Some Music using Voices  created with Myriad Software's "Virtual Singer" 

1.  O Magnum Mysterium  3,883 k.
Gabrieli's  music for two choirs.  The lower choir is changed to Brass players with two voices echoing the top line.  This was partly to avoid many computed voice parts, ( each human voice part has 3 voices, 2 virtual and one "choir aah ".  Some parts had an additional virtual voice)   and partly so I had the excuse to use brass with Gabrieli .

 2. Exultate Justi.    2,196 k.
 Arranged for  choir and brass .. somewhat slower tempo than commonly used.
( information. for myriad users >> the voices are a real singer 'david' for tenor, a 'davidia' for the women, and for bass, the italian singer used in Da Pacem.)
Voices are doubled with a soundfont choir.   The brass in this older version was programmed on Roland D20.   
Comment.. this is so much easier than singing the piece in real life where I kept running out of breath, even when used to singing in a choir.( tenor)  
 

 3.  Amor Vittorioso.  1,046 k. 
 I heard this on the internet sung by the Kings Singers. I liked it, but in my memory it turned into a more solid arrangement, so made this into a reality.   Uses my latest custom brass from Korg, with one additional trombone soundfont.  Timpani from Myriad HA.
( the voices are doubled . A 'david' for tenor plus  a da pacem tenor,  for the women, a 'davidia'  with a 'roberta'.   For bass, the myriad italian singer used in Da Pacem and an 'american bass' . )

4. Kyrie  for two male choirs.  958 k.
  A curious event here.  A kyrie was originally composed in short steps while exploring the possibilities of virtual voices without vibrato.  Later one accidently re-played two versions of the tune at once, but  out of sync.
 Was astonished by the rich, sonourous effect, so worked at  setting them together in an interesting manner.  And here is the result with cathedral  impulse reverberation added .  Voices are  'davids' with HA's  'ragazzo' no-vibrato singers.   Also a soundfont voice made by this person singing 'eee'.
To seriously compose using this double choir effect will take considerable skill and effort.  Perhaps a musician reader  can sense the possibilities.  

5. That old feeling   1,133 k.
Barbershop with four 'daves' of various vocal range, and some added glissando effect using  my  HA script.

6. An extract from Dixit Dominus  1,534 k. 
  A test of Virtual Singer 'real' voices ( mostly based on 'davids') in a baroque work.

NOTE:   The next items are for solo singers. A challenge.  I'm not sure it's a good idea to try and make them sound like real physical singers when they obviously aren't !    Perhaps the hidden expectation makes them seem less satisfactory by comparison?    On the other hand,  if  a listener can hear them on their own terms they may appreciate the effort that has gone into them.

7. The lost chord.   3,871 k.  
  A solo  virtual 'David' baritone.    Some years ago this was orchestrated  for a xmas CD to inflict on the relatives.   At the time I used a software vocoder on a sound font  mens choir, and then added my own real voice singing many times over to make the words clearer.   Here, I have added the virtual voice over the old orchestral audio based on soundfonts and Roland brass etc.    Ideally it should be re-done to make the virtual voice words clearer and have higher quality strings and brass. .. etc.. etc.... but  ...mmm..  Sometime perhaps.    Has potential.

8. Nessun Dorma   2,991 k.  
Well, you have to at least attempt this, don't you !  
 I hope a listener is intrigued and amused as I am by the use of several real audience applause segments blended together and overlayed.  It  adds a 'reality' at the start I think, but I wonder again, does it just make the virtual singer  less real by contrast?
I also put some general ambient audience noise behind the music.   Only to lose most of it in avoiding a sneeze which recurred throughout!

   The voice here is yet another David  ( ie.. it was 'derived' from my voice).. the same one as just above..  but there is a different setting of the voice formants (think of them as equalizer settings) and a few other things.   Perhaps a newcomer can sense how the same data is altered by these 'formants'.  
I have to say that after you have listened to one of the well known real performers, this voice seems biased into a light tone and high note range ,  not  really .. hmm ...  filled with the  passion and physical effort of the blood?
   I have made another attempt since writing that, and the voice has now more vowel  response set in it, and a slight 6 Khz boost added to the file, and is perhaps little better.  The voice is still rather high, but stronger I think.

 It would be interesting to see if the extravagance of some opera styles could eventually be achieved .  
  One can for example alter some of the voice parameters to make it more throaty and forceful. The problem is that my own source voice is light in tone.  Despite the alterations one can make, even reaching down to a bass range, I have the opinion now that it requires an original voice that has some of the desired final qualities.

 9. An animated virtual singer  2,144 kb (2.1megs)  using  H264 mp4 codec

       animated virtual singer    Same file as above using MPEG-4  (mp4v) codec  

Some time ago I got the idea to make an animated singer that operates within Harmony Assistant.
This is possible because it has a built in scripting language (Lua)  which can access much of the program's features and also draw graphics to some extent.
This short video of an Italian  "relaxed popular style " tune, uses only the inbuilt GMSE sounds and FX, so do not expect audio wonders!  None the less, with reverb on it sounds quite reasonable.
The voice is a  virtual David ( my voice) converted to sing Italian. The sequence was provided by Luigi ( musicaresco) on the Myiad forum from an italian source.
The animation figure is roughly modified from  the excellent "Delay Lama" freeware.   

It took this person, who is a slow learner at a new computer language, ages just to get the animation to this stage. And then it seems that it is critical to work through the source sequence/score and make sure that the vowels and words fit correctly on the notes.  This is because  the only way I found to make it all work was to pre-analyze the singers staff on the score before it starts to play. One has to write into an array the start times and end times of sung notes, and the vowel required. And this is NOT always as straightforward as one would like!!  This particular sequence took some work before the thing would even speak correctly.

It is not the only way to make an animation. One can instead, use the current volume of the staff and use it to open and shut a  mouth and move eyebrows etc. This is MUCH simpler to do, but it does not look as interesting.  
 To go any further one would have to master the art of animation drawing and not borrow someone else's work!   I am not sure I have the time left, or the graphic programs to do this, so for now, it all rests.
 

10. She walks in beauty like the night  4,290 kb ( 4.3 meg)

I was struck by the words and feel of this tune, written by an anonymous musician calling herself ( or possibly himself )  'Esmeralda Weatherwax'  ( obviously reads Terry Pratchett)  on the Myriad music forum.
The words written by Byron, are famous. The tune is not!  ' Esmeralda'  seems to have vanished, but has said she /he/they does not mind if anyone uses the music.
Originally it was somewhat 'folk' style, but I have made it  a more somber orchestral version using a fair part of the accompaniment. Changing a fair bit also.  I hope to make  a video of it, but it may be hard to find a woman with suitably innocent and striking beauty and then make beautiful graphic effects to go with it.

 

 

 Link back to Main Music Index