In-band Modem for Non-speech Data Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech codecs are inefficient in transmitting non-speech data due to their design, which is optimized for speech signals, leading to distortion and low data rates when trying to encode signals without periodic or noise-like characteristics, such as tone signals or DTMF signals.
Innovation Solution
A method and apparatus that use an in-band modem to process input data symbols into pulse signals, shape them, and encode them with a speech codec, generating noise-like signals that can be effectively transmitted through a speech codec, including synchronization signals to aid detection and demodulation at the receiver.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional speech codecs are used to transmit non-speech data, then the transmission can be performed through existing infrastructure, but the data rate is low and distortion is severe
Solution Approach 1:
The patent transforms the fundamental parameters of the transmitted signal by converting data into noise-like signals with statistical properties matching speech signals. This includes modifying the signal distribution to have Gaussian characteristics and adjusting the spectral content to match speech frequency profiles, enabling the speech codec to process non-speech data effectively while maintaining acceptable data rates
Solution Approach 2:
The patent introduces an intermediary transformation process that converts digital data into intermediate noise-like waveforms before speech codec encoding. This intermediary representation acts as a bridge between digital data and speech signal processing, allowing the codec to handle non-speech content without severe degradation
2Adaptability or versatility
If traditional speech codecs are used to transmit non-speech data, then transmission through existing speech infrastructure is possible, but distortion is severe
Solution Approach 1:
The patent modifies the statistical parameters of the input signal to match speech signal characteristics. By transforming data into signals with Gaussian amplitude distribution and speech-like spectral density, the speech codec processes the signal with minimal distortion, preserving signal fidelity throughout the transmission chain
Solution Approach 2:
The patent converts the inherent limitation of speech codecs (designed for speech only) into a benefit by exploiting their noise-processing capabilities. By deliberately transforming data into noise-like signals, the patent makes the data invisible to the speech-specific processing pathways that would otherwise cause severe distortion, thereby preserving signal fidelity
3Device complexity
If tone signals are transmitted through speech codecs, then data can be transmitted using simple modulation schemes, but the vocoder cannot effectively model the tones resulting in detection errors
Solution Approach 1:
The patent fundamentally changes the parameter structure of the transmitted signal from deterministic tone waves to stochastic noise-like signals. This transformation maintains the simplicity of the transmission system while dramatically improving detection reliability, as the noise-like characteristics are naturally accommodated by the speech codec's processing architecture
Solution Approach 2:
The patent replaces the mechanical tone modulation approach with a statistical signal transformation method. Instead of using precise frequency and phase modulation of sinusoidal carriers, the system uses statistical modeling to generate noise-like signals with speech-matching properties, thereby improving detection accuracy without increasing system complexity
4Productivity
If data transmission rate is increased by changing tones quickly, then higher data rates are achieved, but vocoder modeling ability is degraded further resulting in more data errors
Solution Approach 1:
The patent employs periodic framing and structured noise-like signal generation that aligns with the speech codec's frame-based processing architecture. By organizing data into periodic structures that match the codec's temporal processing intervals, the system achieves higher effective data rates while maintaining reliability through synchronized processing boundaries
Solution Approach 2:
The patent changes the temporal parameters of signal generation to match the speech codec's processing rate. By generating noise-like signals with statistical properties that are stable over the codec's frame duration, the system enables faster effective data transmission without degrading vocoder modeling ability, thereby reducing data errors
Data Source
Figure 1
Figure 2
Figure 3A~3C
AI summary
A system is provided for transmitting information through a speech codec (in-band) such as found in a wireless communication network. A modulator transforms the data into a spectrally noise-like signal based on the mapping of a shaped pulse to predetermined positions within a modulation frame, and the signal is efficiently encoded by a speech codec. A synchronization sequence provides modulation frame timing at the receiver and is detected based on analysis of a correlation peak pattern. A request/response protocol provides reliable transfer of data using message redundancy, retransmission, and/or robust modulation modes dependent on the communication channel conditions.