Start with a PC at the high end in terms of storage and processing. You’ll be running several memory-intensive software applications, often at the same time. Here is my configuration*:

1. Sibelius — for music score writing.
2. EastWest Symphonic Choirs — for the human singer soundset and the companion software that comes bundled with it for purposes of creating lyrics through the Votox phonetic language.
3. Studio One — the digital audio workstation (DAW) application that “hosts”, edits, and creates the synthesized sound.
4. Garritan Classic Pipe Organs – used for the soundsets that replicate pipe organs (see step 9 below).
5. Microsoft Word (or any word processing application) — to edit lyrics and convert them to Votox.

Applications like Sibelius are primarily designed for creating scores. Owing to this the score itself sometimes requires adjustment prior to generating MIDI file upon which the choral sound set of EastWest Symphonic Choirs will be applied. Points that may require additional attention include:

• Breath marks
• Slurs
• Tempo change & fermatas
• Dynamics
• Missing/mis-aligned lyrics, especially around melismas
• Pickup notes
• Part names

For the MIDI file all you want are, basically, notes, rests and time signatures. Everything else will be added back later in the DAW (see step 8). For the MIDI export:

• Set export for preferred metronome value.
• Generate one MIDI part at a time.
• Audit (listen to) the MIDI files individually and together with score in hand to assure all notes and rhythms exported correctly.
• If errors are encountered revisit Sibelius, correct the problems, and re-generate the MIDI.

• Import one midi “track” or “line” (e.g., Soprano) at a time.
• Be sure the time signature and metronome values are correct since DAW may not necessarily pick these up by default. Pickup notes require similar adjustments.
• Assure part names are correctly assigned.
• Adjust for eccentricities of EastWest, the DAW, or the MIDI file itself. For example, an issue in EastWest is that (a) the “Soprano” can sing no LOWER than D3 and (b) the “Bass” can sing no HIGHER than D3.

• Choose and load correct voices.
• Set up basic configurations.
• Choose language, phonemes, part names.
• Test the “voices” both solo and tutti on a neutral syllable like “ahhhhh” to see if any errors occurred during the earlier MIDI export/import cycle and to get a sense of the overall balance within the DAW.

This step (6) and the the following one (step 7) are the most critical AND time-consuming of tasks. At a high level the roadmap to convert the choral text from English, Latin, or whatever into Votox* syllables is:

• Export text from Sibelius to cover ALL the words for the piece. This is typically 1 or 2 parts (or portions of several parts) depending upon such things as the mix of homophony/polyphony and which parts are tacet during particular stretches.
• In Microsoft Word edit the exported text and convert the text to Votox for one part.
• Add critical measure numbers from the score.
• Bring converted Votox text into EastWest Symphonic Choirs.
• Test the notes with Votox on one part and adjust as necessary.

*NOTE: Votox was developed specifically for use with the EWSC software. It bears some similarity to – but is NOT based on – IPA (or any other phonetic mappings).

If the Votox conversion in step 6 is successful it can serve as the model for the other parts. Differences is rhythms, melismas, rests etc. will require adjustments across parts, and for that reason it is best to work on one part at a time.

Other “gotcha’s” in Votox are similar to things singers always need to be vigilant about:

• Diphthongs
• Final consonants, including shadow vowels and the nasal (“American”) final “R” sound
• Text attack – normal vs. legato vs. staccato

Much of the success (or otherwise) here is a result of trial and error. For example:

• Problem: “K” or hard “G” sounds are coming across too forcefully on a particular note when all 4 voices are singing.
• (Possible) solution: Have just Alto and Bass sing those consonants while Soprano and Tenor go right to the ensuing vowel sound.

Once the choral lyrics and notes are aligned you move on to converting the MIDI parts to audio parts* (e.g., WAV, .MP3). This is where dynamics, expressions, tempo changes, fermatas etc. return to the piece. Working with both the original Sibelius score and the DAW convert each part from MIDI to audio and then:

• Adjust volume levels (dynamics) first across all parts and then …
• … Between parts in measures.
• Add fade-in (-out) on selected measures.
• Adjust panning.
• Add reverberation – i.e., more echo (“wetter”) or less (“drier”)
• Adjust tempo across all parts.
• Produce final ouptut in .WAV, .MP3 or whatever audio format – this is uaually referred to as a “mixdown”.

*For a good “101-type” discussion of MIDI vs. audio see this article.

Unlike the sounds of singers and their words, Sibelius and others have long been able to replicate instrument soundsets with good quality and relative ease. However, replicating pipe organ sounds requires some special considerations. Again, at a high level here are the steps involved:

• The MIDI output is still generated from Sibelius and imported into DAW but …
• … The soundset is now a product like Garritan rather than EastWest.
• Coming from Sibelius it is easiest to think in terms of the initial MIDI file being Organ 1 (mostly right hand/treble clef), Organ 2 (mostly left hand/bass clef), and Pedal.
• In the DAW each organ stop for each part (Organ 1, Organ 2, Pedal) requries its own track, much like a score for band or orchestra.