Why produce our own rehearsal tracks?
After all, there are many resources for choirs on the Internet. Why not just use one of them?
- Other than the simple, unsophisticated examples, they are a significant expense to the user
- Users are limited to the repertoire, and choice of instrument for each voice that is provided by the supplier
- The resulting rehearsal tracks cannot take into account variations in performance details to suit the particular choir.
By producing our own rehearsal tracks, we can overcome these issues, and also tailor the output to suit the choir. One example might be by isolating particularly difficult sections of a score and looping through them, or slowing the tempo. The approach set out below has been to produce each track giving as near as possible to a finished sound, and also with particular voices highlighted (but with the other parts at lower volume to provide cues and harmonies).
The files on this page are examples of the various techniques that have been tried or can be used for the SACS rehearsal tracks, together with comments on how they were produced and the relative difficulty of each. In the main, the mechanisms to produce the outputs are from open source applications, professionally programmed by enthusiasts but effectively free of charge. The only “expense” is the time it takes to create them.
First Steps
Before starting to produced the sound files for rehearsal tracks, the score has to be turned into the correct format. If we are lucky with the choice of music, the score may be available on the Internet in a “MIDI” (Musical Instrument Digital Interface) format that can be imported directly into a scorewriter application (the one used currently is MuseScore). However, if the music is only available as a printed score, first it has to be scanned with an application that can translate it into a MIDI format. The app currently being used for scanning is Neuratron PhotoScore – the output quality of which can be very high, but depends heavily on the clarity of the original printed score.
Once in MuseScore, the various staves are edited to correct transcription errors, add the instruments (including voices), add dynamics, add local performance variations as specified by the choir director and generally to tidy everything up. From this point, edited MIDI files or MusicXML files are produced for use in other players that choir members may have, but by far the main product is MP3 files that can be played on a computer, phone, tablet or sound system without needing any extra applications.
Producing the Basic MP3 Rehearsal Tracks
To date, all rehearsal tracks have been produced directly from MuseScore. The instruments assigned to each of the staves govern the sound, and for simplicity, have been taken from the set provided with MuseScore. For each rehearsal track, an MP3 file with the overall sound, and one with each of the individual voices enhanced is produced. Here are examples from the Durufle Requiem, Kyrie.
Alto Voice Enhanced : http://www.stasaphchoralsociety.co.uk/wordpress/wp-content/uploads/2022/10/Durufle-Requiem-Kyrie-Basic-Alto-Enhanced.mp3
- Tracks are produced directly from MuseScore and so do not need extra steps other than adjusting the sound levels for each of the enhanced voices.
- MuseScore is open source software (i.e. essentially free of charge) that rivals expensive products like Sibelius, and no unusual computing hardware is needed.
- But, the instruments, and particularly those of the vocal parts are rather synthetic.
- And, the vocal sounds are restricted to ‘Aah’ and ‘Ooh’.
Improving the Quality of the Voices
The first experimental step was to use sampled genuine human voices (in this case, from one of Eric Whitaker’s sets of singers). As above, MP3 files for all singers and for each voice enhanced can be produced. Again, using Durufle Requiem Kyrie, the sound is like this.
Alto Voice Enhanced : http://www.stasaphchoralsociety.co.uk/wordpress/wp-content/uploads/2022/10/Durufle-Requiem-Kyrie-Alto-Enhanced-Whitaker.mp3
- As with the basic version above, no unusual computing equipment is needed.
- With the current version of MuseScore, the output has to be transferred to a DAW (Digital Audio Workstation – in this case GarageBand on an Apple Laptop) so the production process is slightly longer.
- The vocal sounds are better, but still are restricted to those in the sample set (sounding a bit like a long ‘Mmm’)
- More exotic and flexible vocal sample sets are available, but do cost money.
Adding Vowels
Some of the feedback from the basic rehearsal tracks was that they would be easier to follow if the actual words were being sung rather than a succession of single sounds. This applied mostly when trying to isolate a particular part of the score to practice. The first attempt here switches on vowel sounds (only) and, provided the listener is familiar with the words, does give an impression of what they are.
- No unusual computing equipment is needed, and the software applications and vocal sample sets are open source.
- The output from MuseScore has to be transferred to a DAW (LMMS in this case) and a sample set of voices with each of the vowel sounds has to be available.
- Each vowel change for each voice has to be identified and assigned to the vowel in the sample set. Depending on the complexity of the score, this can take a significant amount of time and effort.
- Adding consonants to make up full words is not really practical.
Adding Full Words
A number of applications exist that include word builders, so a lyric can be typed in using native languages (English, Latin) or phonetic versions. Most of the applications tried originated from China and Japan, and were orientated to solo singers. While they did actually work, the voices just did not sound right for choral singing.
Eventually, a candidate was found (East West Symphonic Choir) that can (with a bit of skill) produce sounds like this:
- The voices are from extensive sample sets, and the words are very clear.
- The Symphonic Choir word builder software is relatively easy to use, but does need experience to get the best results (there is a lot of guidance on the Internet).
- Producing each set of rehearsal tracks probably would take about double the time of the basic set above.
- The software is relatively expensive (£170 or so), and needs a fairly heavyweight computer (ranging from around £600 for refurbished equipment to over £2,000 new).
- An audio editor (e.g. Audacity) is needed to add the accompaniment (if the audio editor bundled into Symphonic Choirs is used, the software price goes over £600).
There is a hidden message for you. Click to see it.
Views: 31