Contracted with various clients that generated patent pending intellectual property. Licensing of MuseMagic intellectual property. Designed a patent pending watermarking algorithm in two weeks and a low bit rate Vorbis LGPL audio codec profile in one week. Created a second bitstream-watermarking algorithm that is encoder and decoder independent, yet the watermark is bound to the audio signal. Performed significant porting, modifications and evaluation of the MPEG-4 audio VM in various stages of development. Created audio strategy for clients regarding MPEG-4 products. Created a real time streaming audio research test bench involving MFC , Fortran, Microsoft mmsystem calls, custom DSP libraries and optimization on Intel architecture. Research Test bench is a Fortran based DLL with I/O control to and from GUI. Features include multiple sampling rates; input sliders for Fortran based DSP manipulation, mixed language programming and interface. Responsible for a "C only" optimization of MPEG2- AAC audio decoder that improved the code 7% P5-133 from what AT&T experts had already accomplished. Performed research and architecture for an interactive voice solution in a networked virtual environment by utilizing existing VoIP standards and techniques. Researched ZISC neural network technology for related business plan and represented client to Fortune 100 Corporation as technical expert. Performed research and development for improvements in a virtual classroom audio system. Created a multi-channel, multiple outputs per channel digital delay in ADI Sharc and BlackFin assembly. Investigated the OS X kernel architecture and performance metrics for a Pro Audio firewire CoreAudio driver. Architected an AC-3 to AAC transcoding algorithm. Responsible for all aspects of business including obtaining subcontractors, legal council, accounts, taxes and business model of licensing, network and code base.
Represented major corporations in the MPEG-4 ISO/IEC and AES SC0604 international standards bodies, performed technical due diligence of audio startups, researched and presented technical strategy , created various invention ideas and research generating intellectual property owned by Nokia. Invented a frequency domain based embedded MPEG-4 audio watermarking and MIDI error correction scheme for streaming UDP delivery mechanism. Research into scalable synthesis on a thin client limited MIPS platform. Performed lossless compression literature and IP research, including DVD lossless profile, DRM research, MPEG2-AAC FHG liaison. Evaluated Audio compression technologies for price point, robustness, audio quality and product readiness and decision, for Liquid Audio. Was the Technical strategist for the inclusion of MP3/AAC algorithm in downloadable music product. Envisioned Digital Rights Management being the revenue generator for downloadable music and started a group to standardize DRM within MPEG. Performed the initial audio on demand technical strategic research for Nokia Corporation.
Created a C optimization of the AC-3 decoder utilizing architecture and compiler knowledge that performed faster than the Zoran assembly optimized version in two months. Responsible for optimization and general maintenance and bug fixes of all DSP filter modules, including dynamic EQ , Sample Rate Conversion and a Compressor/Limiter and Designed a pitch shift algorithm and implemented physical modeling synthesis in Pentium Assembly language. CCRMA liaison to obtain information and training on Stanford's physical modeling intellectual property and research.
Created major C and assembly language routines, including ACELP search, for the ITU G.723 speech coder floating point standard. Designed assembly optimization implementation of G.723 floating point on the Pentium Processor. Designed and implemented a floating point version of the ETSI GSM speech coder. Nokia Corporation licensed GSM from Intel. Optimized in C and assembly the floating point GSM for Intel Architecture. Reverse engineered an incorrect GSM implementation and further enhanced Intel's' Proshare TM GSM implementation utilizing a Texas Instruments TMS320C31 DSP . Designed re-entrant Ring 0 drivers of speech compression algorithms to operate in Intel's NSP, a RTOS based on IA-SPOX. Implemented G.722 wideband ADPCM and dual-tone generation in AT&T 1600 series DSP Q15 fixed point assembly language.
Created a VLIW audio instruction set design for emerging chip architecture. audio hardware interface, audio application and audio algorithm selections and profiling of AC-3, MPEG-1, Layer III audio compression algorithms. Design was based on VLIW instruction speed up for more computationally intensive algorithm modules and MOT56k architecture. Designed an audio board that utilized an AT&T 1610 DSP and a T7525 linear codec. Redesigned minimal filter hardware and software interface of AT&T 1614 DSP and CXD2570Q Sony codec. Designed software and hardware via a COP840C microcontroller to implement an international telephone dialer for the 1545 Telephone Answering System. Included in hardware board design was clocking, power system, overseeing of PCB board layout, and resistor ladder for dual tone generation interface. Designed Validation VHDL for DCT, IDCT, Motion Estimation, Frame Differencing Blit Assist VHDL computation units and regression testing of custom video VLSI processor. Owner of C based microcode assembler. Experienced in microcode programming.
Performed experiments, design and construction of circuits and software algorithms, measurement automation pertinent to Magnetic Research. Project include Guassian field measurement automation that utilized milling, PC interrupts, stepper motors and digital logic, high frequency biasing of transistors to magnetize sample domain walls, coil construction, complete building of needed electronics for measurement purposes. Created multiple DSP boards for digital telephone serial expansion and processing.