Mike is an L.A. based composer/producer who has written music for the series Sleeper Cell, Bill Nye The Science Guy, Oprah, Maximum Exposure, in addition to many songs and jingles. Starting in 2007, he began a long journey to fill a need he had for a vocal library that was geared to serve a different purpose than others in the marketplace. The result is the Kontakt based library named “Realivox-The Ladies” from his company, Realitone.
Hi Mike, good to see you, thanks for doing this.
You’re quite welcome. Thanks for coming by.
It seems that with many of the great new libraries that are coming out, they are being created by guys who never started out to be developers, but composers, like yourself. What made you decide to take this on?
Originally, I just did it as a lark. Believe it or not, I thought it would be fun to have a NAMM booth and pretend like I was real developer. In 2007, I wanted to create a simple vocal library just to do background “oos” and “ahs” for my own use. I used to write a lot of songs and singers would come in and sing lead and harmonies. I’m not a guy who records a ton of tracks and then chooses the keepers later. I like to work fast and make most of my decisions as we record, so I would often decide, “Nah, we don’t need oos in the chorus.” The singer would leave and then I would realize “Oops, I should have had her sing oos in the chorus!” That happened a lot.
So I got the idea to take one girl and one guy and record them singing oos and ahs chromatically up the keyboard so that when I got in a bind, instead of having to either settle for not having them for those sections or having to call the singers back in to record those parts, I could just play them in from the keyboard. Really simple to do with Logic’s EXS sampler.
Then I discovered the Kontakt scripting forum on V.I. Control. I was a math major in college and I had taken computer programming classes so I understood the concept of how this worked. As I was reading this I realized that I could do the same kind of legato that VSL instruments would do. I could write the code myself. At that time, there weren’t vocal libraries that could do this with legato in a way that impressed me so I did some simple tests of myself singing one octave of oo’s with sampled legato transitions and sure enough, it worked. This was not a script that anyone was going to marvel at, but it was a starting point and proof to myself that I could make it work.
So I decided to do this for real and brought in Julie Griffin, a singer I’ve worked with a lot. Gotta be like 20 years now that I’ve known her.
I just love her timbre, her sound.
Yeah, she’s an amazing singer. She had the hardest job timbre-wise because she’s not singing full voice, but she’s not singing completely softly. I wanted that Brasil ‘66 kind of sound and it’s hard for any singer to sing that way consistently. Anyway, she was the guinea pig as I figured out how to do this library and I must have had her in here for probably 30 sessions while we experimented with various techniques. Toni, who was the last female singer I brought in, was only here for about 5 sessions because by then, I had figured out what really needed to be done, how many legato intervals needed to be recorded, the most streamlined way to do it and what to listen for during the recording process.
The obvious way to record someone is to start on the lowest note and go up chromatically. But what I do is start in the middle to establish the basic tone and then record up, and then go back to that middle note and record down. This way I get a more consistent tone for the meat of the range, which is really important to make the legatos smooth.
That makes perfect sense to me. As a singer myself as I go up the scale I know my voice gets edgier.
Right. Cheryl sings up to this ridiculously high C. Her tone overall is soft and sweet but even she can’t sing that really high C ultra-softly.
And since this is “Realivox” (Mike laughs) you wouldn’t want that because real singer’s voices can’t do that.
And with each of the singers, it was their “money” tone I was looking for. If I hire Patty to sing a song, it’s because I want “that” sound. For Cheryl, I want her angelic sound.
Another challenging thing for the singers and for me is keeping the energy level up, because it is a really boring process recording a sample library, but if the singer loses energy and loses that smile on their face, you actually hear it in the vocal recording.
Were other developers helpful with advice and questions?
Definitely. Hans Adamson (Art Vista,) Andrew Keresztes (Audiobro,) Dan Powell (SonicCouture,) Greg Schaeffler (Orange Tree) and many others gave me valuable tips and encouragement. It’s a very supportive community, at least with the guys I’ve dealt with.
You and I talked about a long time ago how while there were a number of choral libraries aimed at the big “epic’ sound, with Latin phrases, nothing wrong with that, there is clearly a market for it, but there were not the kind of background vocals that are good for pop records, comedies, etc. I particularly have wanted that Beach Boys-Chicago kind of sound…
Which you will get!
So what so impressed me when I started to play with Realivox a little yesterday was that I took a couple of different singers and started to work, and the keyswitching is so consistent and just so intuitive that after only an hour, I felt I could write and arrange effectively for this library. Not as well as a month from now perhaps, when I know it really well, but effectively.
That’s good to hear, because I’ve done a lot of tweaking on the keyswitch programming since I first started five years ago. That’s one nice thing about showing Realivox at the NAMM show for this last few years is that people would give me suggestions right and left, like “it would be really nice if it could do this” or “How come you have the keyswitches set up like this, why don’t you set them up like this instead?”
So I took a lot of those suggestions and that’s how it evolved to the 30 articulations with the keyswitches set up the way it is now.
Some people don’t like keyswitches but what yours allow the user to do is use them in an improvisational way rather than thinking and create things in real time you might not otherwise come up with. You play “ooh ah, ooh ah, shooby do bow” on the spur of the moment and you think, “Oh man, that really works!”
Now, you know that there is a ShooBeDoo page to automate that, right?
Yes, I discovered it and I love it, as well as the drops. It would be fun to try a Swingle Singers kind of thing or like Burt Bacharach’s vocal cue from “Butch Cassidy and The Sundance Kid” once you have released the males as well.
Yes! Once again, that was one of the great things about showing at NAMM. I didn’t even know about the Swingle Sisters until someone there told me I should do their songs using Realivox.
OK now I know that one thing you are concerned about is that this library will be typecast as just doing cute pop stuff and comedic cues and that you feel it is far more versatile than that, and I agree. Please talk a little about some other ways you feel this library can and will be used.
In fact, I didn’t intend for this to be a pop library. While it started with the oohs and ahs, which is straight up pop or rock or R&B, once I introduced the legato I wasn’t thinking pop at all any longer.
I did a theme song for TV a show called “Sleeper Cell” about terrorists cell in the United States. It is kind of a Middle Eastern sounding theme with an Arabic scale for the intro and then goes into this world kind of vocal chant and I hired a singer named Patty Mattson to sing it. About a year later I started fooling around with Realivox and I realized I could recreate what Patty sang on this show’s theme song. It’s in one of the demos you can hear on the website and it sounds almost dead on to what she actually sang on the show. And it was really easy to do. Click on the reverb and delay, start playing an Arabic scale and bam, instant film cue.
Or with Cheryl, you have this very soft, angelic voice that is usable for dreamy ethereal stuff. Even with Toni or Julie, their voices definitely could be used for film score-ish things, soft background or even lead parts. Teresa is obviously a soprano diva singer, but beyond classical, you can even use her to do the original Star Trek theme.
What it comes down to is that the singers don’t really have “styles.” It’s the notes you have them sing that makes the style. Patty can sound like she’s singing prayers in a mosque, or jazz, or techno, or African chants or whatever. The melodies are what determine the style.
One thing I found very interesting is the Voice settings you have, 12 levels of Bright and Dark. My first thought was that they were EQ settings but that is not true, as I read in the manual. They essentially give you a tool for having more singers, correct?
Correct, it is NOT EQ and it has turned out to be surprisingly useful. If I take Toni, for example, and I play a C, then that’s just Toni where she sang a C note. No surprise there. But if I choose “Bright 1,” now when we play a C, we hear where she sang a B pitch, but shifted up a half step. It’s a C, but with a different tone, since it’s really a pitch-shifted B.
Because her vocal placement was different.
Right, and her performance as well. So she’s going to have a little bit brighter timbre. So in the set where I steal notes a half step down and pitch them up a half step, I call that “Bright 1.” I can do this all the way to “Bright 12” and “Dark 12,” which is up an octave and down an octave. What this accomplishes is it can give them a little younger or older sound or even special effects voices. Several of the demos on the website make use of this and it opens up even more possibilities.
The other thing this allows me to do is to double, triple, or even quadruple stack vocals. I can start with Toni “Normal,” then add Toni “Bright 1” on the second track, and Toni “Dark 1” on the third track. That way I have, in effect, many more singers with different tones. I can even create a choir, as Craig Sharmat did in a demo on the website
Well, Mike, in my opinion you have developed a very unique library that fills a gap in the marketplace.
Let’s hope it stays that way. (He laughs)
After the males are released, could you envision yourself doing another library or was this a one shot deal?.
Actually, I enjoy this. I am not saying every minute of it is a joy, but this is a part time thing for me, so it is a diversion from writing songs, doing scores, etc. While I don’t get frustrated writing music generally, I do sometimes get frustrated with clients. The nice thing about this is that the client is me. Obviously, there are customers to consider, but ultimately, I am the client. A sample library can never be perfect so it is left up to me to decide when the library is “perfect” enough.
And then you have to assume if it pleases you, it will please others as well.
Yes. And I’ve been really lucky that while there have been a couple of minor technical issues that users have pointed out to me and that I fixed, I haven’t had any real complaints from those who have bought the library. That is why I took so much time to actually release it. I kept fine-tuning and tweaking it, so that no one would complain about bad notes, bad loops, bad legato, etc.
I have to tell you, I did a quick experiment with Realivox and put it through my UAD Lexicon 224 plug-in and it was gorgeous, just magical.
It has that sheen for the vocals that you expect from a Lexicon. The sound of your singers through that is killer!
Oh, great, now you have me wanting to buy a UAD card! (Both laugh).
One final point about why this needed to be a keyswitch oriented library rather than loading separate articulation on different tracks. When a singer goes from an ooh to an ah or ee, the singer’s mouth will make a W sound before it switches.
(Jay sings) “Oo wah, oo wah, oo wah, oo wah, Why Do Fools Fall In Love”.
(Laughing) Right. I recorded a wonderful bass singer named Juan and I thought we were done, I wanted him to play him singing the “oh-wee-yo” theme that the Wicked Witch’s guards sing in “The Wizard of Oz.” We had the ohs and the ees, but when I played it with the keyswitches, it didn’t sound right because there was no W between oh and ee and you need a Y between ee and oh. So it came out “oh-ee-oh” instead of “oh-wee-yoh” and it just didn’t sound right. It was such a disappointment because a) it didn’t sound the way I wanted it to and b) I now had to record a whole bunch of new legato samples! (Both laugh)
Thanks Mike, this was fun.