KVR Audio

DSmolken · Post by **DSmolken** » Sun Sep 13, 2015 2:50 pm

You can get some amazing "instruments" out of Alter/Ego, too, with the right combination of settings and lyrics - try using "pbpbpbpbpbpbpbpbpbpbpbpb" for the lyrics and slowing down the wave and phoneme rates, for example. Probably not enough instruments for a one-synth challenge, though.

Ah_Dziz · Post by **Ah_Dziz** » Sun Sep 13, 2015 6:07 pm

wagtunes wrote:
Ah_Dziz wrote:
wagtunes wrote:I know I'm gonna get hell for this post but I have to be honest about my feelings on all this Vocaloid stuff and not just this particular product.

It's nice. For anime I guess it can work. But my problem is, I can't sing to save my life and I'd like to put together a demo of one of my songs using artificial intelligence. Problem is, it still sounds robotic. Some of the words I can get but some are totally unintelligible. The vocoder sound is not as prevalent on some words but on others it's very distinguishable.

In short, I can't possibly use this to put together a demo for an industry professional to hear. I'm not sure Vocaloid Diva is that much better but it is an improvement. Some of the demos I've heard I would say are maybe 75% there as far as usability. I'd give this product a 50% at best.

Ironically, the actual vocal quality of the singer on this product I like better than Cyber Diva. But the grammar just isn't there yet.

I don't know if the technology will ever be there, at least not in my life time.

I'm not knocking this product for what it does and I can see some uses for it. But it's not something I can use for what I need to do.
Hire a singer. They are abundant in the world.
Sure, send me the money to pay for him and studio time.

I just assumed you were a professional songwriter since you were talking about submitting songs to industry professionals. That being said vocalists can be hired for very little to record demos. Even with a hobbyist budget you could bring in a vocalist for a full day's work once in a while and knock out vocals for at least ten songs in a day if you're properly prepared. Hell you could use something like this as a demo for your demo vocalist so they could get a better idea of what you want and save even more time.
Either way I think this sounds about as natural as somebody with no singing ability that has been auto tuned into submission. It also has a pretty good range in the actual voice you can have it sing with. The default settings aren't the most natural it's capable of. But all in all it's way easier to use than any other singing plugin I've tried. I had my six year old son making songs with it in about thirty minutes. Thanks for the awesome freebie.

fisherKing · Post by **fisherKing** » Sun Sep 13, 2015 6:50 pm

i had a friend who did a song with a fake sax solo; he spent days getting the sound right, the articulation, bends, etc...all on his keyboard. the artist hated it (it sounded like what it was), and brought in a sax player, who knocked out a solo in 10 minutes.

some things work best when they are what they should be. alter/ego is for fun, for effect. if you need a singer, get a singer.

SampleScience · Post by **SampleScience** » Sun Sep 13, 2015 8:03 pm

I think it sounds really good for fans of anime music and the band "PC Music" (which is quite popular, check their site if you don't know them: https://soundcloud.com/pcmus ). I'd be curious to know more about the synthesis that is used for this plugin. I'm asking because in 1939 Bell's Lab made an instrument called "The Voder" which emulates the sound of the human voice. No samples, it'S 100% synthesis and it sounds quite good for it's time:

That being said, the Voder took at least one year of training to be played correctly. From what I've read, current speech synthesis instruments use more or less the same techniques but they use formant filters instead of band pass filters like the voder used:

For more info: http://www.haskins.yale.edu/featured/he ... voder.html

What Alter/Ego brings that is new to speech synthesis?

davidv@plogue · Post by **davidv@plogue** » Sun Sep 13, 2015 8:22 pm

chipspeech contains many 'singer models', most of which are 100% live synthesis:

Dee Klatt is based on the Klatt algorithm,
VOSIM is based on his namesake algorithm,
Terminal 99 and Lady Parsec (non HD) are LPC10 (TI Speak&Spell core emulator, my research helped improve MAME BTW)
Spencer AL12 is LPC12 (SP0256-AL2 emulation core)
The Voder will be added to chipspeech at one point as another 'synth' singer.

As far as Daisy, its resynthesis of a "Real Voice" with lots of formant and partials logic DSP,
not unlike what UTAU or other modern speech synthesizers do.
However Alter/Ego/chipspeech does that in real-time, without latency, while the others are "offline render" sort of separate timelines.

There are upwards of a hundred methods to generate speech, if you investigate this sort of research seriously.
Each approach having its flaws (read - cool sounds). This is what interests us.

SampleScience · Post by **SampleScience** » Mon Sep 14, 2015 12:14 am

davidv@plogue wrote:chipspeech contains many 'singer models', most of which are 100% live synthesis:

Dee Klatt is based on the Klatt algorithm,
VOSIM is based on his namesake algorithm,
Terminal 99 and Lady Parsec (non HD) are LPC10 (TI Speak&Spell core emulator, my research helped improve MAME BTW)
Spencer AL12 is LPC12 (SP0256-AL2 emulation core)
The Voder will be added to chipspeech at one point as another 'synth' singer.

As far as Daisy, its resynthesis of a "Real Voice" with lots of formant and partials logic DSP,
not unlike what UTAU or other modern speech synthesizers do.
However Alter/Ego/chipspeech does that in real-time, without latency, while the others are "offline render" sort of separate timelines.

There are upwards of a hundred methods to generate speech, if you investigate this sort of research seriously.
Each approach having its flaws (read - cool sounds). This is what interests us.

Thanks! I've done research about the topic, but I obviously didn't go as far as you. You're giving a good hints to go further. I was particularly interested in the Voder when I was searching about speech synthesis, I wanted to create an emulation of it in SE.

I'm sure Alter/Ego is already a success. Cheers!

davidv@plogue · Post by **davidv@plogue** » Mon Sep 14, 2015 1:51 pm

SampleScience wrote:Thanks! I've done research about the topic, but I obviously didn't go as far as you. You're giving a good hints to go further. I was particularly interested in the Voder when I was searching about speech synthesis, I wanted to create an emulation of it in SE.

The interesting bit about the Voder would be to emulate the odd slow articulations that are characteristic of the human controlling the series of keys, the pedal and bar. A vocal tract change can be somewhat precisely modeled by software, but this 'machine' required a human doing 'live automation" of the tract!

Since there are relatively few recordings of the actual thing we would have to guess many such articulations for certain phoneme transitions (We internally use a grid of 36*36=1296 transitions for English).

Fun stuff.

deastman · Post by **deastman** » Mon Sep 14, 2015 2:10 pm

davidv@plogue wrote:The Voder will be added to chipspeech at one point as another 'synth' singer.

Halma · Post by **Halma** » Mon Sep 14, 2015 2:15 pm

A quick "AlterEgo Phrase Vamp". My first attempt into using Alter Ego.

ELEKTRO MUSIK!!!

Dangit I love this synth. Thx Plogue. It took me some time to adopt the phrase playing midi thingy but after some time it became like second nature. Lots of fun.

Cheers
Sebastian

https://soundcloud.com/watch-them-kraut/elektro-musik

wagtunes · Post by **wagtunes** » Mon Sep 14, 2015 2:17 pm

Halma wrote:A quick "AlterEgo Phrase Vamp". My first attempt into using Alter Ego.

ELEKTRO MUSIK!!!

Dangit I love this synth. Thx Plogue. It took me some time to adopt the phrase playing midi thingy but after some time it became like second nature. Lots of fun.

Cheers
Sebastian

https://soundcloud.com/watch-them-kraut/elektro-musik

LOL. Man does that bring me back to the 80s.

Have to admit, that was kind of cool.

CinningBao · Post by **CinningBao** » Mon Sep 14, 2015 2:19 pm

davidv@plogue wrote: The Voder will be added to chipspeech at one point as another 'synth' singer.

When I read this I heard an audience cheer and applause in my mind.
I gave myself the challenge of building my own voder in bidule, and it was bad. Almost playable, but not really. Limited by the mpk49's lack of stops and wrist bar...

Hooray for chipspeech and it's awesome underlying technology!

Hip hip.. <completion not necessary>

Halma · Post by **Halma** » Mon Sep 14, 2015 2:23 pm

wagtunes wrote:
Halma wrote:A quick "AlterEgo Phrase Vamp". My first attempt into using Alter Ego.

ELEKTRO MUSIK!!!

Dangit I love this synth. Thx Plogue. It took me some time to adopt the phrase playing midi thingy but after some time it became like second nature. Lots of fun.

Cheers
Sebastian

https://soundcloud.com/watch-them-kraut/elektro-musik
LOL. Man does that bring me back to the 80s.

Have to admit, that was kind of cool.

Thx. YMMD! I had a rough time but this one gave me a smile.

DSmolken · Post by **DSmolken** » Mon Sep 14, 2015 8:12 pm

Here's something I've been working on... Daisy on lead female vocal and backing male vocal. Backing death metal vocal by the girl whose voice I want to turn into a death metal bank for Alter/Ego. Also a ton of other free plugins and sample libraries used in this.

Iu Gob · Post by **Iu Gob** » Tue Sep 15, 2015 9:12 am

Thank you very much for this free instrument

.
Daisy & Classic Guitar (Guitarra española).
https://soundcloud.com/iu-gob/a-day-a-l ... ra-clasica

kinwie · Post by **kinwie** » Tue Sep 15, 2015 5:36 pm

Hi, sorry for OOT, any news for the multi-out version of sforzando?
If you need a survey, below $100 price is okay to me.

Plogue: Alter/Ego Voice Singing Synth- Free!