Vocal line rendering via command line.

General discussion about chipspeech

Moderators: eric_telemaque, davidv, seb@plogue

jbum
Posts: 5
Joined: Thu Jan 15, 2015 1:32 pm

Vocal line rendering via command line.

Postby jbum » Thu Jan 15, 2015 1:59 pm

First of all, I'm LOVING chipspeech. I can imagine using it for all kinds of projects. I have two questions.

Question 1)

I have a question about automation. I'm an experienced programmer & experimental composer. I have a project in which I am producing choral music by mixing together individual vocal lines using a variety of voices with different settings (similar to the "stack" in chipspeech, but trying to get more variety, and producing 4 SATB choral sections). The music is produced algorithmically (based on a 1650 music algorithm by Athanasius Kircher, the Arca Musarithmica) using scripts, and I can prepare exactly the phonemes I want for each note, as well as produce (via scripts) MIDI files to control the plugin. Given those pieces of input -- I would love to be able automate the rest of it via the command line, since individually rendering 20-30 voices with randomized settings is tedious (especially because my algorithm is capable of producing a LOT of output, and I tend to do a lot of tests where I am changing lots of parameters).

My goal would be to construct a command tool something like this:

$ renderVocalLine notes.mid lyrics.txt [optional settings here] -o sample.wav

I would then run this 30 times and mix the results together, also via the command line.

I was curious if anyone had any ideas about technologies I might use to accomplish this. I can imagine maybe setting this up via Apple script, but if there is a command-line scriptable plug-in host that can do rendering of MIDI, that would be better...

Question 2)

Is there a way to directly input X-Sampa lines via MIDI of some other form of automation? The texts for this music are in latin and would be easier if I can provide entire lines...

Hubert
Posts: 3
Joined: Thu Jan 15, 2015 4:32 pm

Re: Vocal line rendering via command line.

Postby Hubert » Thu Jan 15, 2015 4:54 pm

I'm sorry, I don't think that will be possible. Chipspeech is designed to be used as a VST/AU/AAX plugin, rather than a command line tool. So it can automate many things, but not the lyrics contents, since automating the text position (using CC2 to set line number and CC3 to set the position within the line) is almost always sufficient, and it's not common to have to change the text so many times so often.

jbum
Posts: 5
Joined: Thu Jan 15, 2015 1:32 pm

Re: Vocal line rendering via command line.

Postby jbum » Thu Jan 15, 2015 5:44 pm

Right now I'm looking hard at that "Paste All" button, which perhaps I can automate via Applescript. I can definitely load up the clipboard from Perl or Python...

UPDATE: I've gotten scriptable paste working. FYI, this is a janky Mac-centric solution. Ran into a brick wall with Applescript and Automater, but an app called "Keyboard Maestro" works pretty well. I can initiate a script from Perl/Python that will open the stand-alone ChipSpeech app, and press the "Paste All" button to get my lyrics/phonemes in there. Operating the "Render Offline" dialog is going to be even jankier, but it looks to be doable, and I *think* I'll be able to parameterize my filenames. Would be nice if there were Applescript support for the front-panel buttons, and a high-level applescript event for "Render Offline" where you supply the input and output files. Also would be nice if "Render Offline" would accept an optional lyrics file (text file).


Return to “chipspeech General”

Who is online

Users browsing this forum: No registered users and 1 guest