gibson@unc.UUCP (Bill Gibson) (11/01/85)
> Subject: what's the minimum cost of a good MANY-track set-up? > > What I want - MANY tracks of digital-sampled instruments. > The ability to mimic a decent sized orchestra. Screen/music-notation > oriented editor. And be capable of storing long/dense sequences. > MANY = at LEAST 32 tracks. Go to the Florida State Music Department and beg for time on their university machine (I think it's a big CDC) and associated custom sound hardware. They were doing this sort of stuff around 1980, so they may still be ahead of the average Synclavier-equipped studio. I don't think this exists on micros; I recently got MIDI recording software which was presented to me as the 'latest', and it falls short of your (and my) desired features. I'll admit, it is for a Commodore 64 ("what, you want function out of a 6502?"), and it does have lots of useful features (more than some earlier software I have used). However, I class this program and all the others I have heard about lately as "2nd generation" MIDI software, and I doubt we'll have a quality software system until the 3rd or 4th generation (e.g. 4th = 2nd set of programs done for large-memory 680x0/80x86 micros). I'm not sure what you're getting at with the many-tracks-of-sampling point. MIDI isn't any good for real-time transmission of samples, but "system-exclusive" (i.e. manufacturer-dependent) messages can be used to send stored samples to/from your dedicated sampling box (mirage/akai s612). (Note - somebody here just told me that there were rumors at the last trade show - NAMM? - of standardizing the MIDI codes used for sample dumps. With the number of sampling units out/coming out, this would be a win for the manufacturers and voice-librarian writers.) But the sampled/algorithmically generated sounds will probably all come out of dedicated boxes for the near future (excepting the Amiga). "Mimicry of a decent-sized orchestra" depends on how accurate you want the mimicry to be. Hard-to-distinguish mimicry will probably remain astronomically expensive for several years. The next couple of generations of sampling hardware should allow different samples to be played from different MIDI channels or different keyboard split areas. This is necessary for orchestal-style works, unless you have Lots of tape tracks. But Real Orchestra Music is very complex (in terms of slightly shifting frequencies, harmonic content, and tempo variation) and "good" simulations probably would require hours on large mainframes. I think this is done at the research center in Paris. Screen-editing of music notation - the only excuse for the fact that products don't have this feature is that the publisher wanted to go ahead and get to market. I think most recording software will quantize notes for you, and using the quantized representation for visual editing should be a simple step. Having used Hayden's MusicWorks on a Mac, I realize that it will also be another generation or 2 before editors have the nice little touches that make them Really useful (e.g. search-for-pattern, rhythm and pitch information accessible separately, play (n) last notes as you enter, marking events not representable on a graph (e.g. voice change)). "I have to do everything around here" may be applicable - I may have to write the blasted MIDI patcher/editor/recorder/mumblehack in order to get what I want. And I think it can only be done on something approaching a real hardware system (520ST,Amiga,Mac,IBM (yuck)). 32 tracks - maybe the new MIDI package for the Mac or the stuff coming out for the 520ST/Amiga will have this and all the other neat features. My Syntech Studio 1 C64 program organizes the data into 16 sequences, each with 8 tracks of recording. The 3rd and highest level of abstraction is the song, which consists of an arbitrarily-ordered list of up to (32 I think) sequences. You can bounce tracks indefinitely within a sequence, but there is only 1 MIDI channel per track (bogus). Also, I haven't figured out a nice way to transfer an isolated track from one sequence to another - it's beginning to look as if the only way is to append one sequence to another and then shift the desired track back (96 clock pulses at a time - takes about a minute) to coincide with the start of the target sequence. Unbearably cumbersome. But I think that the organization of recorded MIDI data wouldn't be as useful if it were on a flat 32 tracks. The grouping of "tracks" into "sequences" is a useful abstraction, and all would be bliss if I didn't have these brick-wall boundaries between sequences. > I can't afford 10 mirages. Any other ways around this? > Sync-to-tape? Sync-to-tape is a useful feature, but it eats a tape track and requires multiple passes to accomplish what would be better done with a multi-voice sampling box. Perhaps the engineers/musicians working on the next samplers will put in enough memory to allow multiple independent samples to be played at once (just like 10 mirages!). Bill Gibson gibson@unc ...[akgua,decvax,philabs]!mcnc!unc!gibson
tynor@gitpyr.UUCP (Steve Tynor) (11/01/85)
In article <440@unc.unc.UUCP> gibson@unc.UUCP (Bill Gibson) writes: >> Subject: what's the minimum cost of a good MANY-track set-up? >> >> What I want - MANY tracks of digital-sampled instruments. >> The ability to mimic a decent sized orchestra. Screen/music-notation >> oriented editor. And be capable of storing long/dense sequences. >> MANY = at LEAST 32 tracks. > > Go to the Florida State Music Department and beg for time on their >university machine (I think it's a big CDC) and associated custom sound >hardware. They were doing this sort of stuff around 1980, so they may still >be ahead of the average Synclavier-equipped studio. This is the PLATO synthesizer, developed at the University Of Illinois. It (in it's simplest form) is a 16 voice (oscillator) instrument with screen notation and a very nice editor. I worked on the one up at Illinois, and while the editing features are great, the sound falls short of orchestral mimicry. In order to do FM or additive synthesis you use up 2 (or more) oscillators... > "Mimicry of a decent-sized orchestra" depends on how accurate you >want the mimicry to be. Hard-to-distinguish mimicry will probably remain >astronomically expensive for several years. The next couple of generations >of sampling hardware should allow different samples to be played from >different MIDI channels or different keyboard split areas. This is necessary >for orchestal-style works, unless you have Lots of tape tracks. But Real >Orchestra Music is very complex (in terms of slightly shifting frequencies, >harmonic content, and tempo variation) and "good" simulations probably would >require hours on large mainframes. I think this is done at the research center >in Paris. Take a listen to Walter (sorry Wendy) Carlos's Digital Moonscapes. It's recorded entirely on the DK SYNERGY and GDS synthesizers. It's about as orchestra like as I've heard and the SYNERGY is basically just a Z80 + a custom 16bit waveform processor. Also, the price falls far short of what I'd call 'mainframe' prices: ~$5-6000. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= No problem is so formidable that you can't just walk away from it. Steve Tynor Georgia Instutute of Technology ...{akgua, allegra, amd, harpo, hplabs, ihnp4, masscomp, ut-ngp, rlgvax, sb1, uf-cgrl, unmvax, ut-sally} !gatech!gitpyr!tynor
gibson@unc.UUCP (Bill Gibson) (11/02/85)
In article <967@gitpyr.UUCP> tynor@gitpyr.UUCP (Steve Tynor) writes: >> But Real >>Orchestra Music is very complex (in terms of slightly shifting frequencies, >>harmonic content, and tempo variation) and "good" simulations probably would >>require hours on large mainframes. I think this is done at the research center >>in Paris. > >Take a listen to Walter (sorry Wendy) Carlos's Digital Moonscapes. It's >recorded entirely on the DK SYNERGY and GDS synthesizers. It's about as >orchestra like as I've heard and the SYNERGY is basically just a Z80 + a >custom 16bit waveform processor. Also, the price falls far short of what >I'd call 'mainframe' prices: ~$5-6000. I haven't heard Digital Moonscapes yet, but I have read reviews which say it sounds very orchestral. Agreed, the IC hardware cost may fall short of mainframe prices; however, from the (June?) interview with Wendy Carlos in Keyboard magazine, it sounds as if *lots* of human-brain processing time was spent getting those sounds. Perhaps the work done on the album will generate hardware or software for the rest of us to use in getting "natural"- type sounds. Bill Gibson gibson@unc ...[akgua,decvax,philabs]!mcnc!unc!gibson
tynor@gitpyr.UUCP (Steve Tynor) (11/04/85)
In article <493@unc.unc.UUCP> gibson@unc.UUCP (Bill Gibson) writes: >In article <967@gitpyr.UUCP> tynor@gitpyr.UUCP (Steve Tynor) writes: >>Take a listen to Walter (sorry Wendy) Carlos's Digital Moonscapes. It's >>recorded entirely on the DK SYNERGY and GDS synthesizers. It's about as >>orchestra like as I've heard and the SYNERGY is basically just a Z80 + a >>custom 16bit waveform processor. Also, the price falls far short of what >>I'd call 'mainframe' prices: ~$5-6000. > >I haven't heard Digital Moonscapes yet, but I have read reviews which say >it sounds very orchestral. Agreed, the IC hardware cost may fall short of >mainframe prices; however, from the (June?) interview with Wendy Carlos in >Keyboard magazine, it sounds as if *lots* of human-brain processing time >was spent getting those sounds. Perhaps the work done on the album will >generate hardware or software for the rest of us to use in getting "natural"- >type sounds. I have a SYNERGY, and Carlos's voicing cartridges are available to the consumer. True, s(he) put in lots of thought power to derive those voices, but with her cartidges, it's really not that difficult to get great string sounds. (as well as bassoon, flute, great pipe organ, brass, you name it.) =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Some things have to be believed to be believed. Steve Tynor Georgia Instutute of Technology ...{akgua, allegra, amd, harpo, hplabs, ihnp4, masscomp, ut-ngp, rlgvax, sb1, uf-cgrl, unmvax, ut-sally} !gatech!gitpyr!tynor