand.

(note: this means you won't be able to blend the signals of dry and wet, only use pure vocoded signal. To do this just go from Insert FX "to device" of audio track to modulation input on vocoder, and from carrier output on vocoder to Insert FX "from device" on audio track to not have to hear your original spoken vocals. Obviously the carrier input on vocoder needs to be chosen now. See below. If you just want a bit of vocoded flavor then you can sing and have the vocoding behind it with the "clean" singing taking care of intelligibility, or do it how I described and record a second singing track. If you want pure robotic voice though, you must speak before vocoding!!!)



So get your audio part recorded and your midi info input to pitch the words, then put it on loop start scrolling through the patches for your carrier input until you find one that sounds good. The best I found in the factory sound bank was "Accorditor" which is subtractor patch. I had best luck with FFT set to 512 on vocoder and HF turned all the way up, dry/wet all the way up, attack all the way down, and decay at about 2.5. Shift is a matter of taste I think. Anyways, pick a patch that sounds good (or decent) to start and then start tweaking it.

Not sure how much of that you already knew, but figured I'd give my two cents for those diving into the vocoding pool for the first time.

[joeyluck]

First, a note on recording. If you want your vocals as intelligible as possible, don't sing and then vocode. Speak instead. Speak in a clear tone but with expression, like you're reciting a monologue in tempo with the beat, and overacting. It takes a bit of experimentation to figure out how to fine tune the line between enunciating super-clearly to be understandable and over enunciating to the point where the vocoded signal doesn't sound like singing. But speaking (as opposed to singing) and then vocoding is a definite must for a finished product you can understand.

Good point. I always figure that this is common knowledge/practice when vocoding. However, whenever I have had friends over to play on the vocoder, their natural instinct is to sing. It's all about enunciation.

[EnochLight]

Another vote for Parsec as a vocoder. I can't even touch BV512 since using Parsec. That said - and this is unrelated - but is that Macaulay Culkin in the first video??!

 

[JNeffLind]

Another vote for Parsec as a vocoder. I can't even touch BV512 since using Parsec. That said - and this is unrelated - but is that Macaulay Culkin in the first video??!

Strangely enough, I never much liked using Parsec's vocoder for vocals (as opposed to drums or whatever else), but I did like using some Parsec sounds as carrier input on the BV-512. Just seemed like I couldn't get a pronounced enough effect with the Parsec vocoder on its own. I'm sure it's a matter of personal taste though. For me, when I'm vocoding, it's go big or go home I guess. Maybe I didn't spend enough time with it, but I liked the freedom of being able to run any synth into the bv 512 and having thousands of patches at my disposal to try and tweak.

[Raveshaper]

This thread is giving me another strange idea. It's time to patch.

[MirEko]

I've posted this before but just to show a different angle I made this combi with redrock REs.



But yeah +1 for Parsec.

Also combine more than one method, for example use a bv512 vocoder and duplicate the vocal and run it through a Neptune with max pitch correction and use the midi track to that's triggering the synths and copy the midi notes to trigger the notes on the Neptune to match the bv512. Mix the two tracks together for a much clearer and legible vocoder sound

[Last Alternative]

Yeah that is Macaulay Culkin. Party Monster is an awesome movie. Thanx for the tips you guys. I'll give it a whirl.

[kloeckno]

Check out Zapp and Roger for some more awesome talkbox inspiration!
http://youtu.be/tiuUcUEk3Ck

[Emian]

about Parsec as  vocoder; i messed around with its voice input funcion a bit the other day, but i didn't achieve any results i liked.... i do like Parsecs synths sounds tough...
 
are there any Vocoder presets for it; & if not how come so ?
Any tips for setting up Parsec as a vocoder from any more experienced Parsec users ?
 
Thx in advance !
 
Jan

[Territan]

You want a modulation (instrument part) loaded with harmonics, so a sawtooth gives you the most bang for your buck.

As for the carrier (voice part)? Try turning up the gain and whispering. The airy part of a whisper is loaded with white noise, and will let most of the modulated signal through.

[JNeffLind]

Here's another tip I was reminded of:

If you end up setting up your routing correctly and then just letting it run and auditioning patches, there are some patches that vary drastically depending on what octave they're pitched in. I had some sort of "gangsta" leads (reminiscent of Death Row Records leads from the nineties) that only sounded good a few octaves higher than I'd been pitching everything else. If you find a patch with a good thick sound that you think would be good for vocoding but it doesn't sound good try making a two octave shift in your midi pitching. I just use the tool window to make this easy. Anyway, happy vocoding.

[Last Alternative]

I'm really happy with the sound I have now. It's very clear but I still need to redo my vocals with more enunciation and it will be killer. Appreciate all the tips! This song rocks- can't wait to share it.

[challism]

Predator also has a vocoder.  I can't compare it to Parsec because I haven't used Parsec yet. 

Can anyone here give a good comparison of the different vocoders available in Reason?