I know this isn't Mr. G's forte, but has anyone thought of using this technology for Solo Guitar? Are there great solo guitar samples out there? If so, would they be improved by Mr. G's upcoming Solo Violin technology?
I've got a DSP background, so I know about the theory. I don't know all of the tricks involved in this particular lib, though.
The morphing itself is best for solo instruments that sustain and can swell for a different timbre. Violin is perfect. Horns and woodwinds will be great too. The human voice will probably be awesome. But guitar is an attack/decay instrument. It may not benefit from morphing the way the others do.
That said, guitar is also a vibrato instrument. It's possible that morphing can be used to let you control the vibrato beyond what happens with a mod wheel. Maybe it can be used for bent strings as well.
The real place where guitar will benefit is the MIDI processing for playability. There are so many little inflections in playing lead guitar - hammer ons, slides, mutes, micro stretches, pull offs, no-pick-legato. If these can be mapped to a keyboard in a playable way, that will be awesome!
To me a great sampled guitar needs great articulations and a great way to trigger them. For doublestops it needs to use amp modeling and such. For single lines, recording direct from the amp is fine. I'd prefer a few options. Like libs with multiple mics, I'd like a guitar lib with multiple amps - including the clean option.
If anybody with inside knowledge of Gary's Solo Strad has other insights, let us know!
I essentially agree with Jon. The guitar would not probably gain much advantage from the so called Sonic Morphing, or phase alignment process, since each note mainly consists of an attack/decay. However, phase alignment of adjacent samples may yield smoother transitions when dealing with a limited number of sampled dynamics. In other words, if you've got each note sampled over only four dynamics, you may obtain virtually infinite dynamic levels by layering the corresponding phase aligned samples, their relative amplitudes being proportional to note-on velocities. The same approach will be applicable to any plucked instrument, including piano.
Conversely, guitar samples may gain substantial advantage by the proprietary modal resonance convolution process, developed by myself and Stefano Lucato. In this case, a calculated "anechoich" sound is convoluted by an algorithmic impulse response based on the actual modal resonances of the instrument body. The same principle has been applied to the Stradivari violin with excellent results for vibrato, portamento and note articulations.
The infinite velocity layers concept will be really wonderful for piano, harp and guitar - especially the classical guitar. Applying resonance to obtain vibrato and other articulations was something I hadn't considered. Very cool.
For electric lead guitar the larger challenge remains the playability. Recently, I've been watching my playing closely and thinking about what it would take to sample my lead guitar. I amaze myself with the subtle phrasings and articulations that I use to make the thing sound musical. And I do it all subconsciously! (Not that I'm a virtuoso. Rock lead simply demands such stylings.)
From the video of the Solo Strad, you're fully up to the challenge, should you every apply your work to the electric guitar.
Well, most of the character of the guitar comes in the attack- the picking part. I would think sonic morphic would allow you to have dozens of picks and relatively few sustain portions of a note. Then you can round-robin the picks and have them x-fade into one of a handful of sustains, with and without vibrato. It would be way less byte-hungry than sampling all those notes for the duration. Plus I would think legato techniques like slides and hammer-ons/offs would benefit the same way violin does. But whadda I know?