rap@dana.UUCP (Rob Peck) (05/22/87)
In article <1200@cullvax.UUCP>, csb@cullvax.UUCP (Craig Brown) writes: > > A few months ago, someone mentioned that Robert Peck was working > on some Midi Routines that would be placed in the public domain. > Were they ever released (our news system has been down for a couple > of weeks) ? Has anyone else released routines that will perform > midi functions (or should I write my own) ? > > Craig Just to clarify things, yes, I'm planning to post the PD Audio library I've developed. But it does not include Midi, since I'm not really an expert at the serial stuff. The AmigaWorld article is definitely on tap for the July/Aug issue, and I cannot "publish" the contents of the article until the magazine actually hits the stands (per my agreement with them). So when you see it there, it'll get posted, just as though some other kind soul had typed it in from the mag, as PD goes. The article mentions certain enhancements to the actual published material. Between now and the publication, I may have the time to put in those enhancements, but I do believe that the basic material in the article will be helpful and useful to those who have never tried to use the audio before. (A/W gave my address for those who want to get the listings on disk, along with the enhancements for a nominal fee... so the posting may or may not include enhancements to date, but will at least contain the article's actual printed listings.) --------------------------------------------------------------- On a related topic... at the beginning of April, I spoke to the creators of the Narrator/Translator and told them about the PD audio project. I told them I wanted to make the voice sing! (snicker, guffaw, chortle). I had suggested that I would, as does SpeechToy, continuously send "give me mouth positions" requests to the device while it is speaking, then immediately turn around and modify the device's data structures from which it was getting frequency and/or M/F information so it would be able to change its output for the next mouth position. I realized, of course, that there could be multitasking problems but as we all know, sometimes people turn off multitasking for "important things" (and some even turn it back on again). We discussed it at some length, and it would seem to be do-able if these weren't internal data structures. They also seemed to think that it might require the official support of C-A so that when 1.3 (2.0?) of the operating system came out, the technique would still work. But then, again, why not, if necessary, hack a new version for the new OS, as long as the device still exists. Besides, how high a priority can there be if the device already works (who's gonna fix it if it ain't broke?). So anyway, the technique, from a user standpoint, would be: Send a phrase to the translator and narrator, Use SpeechToy method to get back a list of mouth X,Y size values for that phrase. User examines the mouths (perhaps visual sequencing) to compare mouth positions to words; assigns pitch variation points to mouth positions (multitasking or not, the same total number of mouth position changes comes back for a given phrase... they did at least confirm that.) Save this encoding in some (IFF?) form Playback version sends the phrase to the device, waits in a loop for the mouth info, compares it to the when-should-I-change the-device's parameters, and sings. I have a few other things I have to work on, but I thought somebody else with a bit more experience at disassembling (if need be) devices and so on might be looking for a screwball hack project to work on. (or maybe convince CATS this is a worthwhile hack... maybe to release relevant data structures and how to find them once loaded...?) (I brought this topic up when recently I gave a talk about the PD Audio project and somebody said "Oh Please, when you demo it [singing], don't have it do 'Daisy, daisy, g i v e m e y o u r rrrrr....' (I'm sorry, Dave, I'm afraid I can't do that!) but then that was PRECISELY the first thing that I wanted it to do. His response, understandably, was aaaaaaaaaaaaarrrrrrrrrrrrrgggggggggggghhhhhhhhhh!!!) Or, of course, if nobody else is interested, then this was just good for a laugh. C'est le vie.