Kurzweil Music Systems | New Kurzweil PC3K combines game-changing technology with K series sample library
March 17, 2010
Source: Kurzweil Music Systems — December, 2009 | KMS Staff
Kurzweil Music Systems proudly announces the release of the PC3K keyboard, expanding the PC3 line to include key new features along with backwards compatibility with the legendary K Series product line.
In addition to all of the innovative features that made the PC3 a major success, the PC3K features 128 MB of non-volatile user sample memory. Utilizing a breakthrough in sample flash technology, the PC3K allows user samples to remain intact after a power-cycle, with zero load time upon powering back on. The PC3K can load .WAV files, but more importantly, it can load legacy Kurzweil .K files from the K2000, K2500 and K2600 keyboards. This opens up a whole new world of sonic possibilities; users can now combine the PC3K’s unmatched Dynamic V.A.S.T. synthesis engine with the staggering library of K Series samples generated by users and developers alike for over 15 years.
Along with K Series samples, the PC3K will support the importing of most K Series keymaps, programs and setups, providing seamless integration for existing K series customers and an enticingly large collection of sounds for new Kurzweil users to explore. From Broadway performers and high profile pop stars to working players, the PC3K will allow musicians who’ve relied on Kurzweil for years to continue using their sounds with the next generation of our technology, without having to start over from scratch.
Mike Papa, Kurzweil National Sales Manager for American Music and Sound, couldn’t be more thrilled.
“This is the breakthrough the industry has been waiting for – the added value of 18 years’ worth of K Series sound development is simply immeasurable. It’s mind-blowing to think that artists can continue to build upon the body of work they’ve been creating for just under two decades, when most other manufacturers require users to re-invent the wheel each time a new platform is introduced.”
In addition to the ability to load samples, the PC3K offers basic sample editing and full keymap editing. Of course, all of the features and presets that made the PC3 so popular are present as well; anti-aliasing VA oscillators, improved KB3 Mode, Classic Keys, Orchestral and String Sections sound sets, the renowned Kurzweil Grand Piano, Cascade Mode, 16 independent Riffs and arpeggiators, Song Mode and QA mode, unparalleled MIDI control, and the list goes on.
Another major improvement added to the PC3K is the inclusion of a USB host port to support the use of thumb-drives for data loading and storage as well as OS updates.
Comments (1)
by Ben Rayfield
Yes it has many practical features, but its not hard to beat the competition when music games with very little interaction, like Guitar Hero, profit billions of dollars.
Of course Kurzweil is just doing whats practical, which is to make more money sooner instead of advance technology more in the long-term, but Kurzweil, Guitar Hero, and anyone who would waste 10 years of their life for a billion dollars, are in for a very big suprise…. years from now. I don’t know how many, but I’ll summarize the simple plan that is so logically correct it can’t fail…
My existing prototype ( free and open-source at http://sourceforge.net/projects/audivolv ), I hereby challenge Kurzweil’s music systems to the following: It will evolve an interesting sound that his systems can not generate, and the AI evolves the code. The user just knows what they like and how they like to interact with it.
That’s still far below the level of “Guitar Hero” as judged by amount of profits. A few small improvements and I’ll have it using that same musical-instrument evolution, not on musical instruments, but on electricity patterns between simulated neurons in simulated brains, or more advanced “bayesian networks” and similar exponential types of network (instead of linear neural networks that are well known to not be able to remember certain things about 2 coin flips).
Existing neural-network-like AI research has tried to copy brains or brute-force use more speed. I’m more concerned with accuracy, and I intend to simulate electricity patterns between nodes at 44.1 khz audio quality that you can play on the speakers and hear it buzzing like neurons. My neural-network simulations will be so detailed that you could connect an Emotiv Epoc or OpenEEG or other mind-reading device (for video games they now sell them) and it will be able to communicate with my evolving simulation of a brain… if you added a software-to-analog converter to connect it to that game controller. You get the point. Music quality is a thousand times more detailed granularity of time than most neural networks are simulated at. Between 2 firings of the same neuron, there could be 100,000 numbers in the “audio stream” (or more accurately a “neuron spiking pattern stream”). I have some realtime compiling and permutation optimizations to make it run fast enough.
It really comes down to this. Nobody except me cares enough about the detail of the electricity patterns to throw an algorithm at it that can evolve from whitenoise and radio-static-like sounds to more music-like sounds. If it can do that, it can evolve neuron spiking easy.
A billion dollars is toilet paper compared to that. Instead, the only practical thing for me to do is give it away as open-source software. Try the early version, and lets get an intelligence-related contest going between what I have now (it works, try it) and Kurzweil’s feature-rich keyboard with AI “on the side” added as an afterthought.
I mean no disrespect to Kurzweil though. He’s just making some money. He’ll get back to serious work later.